{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":176,"total_is_capped":false,"direct_labels_cover":44,"predictions_cover":176,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"query_hash":"ac8d0f485ba1","filters":{"venue":"Scientometrics"}},"results":[{"id":"W1767272795","doi":"10.1007/s11192-015-1765-5","title":"The journal coverage of Web of Science and Scopus: a comparative analysis","year":2015,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":4652,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Scopus; Web of science; Directory; Computer science; Scientometrics; Bibliometrics; Data science; Library science; Information retrieval; Political science; MEDLINE","retraction":null,"screen_n_in":null,"score":{"opus":0.6780420299185497,"gpt":0.6098436100295971,"spread":0.06819841988895259,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.1792655,0.000136993,0.0005580933,0.2540573,0.0007816073,0.002532158,0.004579442,0.00005865651,0.00003497793],"category_scores_gemma":[0.1904339,0.0000739299,0.0001518586,0.8145463,0.005190757,0.001076719,0.001376764,0.0002796631,0.00001833525],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002467443,"about_ca_system_score_gemma":0.003039995,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005486572,"about_ca_topic_score_gemma":0.0000245955,"domain_scores_codex":[0.964281,0.0004121839,0.001175953,0.0005176609,0.03291111,0.0007021165],"domain_scores_gemma":[0.9469301,0.01003361,0.001214166,0.001055681,0.03948574,0.001280779],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001534978,0.0004063345,0.8360892,0.00000975065,0.0002341162,0.000006895246,0.003866299,0.001833245,0.004585866,0.005096162,0.0175405,0.1301782],"study_design_scores_gemma":[0.002708204,0.001792103,0.766818,0.00002135183,0.0002694156,0.00005672234,0.01639619,0.1313935,0.01374642,0.01200018,0.05426313,0.0005347896],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9759558,0.003999752,0.004612288,0.000204175,0.0007329189,0.0001741945,0.00003793978,0.000005396721,0.01427749],"genre_scores_gemma":[0.9978505,0.0007961409,0.0008597731,0.00002169246,0.00002108353,0.00000101953,3.463934e-7,0.000002841143,0.000446605],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5604889,"threshold_uncertainty_score":0.9985033,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2165122403","doi":"10.1007/s11192-006-0115-z","title":"Benchmarking scientific output in the social sciences and humanities: The limits of existing databases","year":2006,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":506,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Benchmarking; Bibliometrics; Benchmark (surveying); Database; Bibliographic database; Computer science; Data science; Humanities; Library science; Geography; Management","retraction":null,"screen_n_in":null,"score":{"opus":0.8593883432201264,"gpt":0.609782680977184,"spread":0.2496056622429423,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics","sts"],"category_scores_codex":[0.1096677,0.0001949995,0.0003290084,0.09872509,0.003340084,0.008372738,0.005001927,0.00006397161,0.00005347781],"category_scores_gemma":[0.0424663,0.00009916171,0.0001264725,0.4899217,0.004215502,0.001024178,0.001090991,0.0003571564,0.00002100305],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008810357,"about_ca_system_score_gemma":0.0003217917,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006184105,"about_ca_topic_score_gemma":0.000325578,"domain_scores_codex":[0.9785785,0.0006287548,0.001089237,0.001084996,0.01760306,0.00101544],"domain_scores_gemma":[0.9792731,0.01692205,0.0006137859,0.0008311887,0.002273005,0.00008680503],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00001396402,0.0005389262,0.6458755,0.00003892137,0.000009499257,0.00002085178,0.005744464,0.0002930244,0.0008386985,0.1505293,0.04421008,0.1518868],"study_design_scores_gemma":[0.0005538083,0.0002009237,0.8460857,0.00003439848,0.00001493752,0.00002083643,0.01289672,0.008677776,0.0006360377,0.01239758,0.1181039,0.0003774033],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9622965,0.002488917,0.0007473676,0.0009150481,0.0009381889,0.0003670611,0.00007395211,0.00001454747,0.03215842],"genre_scores_gemma":[0.9974456,0.00004071513,0.000840781,0.000138778,0.0002049772,0.00001050257,0.000006302359,0.000006962935,0.001305384],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3911966,"threshold_uncertainty_score":0.9984944,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2115701185","doi":"10.1007/s11192-007-2036-x","title":"History of the journal impact factor: Contingencies and consequences","year":2009,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":332,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Université du Québec à Montréal","funders":"","keywords":"Impact factor; Citation; Bibliometrics; Computer science; Factor (programming language); Data science; Library science; Operations research; Political science; Law; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.6874658205142047,"gpt":0.5792682361347077,"spread":0.108197584379497,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.02947145,0.0001613262,0.0003517279,0.08029769,0.0003114744,0.001107263,0.003154221,0.00009922497,0.0004132398],"category_scores_gemma":[0.07822573,0.00007991691,0.0002324307,0.2551544,0.001261225,0.0007456327,0.000433842,0.0003877462,0.00001763452],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000628156,"about_ca_system_score_gemma":0.001159677,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005337417,"about_ca_topic_score_gemma":0.000004104567,"domain_scores_codex":[0.9849877,0.000368035,0.0008078605,0.0004022669,0.01279821,0.0006359106],"domain_scores_gemma":[0.9897329,0.003625903,0.0007477188,0.0006658464,0.004657912,0.0005696746],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00002415601,0.0001428063,0.5738164,0.000003878846,0.0000179698,0.000006182272,0.001584694,0.00002667312,0.0170973,0.0008851675,0.06077237,0.3456224],"study_design_scores_gemma":[0.0002984849,0.0003814661,0.9572282,0.000008484637,0.000007951964,0.00007679209,0.0003609042,0.0006048303,0.001266428,0.005663089,0.03395575,0.0001475543],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9750594,0.01293437,0.0008045767,0.0007212727,0.00162887,0.0001633105,0.00002893101,0.00001215818,0.008647101],"genre_scores_gemma":[0.9953232,0.0006607349,0.0003591951,0.0001715631,0.0000451232,3.211648e-7,1.71244e-7,0.000004073224,0.003435623],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3834119,"threshold_uncertainty_score":0.9999297,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2038196424","doi":"10.1007/s11192-013-1221-3","title":"Coverage and adoption of altmetrics sources in the bibliometric community","year":2014,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":266,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Altmetrics; Bookmarking; Social media; Citation; Tracking (education); Scopus; World Wide Web; Download; Computer science; Internet privacy; Psychology; Political science; MEDLINE","retraction":null,"screen_n_in":null,"score":{"opus":0.5482835610101827,"gpt":0.547941311386074,"spread":0.0003422496241086703,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.1713287,0.0002381103,0.000558719,0.7838975,0.0006015102,0.002966345,0.005220999,0.0001921364,0.00006390418],"category_scores_gemma":[0.2467326,0.000145968,0.0001639573,0.9585446,0.0006722743,0.0009260565,0.001274417,0.0009095567,0.00008467943],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001134444,"about_ca_system_score_gemma":0.0001189067,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004068466,"about_ca_topic_score_gemma":0.00004612621,"domain_scores_codex":[0.9744112,0.00269613,0.001423278,0.0007683093,0.01974063,0.0009604779],"domain_scores_gemma":[0.9366072,0.05465752,0.000934993,0.002370915,0.004849922,0.0005794386],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00001223955,0.0004187768,0.6253552,0.00002235055,0.000005348365,0.000002003356,0.0007141193,0.0002022141,0.0001146086,0.003432768,0.001593336,0.368127],"study_design_scores_gemma":[0.0007053017,0.000458395,0.9631696,0.000009004018,0.000007458251,0.00001118682,0.001316197,0.006380897,0.0001797585,0.008942391,0.01862797,0.0001918741],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9725812,0.001970299,0.01310378,0.0003402451,0.0004496927,0.0003536516,0.00003240865,0.00001951674,0.01114919],"genre_scores_gemma":[0.9976739,0.0009003761,0.0008387492,0.0002738863,0.00004887113,0.000009419979,0.000006606601,0.00001199176,0.0002361867],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3679352,"threshold_uncertainty_score":0.9980687,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2143612877","doi":"10.1007/s11192-011-0580-x","title":"Validating online reference managers for scholarly impact measurement","year":2011,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":260,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia; York University","funders":"","keywords":"Citation; Computer science; Citation impact; Sample (material); Citation analysis; Web of science; Information retrieval; Scholarly communication; Impact factor; World Wide Web; Data science; Political science; MEDLINE; Publishing","retraction":null,"screen_n_in":null,"score":{"opus":0.9469149544521874,"gpt":0.6472795575330219,"spread":0.2996353969191655,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.09005316,0.0003228453,0.000503569,0.1864819,0.0006803927,0.006959795,0.005891509,0.0001992522,0.0009468686],"category_scores_gemma":[0.233558,0.0002221434,0.0004149593,0.492865,0.0002484088,0.003027312,0.001114323,0.0004992308,0.0004698049],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000882061,"about_ca_system_score_gemma":0.000587453,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000242794,"about_ca_topic_score_gemma":0.00002989674,"domain_scores_codex":[0.968701,0.0002703744,0.001249935,0.001556273,0.02663901,0.001583426],"domain_scores_gemma":[0.9707146,0.003109583,0.0006914015,0.001836327,0.02224485,0.001403194],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.000120264,0.001197896,0.2390471,0.00002817743,0.00009518251,0.00001899053,0.0006690107,0.00006849066,0.003716741,0.004157508,0.02421278,0.7266679],"study_design_scores_gemma":[0.00209964,0.001655517,0.8873448,0.00003693943,0.00004812987,0.00001396515,0.001552886,0.01046356,0.005317414,0.02425953,0.06617612,0.001031548],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8102085,0.001802986,0.1327956,0.0003402764,0.002484117,0.001569759,0.0005850005,0.0001806727,0.05003309],"genre_scores_gemma":[0.9661549,0.00008863856,0.03145721,0.0001647451,0.00009907644,0.00002988908,0.00001709604,0.00002809965,0.001960298],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7256364,"threshold_uncertainty_score":0.9999664,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2888141929","doi":"10.1007/s11192-018-2889-1","title":"Bibliometric study of family business succession between 1939 and 2017: mapping and analyzing authors’ networks","year":2018,"lang":"en","type":"article","venue":"Scientometrics","topic":"Family Business Performance and Succession","field":"Business, Management and Accounting","cited_by":216,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; HEC Montréal","funders":"","keywords":"Scopus; Scope (computer science); Ecological succession; Family business; Social network analysis; Scientometrics; Bibliometrics; Data science; Citation impact; Sociology; Citation; Computer science; Knowledge management; Library science; Social science; Political science; Management; Social capital; Economics; MEDLINE","retraction":null,"screen_n_in":null,"score":{"opus":0.07934125063747188,"gpt":0.3136777381275538,"spread":0.234336487490082,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics"],"consensus_categories":["bibliometrics"],"category_scores_codex":[0.002291296,0.0002675836,0.0004564325,0.1028561,0.0006938124,0.0008171863,0.000439399,0.0001408011,0.00001633926],"category_scores_gemma":[0.0006324339,0.0002181887,0.00003298894,0.3346522,0.0002591611,0.002674218,0.0008743663,0.0001903653,0.00001453554],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002586037,"about_ca_system_score_gemma":0.00002545187,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009169159,"about_ca_topic_score_gemma":0.00002952703,"domain_scores_codex":[0.9973231,0.00002093377,0.0005667104,0.0005974519,0.0009835189,0.0005082617],"domain_scores_gemma":[0.9977188,0.0001373704,0.000480114,0.0003525073,0.001252444,0.0000587365],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0000156641,0.0001043568,0.9063935,0.0001546945,0.00001637947,0.000002954532,0.0001099602,0.0000387522,0.000205401,0.00002239882,0.000683058,0.09225292],"study_design_scores_gemma":[0.0008596635,0.0000459627,0.9822338,0.0001419484,0.00006718189,8.289336e-7,0.0007048099,0.01322371,0.000007061009,0.00004322008,0.002397935,0.0002739014],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9936663,0.001133085,0.002635673,0.00006094065,0.0009458981,0.0003213004,0.000001838227,0.00007512839,0.001159868],"genre_scores_gemma":[0.9982703,0.0002948027,0.000194816,0.0001306473,0.001006407,0.000005682007,0.00001385074,0.00002443332,0.00005909981],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2317961,"threshold_uncertainty_score":0.9073122,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2139770987","doi":"10.1007/s11192-006-0127-8","title":"Canadian collaboration networks: A comparative analysis of the natural sciences, social sciences and the humanities","year":2006,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":215,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Digital humanities; Humanities; Natural science; The arts; Citation index; Index (typography); Sociology; Social science; Citation; Library science; Political science; Epistemology; Computer