{"meta":{"query_hash":"f029b2e5f93c","filters":{"venue":"Data & Knowledge Engineering"},"cohort_total":41,"direct_labels_cover":1,"predictions_cover":41,"exported":41,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/f029b2e5f93c","api":"https://metacan.xera.ac/api/v1/cohort?venue=Data+%26+Knowledge+Engineering"},"results":[{"id":"W1966678916","doi":"10.1016/j.datak.2009.08.006","title":"Sorting improves word-aligned bitmap indexes","year":2009,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Parallel Computing and Optimization Techniques","field":"Computer Science","cited_by":74,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of New Brunswick; Université du Québec à Montréal","funders":"","keywords":"Bitmap; Lexicographical order; Sorting; Sorting algorithm; sort; Aggregate (composite); Table (database); Construct (python library); Simple (philosophy)","score_opus":0.02252072736554118,"score_gpt":0.27281816931913383,"score_spread":0.25029744195359266,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1966678916","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.3617039,0.0051477733,0.5586286,0.0012592231,0.002498189,0.00040519008,0.0035910797,0.03648591,0.030280167],"genre_scores_gemma":[0.38166526,0.001495872,0.58687764,0.00061650755,0.0004595474,0.00020608747,0.008856897,0.0028518713,0.016970329],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9982309,0.0003093193,0.00023659292,0.00027939607,0.0007070568,0.00023678501],"domain_scores_gemma":[0.9940929,0.0017526752,0.00035036707,0.0018402339,0.0017611088,0.00020269821],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011346228,0.0011300278,0.0017138893,0.0030896186,0.001407459,0.003070682,0.002172641,0.0013688529,0.013999183],"category_scores_gemma":[0.009863682,0.00067327166,0.000763428,0.009103471,0.000631211,0.0062407656,0.0019835813,0.00095460156,0.006433142],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0024664847,0.00075980433,0.0035285389,0.0006320882,0.00014199663,0.0002928868,0.00041139845,0.032972332,0.033684313,0.034082953,0.046006598,0.8450206],"study_design_scores_gemma":[0.0008349061,0.0016357615,0.0034534617,0.00021540614,0.00039028193,0.0008310204,0.001057229,0.6145759,0.13067664,0.18289539,0.06319257,0.00024136055],"about_ca_topic_score_codex":0.003399842,"about_ca_topic_score_gemma":0.007168607,"teacher_disagreement_score":0.013999183,"about_ca_system_score_codex":0.0007840415,"about_ca_system_score_gemma":0.0027621067,"threshold_uncertainty_score":0.046831965},"labels":[],"label_agreement":null},{"id":"W1983768816","doi":"