science; Philosophy; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.4360845501692481,"gpt":0.5388048553623775,"spread":0.1027203051931294,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics","sts"],"category_scores_codex":[0.04525552,0.0001739392,0.0005177922,0.1253716,0.00542787,0.008359406,0.003869884,0.00008531764,0.00007214786],"category_scores_gemma":[0.01271917,0.0000822153,0.0002550263,0.7556335,0.009273396,0.000741326,0.0006760583,0.0002627693,0.000007017853],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003015797,"about_ca_system_score_gemma":0.001416098,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.04476495,"about_ca_topic_score_gemma":0.1363648,"domain_scores_codex":[0.9856075,0.0007139585,0.0007841504,0.0007370177,0.01135666,0.0008007083],"domain_scores_gemma":[0.9883829,0.006461841,0.0006148186,0.0004410485,0.003980749,0.0001185883],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00003043764,0.0001035751,0.5593642,0.000004297252,0.000162324,0.000001600169,0.002907607,0.02553765,0.00005457737,0.3731199,0.03374213,0.00497169],"study_design_scores_gemma":[0.0005735985,0.00006390752,0.6948568,0.00000368098,0.0001239028,0.000001788615,0.003786702,0.2823058,0.00007599826,0.004547621,0.01345998,0.0002002376],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9516325,0.005537967,0.0006421328,0.004157518,0.00134029,0.0006155912,0.0001291744,0.0000155999,0.03592923],"genre_scores_gemma":[0.9976676,0.00003482786,0.0002069176,0.0003662569,0.00008276635,0.000008522271,0.000003751171,0.000002866187,0.001626534],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.630262,"threshold_uncertainty_score":0.995867,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2058498906","doi":"10.1007/s11192-011-0369-y","title":"Sex differences in research funding, productivity and impact: an analysis of Québec university professors","year":2011,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":204,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Université du Québec; Université du Québec à Montréal","funders":"McGill University","keywords":"Productivity; Scientometrics; Political science; Regional science; Sociology; Economics; Economic growth; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.872966831153213,"gpt":0.6254438142378042,"spread":0.2475230169154089,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.09664137,0.0001753771,0.0006941858,0.4405308,0.0003897616,0.0007019419,0.003181933,0.0001542178,0.0004866706],"category_scores_gemma":[0.05953104,0.0001256271,0.0001734338,0.766765,0.001225225,0.001425368,0.001410186,0.0005116217,0.00001835959],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005574236,"about_ca_system_score_gemma":0.0008836918,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.04696523,"about_ca_topic_score_gemma":0.01873166,"domain_scores_codex":[0.9827235,0.001359591,0.0006469616,0.00138475,0.01278458,0.001100574],"domain_scores_gemma":[0.9875354,0.00497048,0.0003451366,0.001278139,0.004956141,0.000914686],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00003939861,0.0005245767,0.9679577,0.000005277978,0.00005193958,0.000009488515,0.003533811,0.00001156052,0.0001515115,0.0005357205,0.0001065642,0.0270724],"study_design_scores_gemma":[0.0002270595,0.0003935306,0.9861859,0.000003647233,0.00002919876,8.675518e-7,0.004411916,0.007311497,0.0003616364,0.0007324847,0.0001932674,0.0001489997],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9955975,0.0003410447,0.0003625681,0.00009150092,0.0001520823,0.0003088051,0.00005928618,0.00001566487,0.003071569],"genre_scores_gemma":[0.9969999,0.00009431289,0.0003836482,0.000003084328,0.00001358111,0.000001911296,0.000002390196,0.000006587633,0.002494563],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3262342,"threshold_uncertainty_score":0.9991739,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":true,"about_ca_topic":true,"confidence":"low"},{"model":"gpt","categories":["metaresearch","bibliometrics"],"domain":"evaluation","study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":true,"confidence":"high"}],"label_agreement":"split"},{"id":"W3082051346","doi":"10.1007/s11192-020-03675-3","title":"Publishing volumes in major databases related to Covid-19","year":2020,"lang":"en","type":"article","venue":"Scientometrics","topic":"Academic Publishing and Open Access","field":"Decision Sciences","cited_by":152,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Thompson Rivers University","funders":"Wellcome Trust","keywords":"Scopus; Coronavirus disease 2019 (COVID-19); Web of science; Publishing; Pandemic; Library science; China; MEDLINE; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Bibliometrics; 2019-20 coronavirus outbreak; Analytics; History; Medicine; Database; Computer science; Political science; Disease; Infectious disease (medical specialty); Virology; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.2913286398736232,"gpt":0.4731993666241887,"spread":0.1818707267505655,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.02228821,0.0001910828,0.0003997858,0.008143477,0.0002977279,0.009177243,0.006798273,0.0001635333,0.001000743],"category_scores_gemma":[0.5404868,0.0001531619,0.00008809343,0.112661,0.000134435,0.01044914,0.002297983,0.0008192403,0.001162552],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002908102,"about_ca_system_score_gemma":0.0006656833,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000635189,"about_ca_topic_score_gemma":0.0001183263,"domain_scores_codex":[0.9908996,0.0003495225,0.001125627,0.001382119,0.005511123,0.0007319807],"domain_scores_gemma":[0.992232,0.003824524,0.0003534103,0.0008349193,0.0005568161,0.002198279],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002080313,0.00003181002,0.3356702,0.000006485586,0.000004051647,0.0000351163,0.003051932,0.0009001834,0.00004725378,0.002302143,0.6290992,0.02883076],"study_design_scores_gemma":[0.0009140386,0.00005186362,0.04505789,0.00001514601,0.000006271859,0.000008311176,0.003772802,0.01013747,0.0001107428,0.004349428,0.9351778,0.0003982522],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7162332,0.0009261876,0.02678407,0.2236959,0.002597289,0.0007685527,0.0003565596,0.0003561246,0.02828214],"genre_scores_gemma":[0.9618907,0.00001294673,0.003063051,0.03171238,0.0001247756,0.00001149825,0.00003490214,0.00001719892,0.003132532],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5181986,"threshold_uncertainty_score":0.9999125,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4392109487","doi":"10.1007/s11192-024-04948-x","title":"Regional disparities in Web of Science and Scopus journal coverage","year":2024,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":145,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Scopus; Web of science; Scientometrics; Data science; Regional science; Bibliometrics; World Wide Web; Computer science; Library science; Information retrieval; Political science; Sociology; MEDLINE","retraction":null,"screen_n_in":null,"score":{"opus":0.5043046486588112,"gpt":0.5676367040364447,"spread":0.06333205537763353,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.09742559,0.0001572989,0.0003509867,0.3817813,0.0004155816,0.008781928,0.00296355,0.00009248902,0.0001595869],"category_scores_gemma":[0.1065742,0.0001093042,0.0001043108,0.7053051,0.003028503,0.002406982,0.001099789,0.0005469694,0.00006264549],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004530215,"about_ca_system_score_gemma":0.003227971,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005421605,"about_ca_topic_score_gemma":0.00001493158,"domain_scores_codex":[0.9708697,0.0001467403,0.0009705138,0.0009175018,0.02617891,0.0009166171],"domain_scores_gemma":[0.9857115,0.005847993,0.0002210204,0.0005897503,0.006801061,0.0008286884],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.00003005771,0.0002467267,0.4302867,0.00006868945,0.00001349334,0.0001775006,0.0006907793,0.0001280257,0.006268474,0.035109,0.03312479,0.4938558],"study_design_scores_gemma":[0.001158415,0.000528455,0.7285021,0.0002127951,0.00001154589,0.0004089527,0.000956007,0.07250423,0.00216641,0.05443011,0.1385454,0.0005755622],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9719712,0.015067,0.001293258,0.0013978,0.002333538,0.000170385,0.00003685352,0.00002085543,0.007709087],"genre_scores_gemma":[0.9956211,0.002382535,0.0006582935,0.0001009395,0.00007221149,0.000002705609,5.178103e-7,0.000008807585,0.001152888],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4932802,"threshold_uncertainty_score":0.9996847,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2226524777","doi":"10.1007/s11192-015-1825-x","title":"How to boost scientific production? A statistical analysis of research funding and other influencing factors","year":2016,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":137,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University","funders":"","keywords":"Quality (philosophy); Work (physics); Production (economics); Relation (database); Impact factor; Statistical analysis; Business; Marketing; Political science; Economics; Computer science; Statistics; Mathematics; Engineering; Microeconomics","retraction":null,"screen_n_in":null,"score":{"opus":0.7558819682205911,"gpt":0.6277056305859157,"spread":0.1281763376346754,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.09974855,0.0002088223,0.0006025062,0.4767476,0.0009271588,0.007461565,0.002574105,0.0001240696,0.0003443215],"category_scores_gemma":[0.3991574,0.0001175394,0.0001643002,0.7982325,0.00181283,0.001135526,0.001701463,0.000261525,0.00009876736],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004343511,"about_ca_system_score_gemma":0.0003677282,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001203945,"about_ca_topic_score_gemma":0.00009319081,"domain_scores_codex":[0.9693173,0.0005634705,0.000879407,0.002003178,0.02569866,0.001538015],"domain_scores_gemma":[0.9652255,0.01815615,0.0003588853,0.001794067,0.0127262,0.001739262],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00001667726,0.0001081464,0.9263775,0.000008778137,0.00008360518,0.000002911884,0.0007880807,0.00002786013,0.01989581,0.0045163,0.004624322,0.04355003],"study_design_scores_gemma":[0.0002493096,0.0002517567,0.9469564,0.0000267177,0.00005707133,0.00000189696,0.002226408,0.0006100925,0.009076896,0.002316291,0.0379461,0.0002810913],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9753835,0.0002480292,0.02024381,0.001788594,0.0009827956,0.0004143104,0.0003321317,0.00002790379,0.0005789458],"genre_scores_gemma":[0.9891614,0.0000179198,0.002309162,0.00002737837,0.00005451293,0.00001044032,0.000002150611,0.00001488962,0.008402177],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3214849,"threshold_uncertainty_score":0.9935688,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2340061204","doi":"10.1007/s11192-016-1953-y","title":"Characterization, description, and considerations for the use of funding acknowledgement data in Web of Science","year":2016,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":121,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"Social Sciences and Humanities Research Council of Canada; Fonds de Recherche du Québec-Société et Culture; Austrian Science Fund; Multiple Sclerosis Society; Impact Fund","keywords":"Acknowledgement; Citation; Computer science; Citation index; Library science; Science Citation Index; Bibliometrics; Index (typography); Scientometrics; Data science; Political science; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.8693361966360683,"gpt":0.5759535144275153,"spread":0.293382682208553,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.04605835,0.00009658866,0.0002424409,0.1029153,0.0003969166,0.001450268,0.002508396,0.00004534177,0.0001144173],"category_scores_gemma":[0.3614376,0.0000492015,0.00003900731,0.3255634,0.002002122,0.002766682,0.001766716,0.00006190809,0.000008367295],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001133572,"about_ca_system_score_gemma":0.0008582813,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001768532,"about_ca_topic_score_gemma":0.00006611066,"domain_scores_codex":[0.9913974,0.0001043187,0.0009364994,0.0007579539,0.006321588,0.0004822118],"domain_scores_gemma":[0.9525431,0.03012498,0.0007329033,0.002299066,0.01395954,0.00034036],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00001694022,0.0002262193,0.6495855,0.00001651977,0.00001035056,4.558748e-7,0.0003348699,0.00002166902,0.2130113,0.01133321,0.004210554,0.1212323],"study_design_scores_gemma":[0.001218793,0.0001942258,0.8849644,0.00005243655,0.00001577198,0.000003450195,0.0003141648,0.0443354,0.01579412,0.004403797,0.04849716,0.0002062664],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8921446,0.0003204375,0.1039187,0.001319129,0.000955629,0.0006445516,0.0006125288,0.00000690909,0.00007751803],"genre_scores_gemma":[0.9939979,0.0004799074,0.004915244,0.00003713551,0.00002435086,0.00001130966,0.000003794743,0.000005120336,0.0005252762],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3153792,"threshold_uncertainty_score":0.9995863,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2734032589","doi":"10.1007/s11192-017-2452-5","title":"Tracking the emergence of synthetic biology","year":2017,"lang":"en","type":"article","venue":"Scientometrics","topic":"Gene Regulatory Network Analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":109,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Office of Naval Research; Office of Defense Programs; National Key Research and Development Program of China; Biotechnology and Biological Sciences Research Council; Engineering and Physical Sciences Research Council; Natural Sciences and Engineering Research Council of Canada; Defense Advanced Research Projects Agency; Directorate for Biological Sciences; National Institutes of Health; Bundesministerium für Bildung und Forschung; Russian Foundation for Basic Research; National Natural Science Foundation of China; Deutsche Forschungsgemeinschaft; Wellcome Trust; Council of Scientific and Industrial Research, India; U.S. Department of Energy; U.S. Department of Defense; European Commission; Canadian Institutes of Health Research; National Science Foundation; American Heart Association; Ministry of Education, Culture, Sports, Science and Technology; Howard Hughes Medical Institute","keywords":"Synthetic biology; Systems biology; Domain (mathematical analysis); Computer science; Set (abstract data type); Benchmark (surveying); Multidisciplinary approach; Data science; Biology; Computational biology; Mathematics; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.03642355739386737,"gpt":0.3285188723528308,"spread":0.2920953149589635,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008364201,0.00007407833,0.0001035051,0.0002390463,0.00031562,0.0000405061,0.0008870257,0.00006813406,0.00003422485],"category_scores_gemma":[0.001027555,0.00005166066,0.0001087546,0.0008411143,0.0003761695,0.000002982556,0.0002522686,0.00004287954,0.000007200519],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000003892438,"about_ca_system_score_gemma":0.00003572448,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001263485,"about_ca_topic_score_gemma":0.00001402099,"domain_scores_codex":[0.9991741,0.00004145056,0.0001529346,0.000245156,0.0001859887,0.000200396],"domain_scores_gemma":[0.9985289,0.00001792181,0.0002095815,0.001023121,0.0001711593,0.00004929112],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00001269235,0.00007590035,0.336045,0.00001702264,0.0001380168,9.535326e-7,0.00008753241,0.0008272759,0.5895551,0.001276123,0.00244226,0.06952211],"study_design_scores_gemma":[0.0003839214,0.0002455995,0.3406213,0.00001559286,0.0001338084,0.00001051757,0.0001679736,0.001811032,0.6046208,0.0009085478,0.05069448,0.0003864427],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9926937,0.001580049,0.003008665,0.000201961,0.0004005225,0.00006538563,0.000005616296,0.00000352596,0.002040538],"genre_scores_gemma":[0.9990503,0.0002103107,0.000242573,0.000025501,0.000096018,0.000002748006,0.000004589408,0.000006416781,0.0003615691],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.06913567,"threshold_uncertainty_score":0.2427526,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2046437697","doi":"10.1007/s11192-007-0205-6","title":"“The Triple-Helix collaboration: Why do researchers collaborate with industry and the government? What are the factors that influence the perceived barriers?”","year":2007,"lang":"en","type":"article","venue":"Scientometrics","topic":"Innovation and Knowledge Management","field":"Business, Management and Accounting","cited_by":108,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Canadian Foundation for Healthcare Improvement; Université Laval","funders":"","keywords":"Government (linguistics); Affect (linguistics); Set (abstract data type); Public relations; Knowledge production; Business; Marketing; Knowledge management; Political science; Sociology; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.03674851862201558,"gpt":0.28574997305715,"spread":0.2490014544351344,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics","sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.009280633,0.0002236665,0.0001591846,0.0005326014,0.003132919,0.008273514,0.001060005,0.0001045318,0.00005066221],"category_scores_gemma":[0.00347139,0.00008841281,0.00003734774,0.02656568,0.001376599,0.001591296,0.0004749602,0.0005629719,0.00002861504],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002064821,"about_ca_system_score_gemma":0.0001030247,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000661365,"about_ca_topic_score_gemma":0.0009446733,"domain_scores_codex":[0.996395,0.00009996162,0.0003128854,0.0003558992,0.002285415,0.0005508529],"domain_scores_gemma":[0.9965748,0.0008151557,0.0004533456,0.0006358324,0.001471772,0.00004910309],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0009235928,0.00008678954,0.71323,0.0001207704,0.000250264,0.00001763708,0.007175857,0.0001248575,0.00005291255,0.2214139,0.04518018,0.01142327],"study_design_scores_gemma":[0.001448285,0.00001838621,0.2076956,0.00004862967,0.00005261408,7.289087e-7,0.327364,0.0005007259,0.00003773671,0.0004297645,0.4621894,0.00021417],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9240853,0.001234391,0.00008831487,0.03901543,0.001433777,0.00178316,0.000009172094,0.00006431073,0.0322861],"genre_scores_gemma":[0.9907241,0.0002008628,0.000009036633,0.005711405,0.0001892963,0.00003925399,0.000003337559,0.00001888322,0.003103827],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5055344,"threshold_uncertainty_score":0.9981649,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3125340956","doi":"10.1007/s11192-021-03868-4","title":"Tracking developments in artificial intelligence research: constructing and applying a new search strategy","year":2021,"lang":"en","type":"article","venue":"Scientometrics","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":106,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Biotechnology and Biological Sciences Research Council; Directorate for Biological Sciences; Social Sciences and Humanities Research Council of Canada; National Research Foundation; National Natural Science Foundation of China; Engineering and Physical Sciences Research Council; Taishan Scholar Foundation of Shandong Province; National Institutes of Health; National Science Foundation","keywords":"Profiling (computer programming); Computer science; Artificial intelligence; Multidisciplinary approach; Benchmark (surveying); Domain (mathematical analysis); Applications of artificial intelligence; Data science; Social science; Sociology","retraction":null,"screen_n_in":null,"score":{"opus":0.4453336730238928,"gpt":0.4467547731298974,"spread":0.001421100106004602,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.006335281,0.0001665677,0.0002230245,0.004123858,0.0004548451,0.001865591,0.001104457,0.0001061077,0.00002667592],"category_scores_gemma":[0.003692796,0.0001855835,0.00003399493,0.03644906,0.0002789177,0.001199151,0.001108782,0.0006624271,0.00009653334],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002922515,"about_ca_system_score_gemma":0.001405264,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002302765,"about_ca_topic_score_gemma":0.0002715779,"domain_scores_codex":[0.9952483,0.0002984579,0.0005888697,0.0009880553,0.001739848,0.001136536],"domain_scores_gemma":[0.9971406,0.001022142,0.00007365074,0.0005012864,0.0008709335,0.0003914236],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000002841876,0.00005312263,0.003937588,0.00001918728,0.00000507505,0.0001746604,0.00376975,0.0006089016,0.003750834,0.1571231,0.00001339691,0.8305415],"study_design_scores_gemma":[0.0002008951,0.0001812462,0.006279571,0.0003086363,0.000005074075,0.0002667342,0.04992527,0.1107273,0.6967097,0.1324782,0.001856644,0.001060734],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4258285,0.001148443,0.563968,0.0005838753,0.0007030542,0.0004989257,0.000001560963,0.000122796,0.007144833],"genre_scores_gemma":[0.9118063,0.00009697664,0.08779979,0.00005712172,0.00005854751,0.00001224214,0.000001210846,0.00001105023,0.0001567202],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8294808,"threshold_uncertainty_score":0.9991705,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2989484692","doi":"10.1007/s11192-019-03301-x","title":"Do articles in open access journals have more frequent altmetric activity than articles in subscription-based journals? An investigation of the research output of Finnish universities","year":2019,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":105,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Turun Yliopisto","keywords":"Altmetrics; Discipline; Scientometrics; Bibliometrics; Social media; Library science; Computer science; Sociology; Social science; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.838152280754456,"gpt":0.6455667087672443,"spread":0.1925855719872117,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.1082317,0.0002111957,0.0006495731,0.2690909,0.0003236766,0.007054362,0.0114311,0.0001994283,0.0004386413],"category_scores_gemma":[0.04319692,0.0001450322,0.000194003,0.6152785,0.0009399091,0.005927231,0.003343812,0.0009296827,0.00002469403],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009683496,"about_ca_system_score_gemma":0.001984661,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004475062,"about_ca_topic_score_gemma":0.0009897391,"domain_scores_codex":[0.9737656,0.002431782,0.001498649,0.001011109,0.02021208,0.001080846],"domain_scores_gemma":[0.9783849,0.008512352,0.001221544,0.001751825,0.009483315,0.0006460279],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0000714087,0.0004190861,0.9661016,0.00002050067,0.000007404975,0.000005940251,0.001483423,0.003067941,0.007809614,0.0001659838,0.0003029797,0.0205441],"study_design_scores_gemma":[0.001086377,0.0002137827,0.9348478,0.0001193549,0.000004348318,0.000001658204,0.008116978,0.01944543,0.02996376,0.005993111,0.00004977768,0.0001576139],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9955313,0.0008126932,0.0002076451,0.001303783,0.0004433122,0.0009106401,0.00006028094,0.000006311641,0.0007239901],"genre_scores_gemma":[0.9989226,0.0001418545,0.0003432622,0.00006284908,0.0000231433,0.000008554824,0.000001501654,0.00001389719,0.0004822791],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3461877,"threshold_uncertainty_score":0.9939764,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics","open_science"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics","open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2127540269","doi":"10.1007/s11192-012-0729-2","title":"Counting citations in the field of business and management: why use Google Scholar rather than the Web of Science","year":2012,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":104,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université Laval","funders":"","keywords":"Web of science; Field (mathematics); Bibliometrics; Personalization; Scientometrics; Computer science; Library science; Data science; Sociology; Psychology; Political science; World Wide Web; Mathematics; MEDLINE; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.4692757887462023,"gpt":0.5425769812812546,"spread":0.07330119253505235,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.09023855,0.0001209673,0.000221977,0.1034174,0.0006354402,0.0022956,0.003830256,0.00005890573,0.00005172826],"category_scores_gemma":[0.1270045,0.00006180266,0.00006694698,0.6459653,0.001444257,0.002604609,0.001065295,0.0002842697,0.00001637487],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005355871,"about_ca_system_score_gemma":0.0001888893,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001883627,"about_ca_topic_score_gemma":0.00003064539,"domain_scores_codex":[0.985712,0.0002854629,0.0006966224,0.0004613438,0.01210154,0.0007430388],"domain_scores_gemma":[0.9795646,0.01402459,0.0004668958,0.001185033,0.004540598,0.0002182795],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000007152057,0.0002335872,0.959743,0.00002060029,0.000006677969,0.000001134498,0.001344905,0.00003757061,0.001051219,0.01276253,0.003891472,0.02090021],"study_design_scores_gemma":[0.0002538662,0.00004062626,0.9702234,0.00001623995,0.000008614816,0.000003024162,0.002744947,0.001261606,0.0006892487,0.0007508964,0.02391418,0.00009333761],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9882324,0.001384422,0.001721464,0.001275167,0.000581145,0.0003995134,0.00001929888,0.000005763698,0.006380812],"genre_scores_gemma":[0.997394,0.0005067901,0.001424867,0.0003692921,0.0000268738,0.000009082928,5.605756e-7,0.000005493387,0.0002630868],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5425479,"threshold_uncertainty_score":0.9987401,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3037469314","doi":"10.1007/s11192-020-03590-7","title":"Coronavirus mapping in scientific publications: When science advances rapidly and collectively, is access to this knowledge open to society?","year":2020,"lang":"en","type":"article","venue":"Scientometrics","topic":"Academic Publishing and Open Access","field":"Decision Sciences","cited_by":101,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Fundação Oswaldo Cruz; Comunidad de Madrid","keywords":"Coronavirus disease 2019 (COVID-19); Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); 2019-20 coronavirus outbreak; Coronavirus; Sociology of scientific knowledge; Open science; Data science; Library science; Political science; Computer science; Sociology; Social science; Virology; Medicine; Physics; Infectious disease (medical specialty)","retraction":null,"screen_n_in":null,"score":{"opus":0.4500580548546065,"gpt":0.5235256239953888,"spread":0.07346756914078229,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.02642713,0.0002557122,0.0004927849,0.009032258,0.001854364,0.04255295,0.01814024,0.0001216524,0.0005081963],"category_scores_gemma":[0.05659683,0.0002149405,0.00007159627,0.1842111,0.0009560685,0.01790772,0.01169926,0.0004521765,0.0005798517],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004920926,"about_ca_system_score_gemma":0.001855322,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001703036,"about_ca_topic_score_gemma":0.00009844056,"domain_scores_codex":[0.9908495,0.0001685832,0.0009522985,0.002610686,0.004456911,0.0009620174],"domain_scores_gemma":[0.992964,0.001549316,0.0003704819,0.0009268783,0.002638642,0.001550655],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001924591,0.0000959028,0.1109818,0.00001231488,0.000005657132,9.028817e-7,0.03763992,0.00006286946,0.0009088907,0.0007961236,0.6457933,0.2036831],"study_design_scores_gemma":[0.0004837293,0.0000671197,0.08170581,0.00004071738,0.000003946322,0.000002109859,0.003381487,0.003622754,0.001188729,0.003885893,0.9052087,0.0004089904],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6750314,0.006266837,0.02588628,0.1514606,0.004680688,0.006189587,0.0004177217,0.0002856229,0.1297813],"genre_scores_gemma":[0.9314362,0.0001088195,0.02457364,0.02491207,0.000189928,0.0002484244,0.000007493717,0.00003537053,0.01848801],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2594154,"threshold_uncertainty_score":0.9994451,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2097269731","doi":"10.1023/b:scie.0000027677.79173.b8","title":"The classification of biomedical journals by research level","year":2004,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":89,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Science North","funders":"","keywords":"Computer science; Citation; Science Citation Index; Information retrieval; Basic research; Data science; Library science","retraction":null,"screen_n_in":null,"score":{"opus":0.9272708597111224,"gpt":0.7093757539334359,"spread":0.2178951057776866,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.2057137,0.000154572,0.0003276869,0.1777989,0.001536793,0.004603872,0.007405369,0.0002101508,0.000207557],"category_scores_gemma":[0.3083346,0.00008479669,0.0002015976,0.7157363,0.002311401,0.0006741668,0.001195092,0.000835082,0.0006112434],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005187356,"about_ca_system_score_gemma":0.001055703,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001161045,"about_ca_topic_score_gemma":0.00001303946,"domain_scores_codex":[0.9392418,0.0007403781,0.001669267,0.0009770918,0.05584848,0.00152296],"domain_scores_gemma":[0.9560933,0.0199717,0.0006784089,0.001840533,0.02019403,0.001222094],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003475098,0.0008469789,0.02419321,0.000007270738,0.00002577331,0.000007267863,0.0003623342,0.00003383488,0.05391213,0.01981578,0.2646721,0.6360886],"study_design_scores_gemma":[0.001374111,0.0006347309,0.4007756,0.00002622899,0.00000609612,0.00001834893,0.002229139,0.001320611,0.01171471,0.08020875,0.5013958,0.0002959356],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8835268,0.01285592,0.06419225,0.02068269,0.003084413,0.0008100205,0.0003111181,0.00004413114,0.01449268],"genre_scores_gemma":[0.9937416,0.001260756,0.001223136,0.00004754034,0.0001021749,0.00001281361,0.000005308968,0.00001266674,0.003593993],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6357926,"threshold_uncertainty_score":0.9997631,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics","metaresearch"],"domain":"evaluation","study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","bibliometrics"],"domain":"evaluation","study_design":"observational","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2114111800","doi":"10.1007/s11192-009-0035-9","title":"Why it has become more difficult to predict Nobel Prize winners: a bibliometric analysis of nominees and winners of the chemistry and physics prizes (1901–2007)","year":2009,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":86,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Citation; Fragmentation (computing); Attribution; Hierarchy; Field (mathematics); Citation analysis; Data science; Sociology; Library science; Political science; Psychology; Computer science; Mathematics; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.2500444637851816,"gpt":0.4672281133466309,"spread":0.2171836495614493,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","bibliometrics","scholarly_communication"],"consensus_categories":["bibliometrics"],"category_scores_codex":[0.0146142,0.0004043294,0.001162774,0.4673313,0.0004008978,0.002126982,0.003657055,0.0002366101,0.0001132819],"category_scores_gemma":[0.07676464,0.0002574035,0.000428227,0.961329,0.001195066,0.000767266,0.00160372,0.0003697445,0.000004913722],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001583258,"about_ca_system_score_gemma":0.0002912022,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001311369,"about_ca_topic_score_gemma":0.00001084774,"domain_scores_codex":[0.9818282,0.0001557809,0.00150288,0.001430598,0.01408601,0.0009965448],"domain_scores_gemma":[0.9856835,0.004478426,0.0009552816,0.001657844,0.00623274,0.0009921911],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0000841349,0.0009345891,0.7577674,0.00009212612,0.0004545572,0.000006762201,0.001668394,0.001730066,0.02493869,0.0001085802,0.02842136,0.1837933],"study_design_scores_gemma":[0.0007807306,0.0003113287,0.969685,0.00003056643,0.0002480829,0.000004843462,0.0006802208,0.01590137,0.005036637,0.000227331,0.006741942,0.0003519508],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9890583,0.001760042,0.003336463,0.003503031,0.0003197615,0.0005460927,0.0003678027,0.00002557026,0.001082942],"genre_scores_gemma":[0.995342,0.0004594644,0.002828321,0.0006746614,0.00007125836,0.000008222213,0.00001110958,0.00001580262,0.0005891867],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4939977,"threshold_uncertainty_score":0.9999878,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2044326206","doi":"10.1007/s11192-008-0220-2","title":"A new look at evidence of scholarly citation in citation indexes and from web sources","year":2007,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":85,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Citation; Web of science; Library science; Citation analysis; Webometrics; Citation impact; Bibliometrics; Scholarly communication; Computer science; Sample (material); Impact factor; World Wide Web; Political science; MEDLINE; Publishing","retraction":null,"screen_n_in":null,"score":{"opus":0.5726813817650799,"gpt":0.5567807477511177,"spread":0.01590063401396213,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.06158305,0.0001795445,0.0003808583,0.2483297,0.0002090949,0.003611983,0.001883733,0.0002070862,0.0002956545],"category_scores_gemma":[0.1921769,0.0001440193,0.00009545132,0.5391781,0.0002872671,0.003999108,0.0009399208,0.0003335925,0.0001327346],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003777984,"about_ca_system_score_gemma":0.0004307377,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008616144,"about_ca_topic_score_gemma":0.0006853588,"domain_scores_codex":[0.9797834,0.0002406213,0.001294367,0.00108498,0.01684294,0.0007536688],"domain_scores_gemma":[0.9724427,0.02094738,0.0006786906,0.0007354334,0.004372299,0.0008234766],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00005946088,0.00004929365,0.8260859,0.000006706121,0.000004549861,0.000005882124,0.0008868973,0.00003081033,0.01218201,0.0002786964,0.001110754,0.1592991],"study_design_scores_gemma":[0.0007069465,0.0001980155,0.9824387,0.0000535971,0.000005681039,0.000001966001,0.0008031926,0.001849006,0.005668396,0.0061341,0.001961292,0.0001791179],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9728765,0.00912236,0.01575318,0.0004166511,0.0005281832,0.0002774931,0.00002254843,0.00001815375,0.0009849267],"genre_scores_gemma":[0.9885912,0.0006787465,0.009100274,0.0000812181,0.00006168662,0.000002291321,0.000003677136,0.000009933621,0.001471007],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2908484,"threshold_uncertainty_score":0.9974223,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics","scholarly_communication"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2004098455","doi":"10.1007/s11192-013-1069-6","title":"Effect of collaboration network structure on knowledge creation and technological performance: the case of biotechnology in Canada","year":2013,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":84,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Concordia University","funders":"","keywords":"Network structure; Productivity; Quality (philosophy); Production (economics); Knowledge management; Field (mathematics); Network analysis; Biotechnology; Business; Computer science; Data science; Engineering; Biology; Economics","retraction":null,"screen_n_in":null,"score":{"opus":0.1136474693055103,"gpt":0.4694544594801149,"spread":0.3558069901746047,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":["bibliometrics"],"category_scores_codex":[0.01540162,0.000130408,0.0003457652,0.05063199,0.0001753469,0.000285279,0.001128622,0.0001924777,0.00009330339],"category_scores_gemma":[0.03425495,0.00006740615,0.00003072491,0.4314651,0.0005514974,0.000243503,0.0004486114,0.0003088981,0.000007501409],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003298815,"about_ca_system_score_gemma":0.0005209459,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.05370288,"about_ca_topic_score_gemma":0.07347152,"domain_scores_codex":[0.9953133,0.000356166,0.0006361364,0.0004599351,0.002798569,0.0004358876],"domain_scores_gemma":[0.9895739,0.00773305,0.0003446456,0.0006007021,0.001624832,0.0001228694],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.00002286156,0.00003863651,0.2841277,0.00002394552,0.000005109296,0.00001150405,0.00006510794,0.0009922805,0.0009183518,0.001664325,0.001864912,0.7102652],"study_design_scores_gemma":[0.001534187,0.002950004,0.740337,0.0000499169,0.00001660216,0.0001227433,0.001751652,0.150799,0.09668067,0.00259415,0.002805426,0.0003586424],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9970747,0.0008125238,0.00008144907,0.0005073469,0.0003035494,0.0005588941,0.00002506945,0.000007662398,0.0006287817],"genre_scores_gemma":[0.9995803,0.0001863151,0.0001270948,0.00001459261,0.00001331321,0.00001261571,0.000001521015,0.000004052707,0.00006016437],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7099066,"threshold_uncertainty_score":0.9738799,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":true,"about_ca_topic":true,"confidence":"low"},{"model":"gpt","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":true,"about_ca_topic":true,"confidence":"low"}],"label_agreement":"agree"},{"id":"W2064596000","doi":"10.1007/s11192-013-1118-1","title":"The effect of characteristics of title on citation rates of articles","year":2013,"lang":"en","type":"article","venue":"Scientometrics","topic":"Academic Writing and Publishing","field":"Arts and Humanities","cited_by":79,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University Health Centre","funders":"","keywords":"Citation; Acronym; Information retrieval; Computer science; Sentence; Scopus; Citation analysis; Statistics; Library science; Mathematics; Natural language processing; MEDLINE; Linguistics; Biology","retraction":null,"screen_n_in":null,"score":{"opus":0.03765010945226892,"gpt":0.277116055331689,"spread":0.2394659458794201,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005420418,0.00002143363,0.00005285872,0.0002309674,0.00004021628,0.00004258592,0.00007991374,0.00001217201,0.0002368252],"category_scores_gemma":[0.00132267,0.00001251435,0.00001729917,0.0003192493,0.0001024981,0.00005964711,0.00001042864,0.00003654135,0.00003477114],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000003646285,"about_ca_system_score_gemma":0.000005215156,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002788616,"about_ca_topic_score_gemma":5.997167e-7,"domain_scores_codex":[0.9996192,0.00001897528,0.0001061224,0.00003046672,0.0001754234,0.00004978087],"domain_scores_gemma":[0.9990969,0.0006151291,0.00009574262,0.00005250202,0.0001288462,0.00001088288],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002420649,0.00007092529,0.03634286,0.0003542772,0.00003614419,1.549518e-7,0.0130109,0.000007677155,0.003374937,0.5112465,0.1234184,0.312113],"study_design_scores_gemma":[0.002166146,0.004159965,0.3329673,0.000838666,0.0001761235,0.00000146568,0.01059985,0.008710251,0.2329959,0.04281281,0.3637451,0.0008264786],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9626379,0.00008224913,0.000009050928,0.00005979588,0.0002213858,0.00004139677,0.000009008309,0.000004099501,0.03693513],"genre_scores_gemma":[0.9982958,0.000006745685,0.00001092834,0.000007659738,0.00004269054,0.000001010918,0.000001565003,0.00000189681,0.00163168],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4684337,"threshold_uncertainty_score":0.2593069,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2064363624","doi":"10.1023/b:scie.0000034389.14825.bc","title":"Links to commercial websites as a source of business information","year":2004,"lang":"en","type":"article","venue":"Scientometrics","topic":"Web visibility and informetrics","field":"Computer Science","cited_by":78,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Webometrics; Hyperlink; Business; Revenue; Scopus; China; Link analysis; Profit (economics); Competitive intelligence; Business information; Marketing; World Wide Web; Computer science; Web page; Accounting; Economics","retraction":null,"screen_n_in":null,"score":{"opus":0.02091219160032057,"gpt":0.2720124649186993,"spread":0.2511002733183787,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.001308994,0.0001162986,0.0001786999,0.006098521,0.000142344,0.000366619,0.001366579,0.0001348917,0.00001405484],"category_scores_gemma":[0.004229892,0.0001086039,0.00006137189,0.05662442,0.00007244525,0.002518973,0.0005601599,0.0001838577,0.0007766339],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001497999,"about_ca_system_score_gemma":0.0003412076,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000252955,"about_ca_topic_score_gemma":0.000009367783,"domain_scores_codex":[0.9978217,0.00001535905,0.0004376533,0.0002036583,0.001189227,0.0003324165],"domain_scores_gemma":[0.9978164,0.0001360206,0.0001739579,0.0005904721,0.001066791,0.0002162927],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.00003181941,0.0004927347,0.02025308,0.0001909801,0.00001952481,0.000002511484,0.008781149,0.03431188,0.0003087283,0.1092635,0.003084969,0.8232591],"study_design_scores_gemma":[0.002745883,0.001231109,0.6695571,0.0001206009,0.00001900597,0.00002355446,0.0003114386,0.01097263,0.02167599,0.006867254,0.2854566,0.001018863],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.272632,0.00002821012,0.7203693,0.0009758401,0.0005581227,0.0001893878,0.000003564056,0.00008975972,0.005153792],"genre_scores_gemma":[0.9470721,0.00001129737,0.05067949,0.00207991,0.00004066722,0.000004267336,0.000005641753,0.000003893517,0.0001026777],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8222402,"threshold_uncertainty_score":0.9982316,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2954205856","doi":"10.1007/s11192-019-03157-1","title":"How collaboration type, publication place, funding and author’s role affect citations received by publications from Africa: A bibliometric study of LIS research from 1996 to 2015","year":2019,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":77,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Citation; Incentive; Library science; Political science; Citation impact; Bibliometrics; Public relations; Economics; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.5915490014438257,"gpt":0.5835091969205798,"spread":0.00803980452324593,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.04756975,0.0003101508,0.0006698212,0.7025377,0.0007846878,0.01561464,0.003798359,0.0002941785,0.0004759513],"category_scores_gemma":[0.263194,0.0002631141,0.00008684335,0.9555584,0.0002354218,0.003501321,0.001679258,0.0006130276,0.0006081595],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000587261,"about_ca_system_score_gemma":0.0006790253,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001962771,"about_ca_topic_score_gemma":0.0003318974,"domain_scores_codex":[0.9717721,0.00144467,0.001237409,0.002256005,0.02205972,0.001230053],"domain_scores_gemma":[0.9325635,0.02581298,0.0009091271,0.002602417,0.03643404,0.001677871],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00007055468,0.001175067,0.6169953,0.000006739959,0.00006161159,7.375871e-7,0.003891633,0.00005384746,0.01455891,0.0006094109,0.3328057,0.02977046],"study_design_scores_gemma":[0.001399413,0.001552733,0.8760884,0.00001469993,0.00002793457,5.181445e-7,0.02288548,0.004670325,0.0007849813,0.002650301,0.08950081,0.0004244008],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9819898,0.002947714,0.001041317,0.00606135,0.001000316,0.002474082,0.001133015,0.00007582845,0.003276548],"genre_scores_gemma":[0.9829507,0.0001398978,0.001991683,0.00004537167,0.00009510496,0.0001345388,0.0002207615,0.00003292381,0.01438902],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.259093,"threshold_uncertainty_score":0.9999821,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2998363348","doi":"10.1007/s11192-019-03339-x","title":"The role of Web of Science publications in China’s tenure system","year":2020,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":72,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; Université de Montréal","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"China; Web of science; Political science; World Wide Web; Library science; Computer science; Sociology; MEDLINE; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.4024573698676067,"gpt":0.5294233401406908,"spread":0.1269659702730841,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.05203402,0.0001378227,0.0003878806,0.1496863,0.000492373,0.001609019,0.007038584,0.00008051716,0.00005092049],"category_scores_gemma":[0.1776651,0.00007896608,0.0001501609,0.8136093,0.002133068,0.0008680943,0.001515258,0.0002997782,0.00007004895],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00024868,"about_ca_system_score_gemma":0.00159803,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001027676,"about_ca_topic_score_gemma":0.00001544651,"domain_scores_codex":[0.9761562,0.0001933559,0.001374925,0.0008745887,0.02055713,0.0008437501],"domain_scores_gemma":[0.9851052,0.003870218,0.0007967435,0.001334469,0.008137533,0.0007558106],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00002422099,0.0002334618,0.649128,0.00002910606,0.000008872534,0.000002185553,0.001111199,0.0002222919,0.0191752,0.1592269,0.001981582,0.1688571],"study_design_scores_gemma":[0.0005283738,0.0002445894,0.8062006,0.0000179467,0.000005527257,0.00000468912,0.005601948,0.1132713,0.01268389,0.001959797,0.05928292,0.0001984597],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9127955,0.003052676,0.0008228233,0.003567369,0.0007272805,0.0005600955,0.0001011387,0.00004094405,0.07833216],"genre_scores_gemma":[0.999057,0.00009355951,0.0005411955,0.00002161713,0.00002847184,0.000009041754,8.665044e-7,0.000006950622,0.0002412964],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6639231,"threshold_uncertainty_score":0.9994274,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2906809950","doi":"10.1007/s11192-018-2987-0","title":"Publishing in English or another language: An inclusive study of scholar’s language publication preferences in the natural, social and interdisciplinary sciences","year":2019,"lang":"en","type":"article","venue":"Scientometrics","topic":"Discourse Analysis in Language Studies","field":"Arts and Humanities","cited_by":70,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Publishing; Publication; Reputation; English language; Contrast (vision); Sociology; Linguistics; First language; Inclusion (mineral); Social science; Psychology; Computer science; Political science; Mathematics education","retraction":null,"screen_n_in":null,"score":{"opus":0.06136419564155948,"gpt":0.3791982291499634,"spread":0.3178340335084039,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.003775318,0.0001142867,0.0002115394,0.002747663,0.0004351794,0.001961812,0.0008119029,0.00003179813,0.0002928041],"category_scores_gemma":[0.002230651,0.00006612758,0.00002647389,0.004022789,0.0004476269,0.003387478,0.0005769239,0.0002948603,0.000002554951],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000466233,"about_ca_system_score_gemma":0.00005362518,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009359983,"about_ca_topic_score_gemma":0.02578174,"domain_scores_codex":[0.9980698,0.0002553979,0.0002614213,0.0003470014,0.0008352449,0.0002310975],"domain_scores_gemma":[0.9989723,0.0003349571,0.0001827003,0.0002135448,0.0002778932,0.00001861217],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.00001264097,0.0004000026,0.03477782,0.00002086851,0.00001477191,0.000003854687,0.9524078,0.000003645379,0.00001840428,0.004336594,0.0003778939,0.00762569],"study_design_scores_gemma":[0.0003578679,0.0002405004,0.04689897,0.00002115765,0.00001430555,3.55576e-7,0.951802,0.0001114424,0.000007673232,0.0001228749,0.0003171778,0.0001056987],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9628813,0.000783492,3.065516e-7,0.0004208078,0.0002974079,0.0003213685,0.00001407922,0.00001945892,0.03526176],"genre_scores_gemma":[0.9990767,0.000005662615,0.00002989128,0.0001378416,0.0001664964,0.00001852205,0.000008809657,0.000006082614,0.0005499512],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03619542,"threshold_uncertainty_score":0.9990742,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1998258485","doi":"10.1023/b:scie.0000045122.93018.2a","title":"Exploring website features for business information","year":2004,"lang":"en","type":"article","venue":"Scientometrics","topic":"Web visibility and informetrics","field":"Computer Science","cited_by":64,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Webometrics; Ranking (information retrieval); Data collection; Business; China; Marketing; Computer science; Information retrieval; World Wide Web; Statistics; Political science; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.122088333432935,"gpt":0.283840460132473,"spread":0.161752126699538,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.00122144,0.0001185345,0.0001250207,0.004724024,0.0002666374,0.0008469886,0.001031664,0.00005474643,0.000001924129],"category_scores_gemma":[0.002808766,0.0001070775,0.00007024513,0.0349849,0.00004297145,0.009716222,0.0002507407,0.00009948926,0.0001123663],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000209114,"about_ca_system_score_gemma":0.0001981809,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002878189,"about_ca_topic_score_gemma":0.000003661816,"domain_scores_codex":[0.9982677,0.00000677829,0.0002854616,0.0002329123,0.0008051094,0.0004021049],"domain_scores_gemma":[0.9982566,0.0001370986,0.0001198105,0.0005024792,0.0008417224,0.0001423128],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.00001761821,0.0001662923,0.001312844,0.0002219205,0.00001596967,0.000001785316,0.003234059,0.01164525,0.0001267869,0.4309795,0.002161955,0.5501159],"study_design_scores_gemma":[0.00529793,0.00068072,0.4864403,0.0001084054,0.00002441163,0.00003305542,0.0004914824,0.01392745,0.02318722,0.02732241,0.440958,0.001528614],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05103415,0.0000795457,0.9433129,0.0007295134,0.001854828,0.0003039762,0.00001044352,0.0002122993,0.002462287],"genre_scores_gemma":[0.7692963,0.0001248348,0.229259,0.0009416692,0.0001399807,0.00007281841,0.00002431399,0.000008235394,0.0001329186],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7182621,"threshold_uncertainty_score":0.985527,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2585136728","doi":"10.1007/s11192-017-2252-y","title":"Development trend forecasting for coherent light generator technology based on patent citation network analysis","year":2017,"lang":"en","type":"article","venue":"Scientometrics","topic":"Innovation Diffusion and Forecasting","field":"Decision Sciences","cited_by":63,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"National University of Defense Technology; National Natural Science Foundation of China","keywords":"Trademark; Computer science; Technology forecasting; Autoregressive integrated moving average; Citation; Network analysis; Development (topology); Patent office; Data science; Operations research; Artificial intelligence; Time series; Engineering; Machine learning; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.3621151555736334,"gpt":0.4023682443934626,"spread":0.04025308881982914,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["bibliometrics"],"category_scores_codex":[0.007963471,0.000193415,0.0003642885,0.01187971,0.002149055,0.0012758,0.001204031,0.0001410237,0.0001745762],"category_scores_gemma":[0.01718171,0.000144538,0.0001891462,0.02979638,0.00009796325,0.0002114149,0.0002182095,0.0001276762,0.00004965537],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001689057,"about_ca_system_score_gemma":0.0001401435,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001682797,"about_ca_topic_score_gemma":0.0000668455,"domain_scores_codex":[0.9954281,0.00004350357,0.0009398152,0.0007510743,0.002284245,0.0005532315],"domain_scores_gemma":[0.9957308,0.0006249514,0.001169048,0.0009380627,0.001384498,0.0001526274],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00004947263,0.0001691228,0.1350375,0.000006946086,0.00007968087,0.000004484711,0.0002864959,0.01738448,0.0004211765,0.004247686,0.007818945,0.834494],"study_design_scores_gemma":[0.00100389,0.0001328509,0.0514077,0.00002584885,0.00006351523,7.722583e-7,0.0001812672,0.8349457,0.002533949,0.001543164,0.1078258,0.0003355742],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6810729,0.00004642702,0.3087996,0.00170674,0.002308436,0.0005585608,0.00003014335,0.00009776165,0.005379504],"genre_scores_gemma":[0.922232,6.749288e-7,0.07598914,0.0003219713,0.0001394045,0.0000668422,0.00003387668,0.00001250384,0.001203539],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8341584,"threshold_uncertainty_score":0.999761,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2075464897","doi":"10.1007/s11192-012-0859-6","title":"Characterizing a scientific elite (B): publication and citation patterns of the most highly cited scientists in environmental science and ecology","year":2012,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":62,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Citation; Publishing; Publication; Elite; Productivity; Citation analysis; Library science; Ecology; Political science; Sociology; Biology; Computer science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.2644749148570341,"gpt":0.4612345109392629,"spread":0.1967595960822288,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.06896877,0.0001695564,0.0002728926,0.1482552,0.0009451085,0.004050986,0.002212479,0.0001046378,0.0000641103],"category_scores_gemma":[0.06760544,0.0001168383,0.00004837319,0.4536093,0.002651048,0.003223875,0.002137082,0.0002516458,0.00002869701],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000433768,"about_ca_system_score_gemma":0.0003471444,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005557412,"about_ca_topic_score_gemma":0.00002216032,"domain_scores_codex":[0.9837248,0.0002661919,0.0008648708,0.001073345,0.01297681,0.001093919],"domain_scores_gemma":[0.9935562,0.001803566,0.0005942655,0.0009216093,0.00244432,0.0006800312],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00000458263,0.0001471349,0.8878684,0.000007167814,0.00000163685,3.326316e-7,0.0008201284,0.000001193899,0.04148963,0.0006697598,0.0001076892,0.06888234],"study_design_scores_gemma":[0.0003440464,0.00003760991,0.9871478,0.00001071367,0.000003806649,0.00001018024,0.0003664066,0.002541281,0.004473042,0.0001965574,0.004734856,0.0001336748],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9956776,0.0006994798,0.0001346355,0.0007187071,0.001670091,0.0004526699,0.0001011623,0.00001010916,0.0005355787],"genre_scores_gemma":[0.9988313,0.0001383028,0.000342686,0.0001629851,0.00003054688,0.00001347827,0.000008483442,0.000008122991,0.0004640881],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3053541,"threshold_uncertainty_score":0.9969829,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics","metaresearch"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics","metaresearch"],"domain":"evaluation","study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W1987570886","doi":"10.1007/s11192-010-0162-3","title":"The discipline dependence of citation statistics","year":2010,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":60,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"National Academy of Sciences; National Science Council","keywords":"Citation; Field (mathematics); Citation analysis; Value (mathematics); Index (typography); Citation index; Scientometrics; Computer science; Data science; Statistics; Library science; Mathematics; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.5142649586072647,"gpt":0.6001517069820625,"spread":0.08588674837479782,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.05536565,0.0001344871,0.0002331443,0.07803998,0.0007930655,0.002939895,0.004670041,0.0001094759,0.0003254991],"category_scores_gemma":[0.2715822,0.00007410849,0.0001098252,0.4760623,0.0008723497,0.0005462879,0.000974177,0.0004958269,0.0002959979],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004853111,"about_ca_system_score_gemma":0.000293336,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003567485,"about_ca_topic_score_gemma":0.000122418,"domain_scores_codex":[0.9777182,0.0001507015,0.001019046,0.0006342519,0.0197705,0.00070729],"domain_scores_gemma":[0.968447,0.01899838,0.0006386715,0.001449353,0.009935386,0.0005312515],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.00002051429,0.0001764024,0.1346284,0.000005681876,0.00001081701,0.000006222821,0.0002141676,0.00006139318,0.01155115,0.0645676,0.02087903,0.7678787],"study_design_scores_gemma":[0.0005558886,0.0002556962,0.7875187,0.000003428841,0.00001090559,0.00001692974,0.0006101981,0.03280526,0.007214009,0.09064214,0.08005556,0.0003112186],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8341965,0.0005062174,0.1478002,0.001128859,0.005441917,0.0003842601,0.0003074894,0.00003372031,0.01020085],"genre_scores_gemma":[0.9823861,0.0001304072,0.0140798,0.00002529612,0.00007393467,0.000005067968,0.000005821534,0.000009736238,0.003283833],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7675675,"threshold_uncertainty_score":0.9980952,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2027998681","doi":"10.1007/s11192-009-0424-0","title":"Citation structure of an emerging research area on the verge of application","year":2008,"lang":"en","type":"article","venue":"Scientometrics","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":57,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Thomson Reuters (Canada)","funders":"","keywords":"Pace; Commercialization; Demise; Citation; Bibliometrics; Nanotechnology; Data science; Engineering ethics; Computer science; Business; Engineering; Political science; Library science; Materials science; Marketing; Physics","retraction":null,"screen_n_in":null,"score":{"opus":0.1288690603491131,"gpt":0.3997077030823721,"spread":0.270838642733259,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005074818,0.00008279118,0.0001400937,0.001601755,0.0003785123,0.00004810024,0.0009330355,0.0000492422,0.0003642982],"category_scores_gemma":[0.00386688,0.00005622538,0.00002551192,0.01001083,0.0005976851,0.0002229527,0.0001267893,0.0001466383,0.00003408414],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000537939,"about_ca_system_score_gemma":0.00008777156,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001209281,"about_ca_topic_score_gemma":0.000006550084,"domain_scores_codex":[0.996573,0.0003584166,0.0003093858,0.0003415595,0.002128602,0.0002890663],"domain_scores_gemma":[0.9978175,0.0005661933,0.0002661452,0.0006064281,0.0006740112,0.00006978986],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00001437424,0.00004473226,0.002252221,0.00001762551,6.617386e-7,3.96544e-7,0.001037607,0.005109129,0.9871975,0.003608203,0.0001286177,0.0005889197],"study_design_scores_gemma":[0.000121867,0.0002249442,0.05399594,0.00001656301,0.000002502573,0.000004190019,0.0002190052,0.01602959,0.9271874,0.001785341,0.0003216179,0.00009097543],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9967789,0.00001886253,0.001793283,0.0001334936,0.0003419934,0.0002277123,0.00002861092,0.00002450438,0.0006526455],"genre_scores_gemma":[0.9954367,0.000006252606,0.004419215,0.00002461048,0.00003793994,0.000005715405,0.000005948312,0.000007532196,0.00005603437],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.06001005,"threshold_uncertainty_score":0.4809872,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2080497488","doi":"10.1007/s11192-014-1457-6","title":"Exploring the interdisciplinary evolution of a discipline: the case of Biochemistry and Molecular Biology","year":2014,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":56,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; Université de Montréal","funders":"","keywords":"Discipline; Variety (cybernetics); Interdisciplinarity; Identification (biology); Epistemology; Sociology; Data science; Engineering ethics; Biology; Social science; Computer science; Ecology; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.4283907410291913,"gpt":0.5475682690263628,"spread":0.1191775279971715,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.03298892,0.0001368364,0.0002925837,0.03702204,0.0004097176,0.0003903373,0.002299743,0.00006754036,0.00002067387],"category_scores_gemma":[0.05051106,0.00006417531,0.0001489777,0.2764794,0.00136782,0.0002946162,0.003456335,0.0002438446,0.000007794844],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008286251,"about_ca_system_score_gemma":0.0001066888,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009147661,"about_ca_topic_score_gemma":0.000009920851,"domain_scores_codex":[0.9947879,0.0003404771,0.0007558698,0.0005908344,0.003080929,0.0004439789],"domain_scores_gemma":[0.9885278,0.007329972,0.0004551222,0.001226484,0.002243359,0.0002172039],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.0000600623,0.0003919994,0.09868237,0.00008858828,0.00006932305,0.0000459966,0.002317844,0.0003177717,0.09614258,0.05640392,0.001707912,0.7437716],"study_design_scores_gemma":[0.00269458,0.002241294,0.2997963,0.0001190729,0.0001507949,0.001131705,0.03159652,0.1595343,0.2225431,0.2734025,0.005566349,0.001223489],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9571957,0.001320991,0.0384356,0.001099249,0.0004922299,0.0001725619,0.00003143221,0.000007765928,0.001244464],"genre_scores_gemma":[0.9991814,0.00005474352,0.000580131,0.00001811711,0.00004251545,0.00001267962,0.000001082825,0.00000667514,0.0001026829],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7425482,"threshold_uncertainty_score":0.9957414,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2035792097","doi":"10.1007/s11192-009-0008-z","title":"Discovery of factors influencing patent value based on machine learning in patents in the field of nanotechnology","year":2009,"lang":"en","type":"article","venue":"Scientometrics","topic":"Intellectual Property and Patents","field":"Business, Management and Accounting","cited_by":56,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Selection (genetic algorithm); Field (mathematics); Patent analysis; Value (mathematics); Data science; Computer science; Feature selection; Set (abstract data type); Artificial intelligence; Machine learning; Knowledge management; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.1052886624033666,"gpt":0.2640623858484547,"spread":0.1587737234450881,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001071215,0.0001125209,0.0001883372,0.003832084,0.0000548514,0.00005994315,0.0004218513,0.00008734403,0.00001827388],"category_scores_gemma":[0.002925549,0.00007092739,0.00005821371,0.008873057,0.00003982148,0.0003373132,0.00006621038,0.0002947029,0.000007120503],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003366701,"about_ca_system_score_gemma":0.0000169484,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001457638,"about_ca_topic_score_gemma":0.00004496846,"domain_scores_codex":[0.9986776,0.00003309224,0.0003336042,0.0001880655,0.0005225496,0.0002450839],"domain_scores_gemma":[0.999263,0.0002824431,0.0001930813,0.0001777993,0.00007914849,0.000004548577],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001444204,0.0005322554,0.9641641,0.0001161345,0.000004948662,0.00001093226,0.0006069535,0.01816598,0.003193869,0.001425616,0.00006249646,0.01157233],"study_design_scores_gemma":[0.001973884,0.0007359416,0.7907759,0.0004892279,0.00002584669,4.820376e-7,0.001064068,0.1794335,0.02273688,0.001488524,0.0008541521,0.0004215843],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9974535,0.00004668967,0.0002359328,0.0003198988,0.0001499156,0.0001630698,0.000001357925,0.00001476288,0.001614951],"genre_scores_gemma":[0.9985825,0.000008567377,0.00002012057,0.001329632,0.00001487578,0.00000158567,0.000007906641,0.000004985706,0.0000297777],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1733882,"threshold_uncertainty_score":0.4263209,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4221058759","doi":"10.1007/s11192-022-04331-8","title":"Music information visualization and classical composers discovery: an application of network graphs, multidimensional scaling, and support vector machines","year":2022,"lang":"en","type":"article","venue":"Scientometrics","topic":"Music and Audio Processing","field":"Computer Science","cited_by":54,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Multidimensional scaling; Computer science; Visualization; Pairwise comparison; Similarity (geometry); Music theory; Musical; Musicology; Realm; Musical composition; Classical music; Artificial intelligence; Art; Machine learning; Visual arts; History","retraction":null,"screen_n_in":null,"score":{"opus":0.01655784240464562,"gpt":0.2749442264079385,"spread":0.2583863840032928,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008894632,0.00008437691,0.0001180026,0.0008066913,0.000410253,0.0002012074,0.0002301502,0.0000257405,0.000005024393],"category_scores_gemma":[0.00006086503,0.00007971075,0.00001910931,0.005131799,0.0001002766,0.001663627,0.0003869005,0.00008082746,5.833529e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003087498,"about_ca_system_score_gemma":0.00005974278,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001613358,"about_ca_topic_score_gemma":0.00000222879,"domain_scores_codex":[0.9986306,0.00005722068,0.0002409623,0.0002414507,0.0006630687,0.0001666599],"domain_scores_gemma":[0.9993433,0.00006809102,0.0002186593,0.0001685847,0.0001084522,0.00009291749],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00004341108,0.0003398459,0.0984228,0.0001812843,0.00002149372,0.000001688822,0.0065832,0.01119957,0.002654388,0.4435586,0.002374953,0.4346188],"study_design_scores_gemma":[0.0003648667,0.0001424795,0.07519391,0.000005842621,0.000007877733,0.00001132734,0.00007431513,0.9173726,0.0001514842,0.001308707,0.005219218,0.0001473056],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4528466,0.00008598367,0.5464534,0.0001086028,0.0002581769,0.0001393806,0.00001069233,0.00004027624,0.00005693545],"genre_scores_gemma":[0.9876592,0.000007756206,0.01178516,0.0004211369,0.00002475194,0.00001004503,0.00007475657,0.000003752534,0.00001349543],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9061731,"threshold_uncertainty_score":0.325051,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1965490711","doi":"10.1007/s11192-006-0100-6","title":"Why are hyperlinks to business Websites created? A content analysis","year":2006,"lang":"en","type":"article","venue":"Scientometrics","topic":"Web visibility and informetrics","field":"Computer Science","cited_by":53,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Hyperlink; Competitor analysis; Link analysis; Context (archaeology); Link (geometry); Content analysis; Business information; Business; Sample (material); Computer science; World Wide Web; Marketing; Advertising; Web page; Sociology; Geography","retraction":null,"screen_n_in":null,"score":{"opus":0.05904947856334476,"gpt":0.2712085113588207,"spread":0.2121590327954759,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics","scholarly_communication"],"consensus_categories":["bibliometrics"],"category_scores_codex":[0.001502558,0.0002207719,0.0004001511,0.01499682,0.0002680499,0.001101478,0.001799294,0.0001227613,0.00005547744],"category_scores_gemma":[0.002102505,0.00019184,0.0002148172,0.1944064,0.00008021262,0.001008838,0.0005689597,0.0001564037,0.0003438093],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002200326,"about_ca_system_score_gemma":0.00009476198,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000939315,"about_ca_topic_score_gemma":0.0001985891,"domain_scores_codex":[0.9964789,0.00004412284,0.0005171584,0.0007628429,0.001548626,0.000648408],"domain_scores_gemma":[0.9963806,0.0002814756,0.0002054839,0.00114016,0.001673225,0.0003191089],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00001732846,0.0009241123,0.8901268,0.00006374026,0.0002085327,0.00004831605,0.0003350347,0.01730595,0.0008610658,0.01601117,0.05537192,0.01872602],"study_design_scores_gemma":[0.0003108218,0.00008395342,0.9069355,0.00001053038,0.00007965986,0.000004255188,0.00004441336,0.02716839,0.001020737,0.000288133,0.06367464,0.0003790033],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2898587,0.0003309347,0.7007824,0.001941301,0.0006153736,0.0002633019,0.00002527955,0.000272604,0.005910061],"genre_scores_gemma":[0.9467893,0.00001987651,0.04609479,0.004256386,0.0001218293,0.00001657916,0.00002286419,0.00001215428,0.002666199],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6569306,"threshold_uncertainty_score":0.9999354,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2524791359","doi":"10.1007/s11192-016-2134-8","title":"The linguistic patterns and rhetorical structure of citation context: an approach using n-grams","year":2016,"lang":"en","type":"article","venue":"Scientometrics","topic":"Topic Modeling","field":"Computer Science","cited_by":50,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Université du Québec à Montréal","funders":"","keywords":"Rhetorical question; Citation; Linguistics; Context (archaeology); Computer science; Linguistic context; Function (biology); Natural (archaeology); Natural language processing; Information retrieval; Linguistic analysis; History; Library science; Philosophy; Biology","retraction":null,"screen_n_in":null,"score":{"opus":0.08286568018633636,"gpt":0.3170019493522849,"spread":0.2341362691659486,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007221449,0.00007224044,0.00009236571,0.000492243,0.0001561838,0.000180824,0.000571169,0.00004254263,8.407481e-7],"category_scores_gemma":[0.001240445,0.00003958286,0.00002012096,0.002405575,0.0000808344,0.0002217495,0.000165448,0.00005785301,3.614922e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005509664,"about_ca_system_score_gemma":0.00004327997,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004154988,"about_ca_topic_score_gemma":0.000007373903,"domain_scores_codex":[0.9986974,0.0000668227,0.0001863581,0.0002959526,0.0005499657,0.0002034971],"domain_scores_gemma":[0.9988305,0.0003087219,0.0001082421,0.0004034295,0.0002509523,0.00009814324],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000004558396,0.0000559617,0.06857125,0.00003529806,0.00001242354,0.000001826654,0.003293814,0.0002860831,0.006552438,0.1421149,0.000008265692,0.7790632],"study_design_scores_gemma":[0.0005367494,0.0001068045,0.03740083,0.00003714938,0.00001176787,0.00001884729,0.0001949255,0.946605,0.001767439,0.01270424,0.0003728088,0.0002434231],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4092207,0.00008304675,0.5901917,0.00003396341,0.000377021,0.00005003508,0.000003036854,0.00001535317,0.00002511231],"genre_scores_gemma":[0.9626554,0.000008811357,0.03722548,0.00002188315,0.00006384227,5.666193e-7,4.369285e-7,0.000003729841,0.00001984146],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9463189,"threshold_uncertainty_score":0.174369,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1993126323","doi":"10.1556/scient.67.2006.3.7","title":"Indirect patent citations","year":2006,"lang":"en","type":"article","venue":"Scientometrics","topic":"Intellectual Property and Patents","field":"Business, Management and Accounting","cited_by":50,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Center for Interuniversity Research and Analysis on Organizations; University of Ottawa; Statistics Canada","funders":"","keywords":"Index (typography); Quality (philosophy); Citation; Measure (data warehouse); Variable (mathematics); Citation analysis; Computer science; Information retrieval; Mathematics; Data mining; Library science; Epistemology; World Wide Web; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.2538094299036781,"gpt":0.2522532225220449,"spread":0.001556207381633257,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0004973737,0.00009074307,0.00008620462,0.002773249,0.0002672363,0.0005200324,0.000223346,0.00003914983,0.0005256935],"category_scores_gemma":[0.0004841019,0.00007289508,0.00005370519,0.01163498,0.00004641608,0.0005555399,0.00009036941,0.00007825599,0.003062908],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003257796,"about_ca_system_score_gemma":0.00001261618,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005490694,"about_ca_topic_score_gemma":0.00003167677,"domain_scores_codex":[0.9988542,0.000004953699,0.0001659317,0.000197611,0.0004992402,0.0002781129],"domain_scores_gemma":[0.9994577,0.00004334864,0.00007759721,0.0001303021,0.0002814727,0.000009628807],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005166608,0.001006417,0.3591878,0.0002187514,0.00004485017,0.00003935561,0.0002310874,0.001675166,0.002191654,0.09812512,0.4781319,0.05909617],"study_design_scores_gemma":[0.0007371397,0.00002603009,0.2533346,0.00002969126,0.0000482626,0.000002596595,0.0001108422,0.01146794,0.0007843666,0.01124401,0.7216616,0.000552959],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6574904,0.000173239,0.001662196,0.0003531256,0.001561446,0.0002007163,0.000003948608,0.0001950443,0.3383599],"genre_scores_gemma":[0.9935892,0.000002316413,0.0001886621,0.0008696354,0.0005142645,0.000006574587,0.00003779459,0.00001085207,0.004780687],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3360988,"threshold_uncertainty_score":0.9977133,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"agree"},{"id":"W2102705305","doi":"10.1007/s11192-008-0106-3","title":"The declining scientific impact of theses: Implications for electronic thesis and dissertation repositories and graduate studies","year":2007,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":50,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; McGill University","funders":"","keywords":"Library science; Altmetrics; Graduate students; Peer review; Sociology; Political science; Public relations; Computer science; Pedagogy; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.7238816379522245,"gpt":0.6807475618396931,"spread":0.04313407611253139,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.07345004,0.0001568206,0.0003083725,0.04414637,0.00204759,0.003741698,0.001187846,0.00006487826,0.000001544644],"category_scores_gemma":[0.1663635,0.00008314037,0.0001606245,0.2367376,0.001217727,0.000598426,0.0004067023,0.0001499923,0.0000019878],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001881631,"about_ca_system_score_gemma":0.0003140082,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004325264,"about_ca_topic_score_gemma":0.00008884861,"domain_scores_codex":[0.9932266,0.0001110061,0.0008981966,0.0007831079,0.004088397,0.0008926188],"domain_scores_gemma":[0.9441748,0.0442098,0.0006803746,0.0008826781,0.00966383,0.0003885278],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00006966767,0.00007401209,0.5932082,0.00001408628,0.00008713714,3.598331e-7,0.002317592,0.00001734463,0.003610668,0.01616162,0.0009971547,0.3834422],"study_design_scores_gemma":[0.0002467569,0.0003743026,0.941357,0.000007885856,0.00002087444,0.000006007333,0.004757581,0.0009631797,0.002044254,0.04886956,0.001223012,0.0001295736],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9818325,0.009835767,0.006632156,0.0005102321,0.0005240319,0.0004507179,0.0000269202,0.00001534268,0.0001723227],"genre_scores_gemma":[0.9971282,0.001342463,0.0009149256,0.000008130403,0.00004002873,0.00002015518,0.000002321177,0.00000938659,0.0005343276],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3833126,"threshold_uncertainty_score":0.9992516,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1990436954","doi":"10.1007/s11192-014-1386-4","title":"Forty years of gender disparities in Russian science: a historical bibliometric analysis","year":2014,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":50,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; Université de Montréal","funders":"National Science Foundation","keywords":"Bibliometrics; Scientometrics; Political science; Sociology; Library science; Social science; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.4898121462066615,"gpt":0.5481804217127789,"spread":0.0583682755061174,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.1063659,0.0002678087,0.001058005,0.9673125,0.0003187365,0.002787581,0.006775371,0.0002129837,0.0003331427],"category_scores_gemma":[0.1949315,0.0002168584,0.0005343935,0.994532,0.00115386,0.001603256,0.001717935,0.0004808527,0.0001260794],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001534838,"about_ca_system_score_gemma":0.0006794594,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009924469,"about_ca_topic_score_gemma":0.0000649958,"domain_scores_codex":[0.960107,0.0003618723,0.001893431,0.001768663,0.03411842,0.001750655],"domain_scores_gemma":[0.9827145,0.008040538,0.0008120106,0.002179148,0.004898362,0.001355455],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000009892534,0.0002945871,0.8801878,0.000008443941,0.00002509572,0.000005218426,0.0001179889,0.0003599234,0.0001566562,0.004329752,0.0009458121,0.1135588],"study_design_scores_gemma":[0.0004781399,0.0001489762,0.9695664,0.000003248092,0.00003810822,0.000002094445,0.00006832961,0.01747897,0.0002242575,0.005777599,0.005944925,0.0002689192],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.961884,0.001317306,0.02042077,0.0001783864,0.001292742,0.0002601983,0.00003084205,0.000036368,0.01457936],"genre_scores_gemma":[0.9953027,0.0001948203,0.003213406,0.00007944545,0.00005870073,0.00001088398,0.000003628123,0.00001527516,0.001121081],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1132899,"threshold_uncertainty_score":0.9985985,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2775516587","doi":"10.1007/s11192-017-2593-6","title":"Assessing the effect of the United States’ “citation advantage” on other countries’ scientific impact as measured in the Web of Science (WoS) database","year":2017,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":49,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Citation; Web of science; Relation (database); Citation database; Visibility; Citation analysis; Computer science; Database; Citation impact; Data science; Political science; Regional science; Library science; Scopus; Sociology; Geography; Law; MEDLINE","retraction":null,"screen_n_in":null,"score":{"opus":0.4570312116954996,"gpt":0.6150197274678524,"spread":0.1579885157723528,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics","sts"],"category_scores_codex":[0.2244995,0.0002839547,0.0004604514,0.1217754,0.003741613,0.01546373,0.01670086,0.00009725178,0.0001120579],"category_scores_gemma":[0.2981451,0.0001108133,0.0002794201,0.5253583,0.008851208,0.002672829,0.001675385,0.0005848216,0.00007838703],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003891235,"about_ca_system_score_gemma":0.001763854,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008109189,"about_ca_topic_score_gemma":0.00006985458,"domain_scores_codex":[0.9587404,0.001728949,0.001157134,0.001193126,0.03602387,0.00115652],"domain_scores_gemma":[0.9596142,0.02317583,0.002185164,0.006391074,0.008257183,0.0003766158],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001131474,0.000300284,0.9568405,0.00003531785,0.00001898952,0.000007938786,0.001732373,0.001898575,0.01502512,0.001551247,0.00347185,0.01900466],"study_design_scores_gemma":[0.001467529,0.000692023,0.9217571,0.000115877,0.00002756931,0.0000085825,0.002704873,0.03122532,0.03181839,0.001203832,0.0087111,0.0002677793],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9921231,0.0002424871,0.0003370418,0.0009744827,0.00190051,0.0009372737,0.0002002702,0.000009458673,0.00327537],"genre_scores_gemma":[0.9994054,0.00004531843,0.00006423931,0.0001471989,0.0000314869,0.00001333334,0.000007033022,0.0000137132,0.0002722459],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.403583,"threshold_uncertainty_score":0.9975554,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2140162241","doi":"10.1007/s11192-011-0589-1","title":"Author disambiguation using multi-aspect similarity indicators","year":2011,"lang":"en","type":"article","venue":"Scientometrics","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":48,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada; Genome Canada","keywords":"Similarity (geometry); Recall; Precision and recall; Computer science; Task (project management); Measure (data warehouse); Information retrieval; Key (lock); Data mining; Artificial intelligence; Natural language processing; Psychology; Cognitive psychology","retraction":null,"screen_n_in":null,"score":{"opus":0.2003679371428773,"gpt":0.3850923397812263,"spread":0.184724402638349,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.001450396,0.000154435,0.0001825386,0.005632685,0.0002468751,0.0001707102,0.001437407,0.00009159531,0.00002730618],"category_scores_gemma":[0.0009301903,0.0001457812,0.0001027673,0.0292352,0.0001241912,0.00112845,0.0005025721,0.0001777672,0.00003255807],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002435515,"about_ca_system_score_gemma":0.00008151416,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007318229,"about_ca_topic_score_gemma":0.00001171848,"domain_scores_codex":[0.9977328,0.00007072586,0.0003002518,0.0005834282,0.0009090232,0.0004037952],"domain_scores_gemma":[0.9984842,0.00006259727,0.0002368593,0.0008169024,0.0001971705,0.0002023053],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000009446847,0.001864625,0.3969854,0.00004410848,0.0001100204,0.00007000408,0.01015378,0.0002472493,0.008101798,0.125403,0.0006532187,0.4563573],"study_design_scores_gemma":[0.0005977476,0.0001790324,0.3545432,0.00002753197,0.00007081143,0.00001526575,0.0001162244,0.5287514,0.09066588,0.02152439,0.002424808,0.001083638],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06664888,0.00007596157,0.9314774,0.00003842164,0.0002942264,0.0001413117,0.000002071367,0.0004454151,0.0008762671],"genre_scores_gemma":[0.5548025,0.000003072061,0.4450245,0.00005052084,0.00001237458,0.000003334687,0.000001051423,0.00000629034,0.00009630904],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5285042,"threshold_uncertainty_score":0.9913989,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2035191842","doi":"10.1007/s11192-006-0133-x","title":"Comparing business competition positions based on Web co-link data: The global market vs. the Chinese market","year":2006,"lang":"en","type":"article","venue":"Scientometrics","topic":"Web visibility and informetrics","field":"Computer Science","cited_by":47,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Multidimensional scaling; Competition (biology); Similarity (geometry); Business; Competitive advantage; Link (geometry); Web site; Market share; Industrial organization; Marketing; Computer science; The Internet; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.0310848158395885,"gpt":0.2980595454444924,"spread":0.2669747296049039,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.005143519,0.0002542378,0.000238723,0.001445927,0.001124578,0.001686889,0.004861056,0.00008810701,0.00008963929],"category_scores_gemma":[0.001677514,0.0001492452,0.00009569216,0.04925736,0.0003507506,0.001232179,0.0009241946,0.0003106436,0.0001003236],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003140129,"about_ca_system_score_gemma":0.0003097867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002077989,"about_ca_topic_score_gemma":0.0001266101,"domain_scores_codex":[0.9961457,0.0002414302,0.0004741487,0.0006806065,0.001893774,0.0005642959],"domain_scores_gemma":[0.9950925,0.001313894,0.000226017,0.002721716,0.0005190619,0.0001268182],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001482008,0.001440352,0.5438882,0.0001333397,0.00004291904,0.00002474909,0.00009040601,0.01937346,0.00002094004,0.1051034,0.3074465,0.02228748],"study_design_scores_gemma":[0.0003146132,0.00003400122,0.4882874,0.00001351305,0.000007377316,0.000007970184,0.000006744972,0.4847694,0.000004076539,0.0006825743,0.02574037,0.0001318945],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04637431,0.0002796288,0.6097994,0.01797105,0.00327514,0.001012933,0.0007309603,0.000450414,0.3201061],"genre_scores_gemma":[0.9929336,0.00002036455,0.00432794,0.00195857,0.000275738,0.00001073215,0.0001720753,0.000008337264,0.000292648],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9465593,"threshold_uncertainty_score":0.9993495,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2098507569","doi":"10.1007/s11192-015-1788-y","title":"Bibliometric analysis of output and impact based on CRIS data: a case study on the registered output of a Dutch university","year":2015,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":47,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Bibliometrics; Regional science; Sociology; Operations research; Computer science; Econometrics; Data science; Library science; Economics; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.8230981755742243,"gpt":0.5877561517828117,"spread":0.2353420237914126,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.08615682,0.0003086135,0.001003928,0.9048277,0.0003301841,0.001443172,0.005802395,0.0001410887,0.00007358399],"category_scores_gemma":[0.1827513,0.0001795105,0.0003646572,0.9811693,0.0006142933,0.0007273988,0.002309362,0.0003870095,0.00002673166],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003396087,"about_ca_system_score_gemma":0.0007678303,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002709073,"about_ca_topic_score_gemma":0.000117376,"domain_scores_codex":[0.9740435,0.001167949,0.00111391,0.001593014,0.02134882,0.0007327967],"domain_scores_gemma":[0.9622369,0.02115999,0.001233626,0.005158831,0.009058973,0.001151648],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002068575,0.002056023,0.9264165,0.000008197001,0.0003775668,0.0004309792,0.0008712417,0.002048126,0.00001127176,0.00006946288,0.01462197,0.05288185],"study_design_scores_gemma":[0.00177856,0.002863061,0.825218,0.000008988508,0.0003901591,0.00001944393,0.01268499,0.154682,0.00004102434,0.00006035463,0.001984555,0.0002688932],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9924242,0.0003158368,0.003008229,0.0003371967,0.0003046703,0.0006589211,0.0009107823,0.00001806854,0.002022089],"genre_scores_gemma":[0.9988115,0.00004145619,0.0002826918,0.00006080344,0.00001843423,0.000001585766,0.00001218566,0.00001236445,0.0007589151],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1526339,"threshold_uncertainty_score":0.9995934,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2109757953","doi":"10.1007/s11192-011-0515-6","title":"Exploring the web visibility of world-class universities","year":2011,"lang":"en","type":"article","venue":"Scientometrics","topic":"Web visibility and informetrics","field":"Computer Science","cited_by":44,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"University of British Columbia","keywords":"Visibility; Proxy (statistics); The Internet; Computer science; World Wide Web; Test (biology); Corporate governance; Information retrieval; Geography; Business","retraction":null,"screen_n_in":null,"score":{"opus":0.2267615598288774,"gpt":0.2801388686842395,"spread":0.05337730885536215,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.002879748,0.0001145096,0.0001617489,0.004310058,0.0002076769,0.0001146222,0.00232329,0.00002712317,0.00004161273],"category_scores_gemma":[0.0008330942,0.00008257484,0.0001132919,0.04230254,0.0003069126,0.002138501,0.0007919016,0.0001725259,0.00004653787],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001670407,"about_ca_system_score_gemma":0.0002935723,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008843798,"about_ca_topic_score_gemma":0.00002615241,"domain_scores_codex":[0.9979497,0.00006681293,0.0003286779,0.000339001,0.0009580336,0.0003577858],"domain_scores_gemma":[0.9975728,0.0004061844,0.0001622063,0.001135931,0.00059703,0.0001257927],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.00003342226,0.000601479,0.05422524,0.00013973,0.00005788921,0.00000983531,0.02135672,0.0001134469,0.0002436468,0.7705525,0.002718843,0.1499473],"study_design_scores_gemma":[0.001113652,0.0007071398,0.824467,0.00005757432,0.00004156199,0.000007844661,0.005198156,0.04831082,0.0359715,0.01114433,0.07205407,0.0009263305],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8619646,0.0001235182,0.04310448,0.0001895572,0.002349965,0.0002049669,0.000009140083,0.0001524498,0.09190129],"genre_scores_gemma":[0.9891647,0.0000423004,0.009839698,0.00009632475,0.00002499492,0.000002727822,3.985235e-7,0.00000366494,0.0008252643],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7702418,"threshold_uncertainty_score":0.9780538,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W1993018298","doi":"10.1007/s11192-008-2064-1","title":"The relationship between reviewers’ quality-scores and number of citations for papers published in the journal Physics in Medicine and Biology from 2003–2005","year":2009,"lang":"en","type":"article","venue":"Scientometrics","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":44,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University; Juravinski Cancer Centre","funders":"","keywords":"Impact factor; Citation; Quality (philosophy); Statistics; Mathematics; Medicine; Psychology; Computer science; Library science; Biology; Philosophy; Epistemology","retraction":null,"screen_n_in":null,"score":{"opus":0.8825900474161539,"gpt":0.626060316233719,"spread":0.2565297311824348,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2101405,0.0001156354,0.001152583,0.001173738,0.0002422672,0.0004643724,0.0009572013,0.00005643813,0.0001385351],"category_scores_gemma":[0.393099,0.00004203422,0.0001512442,0.01771089,0.000251028,0.0002647469,0.00003625482,0.0001946778,0.00000705657],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002289267,"about_ca_system_score_gemma":0.00005842571,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003503761,"about_ca_topic_score_gemma":0.0001008247,"domain_scores_codex":[0.9853525,0.006903298,0.004737137,0.0004056272,0.002349989,0.0002514349],"domain_scores_gemma":[0.9202477,0.07382497,0.003133577,0.001215944,0.001452783,0.0001250047],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00000200282,0.000012484,0.9491445,0.000007265998,0.00001110978,6.61503e-8,0.001588608,0.000003219115,0.000004709561,0.01467165,0.01274917,0.02180522],"study_design_scores_gemma":[0.0002570283,0.00002723302,0.8674288,0.00003180399,0.0000573269,0.000001112706,0.001480971,0.0001218415,2.694507e-7,0.1088328,0.02171513,0.00004570311],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9327953,0.02240387,0.01557538,0.02452932,0.0005049249,0.001420386,0.00006921892,0.000002176934,0.002699381],"genre_scores_gemma":[0.9948215,0.0008311295,0.00363151,0.0003484742,0.0001000152,0.000008058087,0.00001049769,0.000002426872,0.0002463713],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1829585,"threshold_uncertainty_score":0.8509498,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1979812464","doi":"10.1007/s11192-010-0249-x","title":"Mathematics 1868–2008: a bibliometric analysis","year":2010,"lang":"en","type":"article","venue":"Scientometrics","topic":"Complex Systems and Time Series Analysis","field":"Economics, Econometrics and Finance","cited_by":43,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Productivity; Distribution (mathematics); Range (aeronautics); Mathematics; Bibliometrics; Quarter (Canadian coin); Statistics; Mathematical economics; Econometrics; History; Library science; Economics; Computer science; Economic growth","retraction":null,"screen_n_in":null,"score":{"opus":0.04005720797298853,"gpt":0.2584068437551288,"spread":0.2183496357821403,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","bibliometrics","insufficient_payload"],"consensus_categories":["bibliometrics","insufficient_payload"],"category_scores_codex":[0.002787555,0.0002387382,0.0008427789,0.3593729,0.000272976,0.000619211,0.0008242433,0.0001571821,0.004938465],"category_scores_gemma":[0.001464239,0.0002493479,0.0006385334,0.7178043,0.0001283555,0.0003612175,0.0002095707,0.0003133886,0.002029994],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007385346,"about_ca_system_score_gemma":0.00002926407,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001207156,"about_ca_topic_score_gemma":0.00030495,"domain_scores_codex":[0.9972323,0.00001045569,0.001069627,0.0007236893,0.0003451391,0.0006187649],"domain_scores_gemma":[0.997393,0.0001928768,0.0006531982,0.001148919,0.0002661692,0.0003458723],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000002248869,0.0004435472,0.811353,0.00004496218,0.00108076,0.00001398128,0.0005758913,0.0001985144,0.0001349184,0.1754548,0.002654193,0.008043173],"study_design_scores_gemma":[0.0007733971,0.0001139688,0.5448301,0.000006138839,0.0004372613,0.00002934611,0.0004539177,0.05939784,0.00008568605,0.01284718,0.379807,0.00121818],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9140246,0.001832613,0.02715698,0.0001885805,0.001313535,0.0002139754,0.000169538,0.0001319803,0.05496819],"genre_scores_gemma":[0.9838277,0.00007992041,0.01016022,0.00007116372,0.0001453048,0.00001526735,0.0000215763,0.00002773715,0.005651088],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3771528,"threshold_uncertainty_score":0.9999959,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics"],"domain":null,"study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W1724869758","doi":"10.1023/a:1014256102041","title":"Collaboration and Cognitive Structures in Social Science Research Fields. Towards Socio-Cognitive Analysis in Information Systems","year":2001,"lang":"en","type":"article","venue":"Scientometrics","topic":"Knowledge Management and Sharing","field":"Social Sciences","cited_by":43,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Mainstream; Consolidation (business); Cognition; Field (mathematics); Sociology; Value (mathematics); Social network analysis; Knowledge management; Epistemology; Social science; Computer science; Psychology; Political science; Business; Social capital","retraction":null,"screen_n_in":null,"score":{"opus":0.09573227246461083,"gpt":0.4494395081219599,"spread":0.3537072356573491,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["bibliometrics","scholarly_communication"],"consensus_categories":["bibliometrics"],"category_scores_codex":[0.009587261,0.00007555395,0.0001627639,0.01394532,0.001121019,0.0013179,0.0002721867,0.000102891,0.00002303114],"category_scores_gemma":[0.005288403,0.00007779912,0.00002486976,0.08850171,0.0008013732,0.001986122,0.000176643,0.0002215058,0.00001079125],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004541039,"about_ca_system_score_gemma":0.0003747272,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002198567,"about_ca_topic_score_gemma":0.00355172,"domain_scores_codex":[0.9969449,0.0002643228,0.0002603224,0.0002679544,0.001748899,0.0005135717],"domain_scores_gemma":[0.9980844,0.0003125094,0.00008765933,0.00005578708,0.001367161,0.00009247985],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00003710567,0.00005849356,0.6752026,0.00003205007,0.00003249327,0.000007302497,0.1890233,0.00005790975,0.000005881931,0.04560138,0.0001015824,0.08983993],"study_design_scores_gemma":[0.0005883457,0.00003281542,0.7922537,0.00002757764,0.00003803397,1.250148e-7,0.2011205,0.003173427,0.00001524994,0.00143819,0.001147673,0.0001642822],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8832985,0.0001846134,0.0005126857,0.0001987395,0.000229792,0.0004538346,0.000008257317,0.00001891726,0.1150947],"genre_scores_gemma":[0.9993936,0.0001122909,0.00001763985,0.00002969871,0.00008288825,0.00002116191,0.00001090548,0.000002132394,0.0003297219],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1170511,"threshold_uncertainty_score":0.9997188,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["bibliometrics"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics"],"domain":null,"study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4220877524","doi":"10.1007/s11192-022-04338-1","title":"Interdisciplinarity and impact: the effects of the citation time window","year":2022,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":42,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Université du Québec à Montréal","funders":"National Planning Office of Philosophy and Social Science","keywords":"Citation; Impact factor; Window (computing); Citation analysis; Scientometrics; Computer science; Bibliometrics; Data science; Citation impact; Statistics; Data mining; Mathematics; Library science; Political science; World Wide Web; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.2744588650714582,"gpt":0.5399536004006232,"spread":0.265494735329165,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.03956011,0.0001324373,0.0002468147,0.06512869,0.001478551,0.001309289,0.004119882,0.00004400824,0.0003657265],"category_scores_gemma":[0.09398838,0.0000609194,0.0002330613,0.547464,0.0005790234,0.0003567418,0.005891925,0.000455791,0.00004043234],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002523708,"about_ca_system_score_gemma":0.0002350215,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000510144,"about_ca_topic_score_gemma":0.000002987311,"domain_scores_codex":[0.9820014,0.0007625456,0.0005336002,0.000574779,0.01561731,0.0005104088],"domain_scores_gemma":[0.9806128,0.0153376,0.0005072693,0.00125209,0.002010617,0.0002796394],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00004855092,0.0004722862,0.6917716,0.0000302875,0.00004792501,0.000006802905,0.003469947,0.000586116,0.01055285,0.001568631,0.05059177,0.2408533],"study_design_scores_gemma":[0.0003983947,0.0003706958,0.9784735,0.000004056541,0.00001127431,0.00001340697,0.0005296089,0.006640126,0.001080068,0.008498293,0.003869749,0.0001108377],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9933278,0.001145712,0.0004284798,0.001429759,0.001393721,0.0005191043,0.00005880042,0.00001135446,0.0016853],"genre_scores_gemma":[0.9969249,0.00002668843,0.00007250484,0.0001155192,0.00002902099,0.00002020112,0.000001510389,0.000007270445,0.00280243],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4823353,"threshold_uncertainty_score":0.9998214,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics"],"domain":null,"study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2900764028","doi":"10.1007/s11192-018-2959-4","title":"Capturing the economic value of triadic patents","year":2018,"lang":"en","type":"article","venue":"Scientometrics","topic":"Intellectual Property and Patents","field":"Business, Management and Accounting","cited_by":42,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Value (mathematics); Patent analysis; Cover (algebra); Test (biology); European patent office; Patent application; Business; Computer science; Regional science; International trade; Data science; Political science; Sociology; Engineering; Law; Machine learning","retraction":null,"screen_n_in":null,"score":{"opus":0.177109436578545,"gpt":0.2610980654164781,"spread":0.08398862883793304,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001198201,0.00009262584,0.0001178838,0.001254023,0.0002510829,0.0001896577,0.0005174353,0.00003925306,0.0005552039],"category_scores_gemma":[0.000594839,0.00005714381,0.00006796859,0.003319296,0.0001959947,0.0004062807,0.0002125652,0.00007867107,0.002186112],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005203119,"about_ca_system_score_gemma":0.0000283689,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008740136,"about_ca_topic_score_gemma":0.00002138531,"domain_scores_codex":[0.9989766,0.00001053099,0.0002295918,0.0001905112,0.0003327577,0.0002599667],"domain_scores_gemma":[0.9993346,0.00006756844,0.0001632857,0.0002356159,0.0001873976,0.00001147741],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00102519,0.0009136067,0.3594777,0.001041944,0.000576424,0.00002181553,0.006285347,0.003257859,0.0065138,0.1771792,0.2078862,0.2358209],"study_design_scores_gemma":[0.003638256,0.0002523979,0.113529,0.0001787487,0.0002419339,0.000008204767,0.000939181,0.2317584,0.01578495,0.01181125,0.6206933,0.001164347],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.935585,0.00007873015,0.000250739,0.0001655923,0.002297948,0.0001712639,0.000002533049,0.00003226893,0.06141598],"genre_scores_gemma":[0.9974622,0.000006689481,0.00003649023,0.0006642556,0.0008614958,0.000002599785,0.000003083763,0.00001046461,0.0009527361],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4128071,"threshold_uncertainty_score":0.9985908,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}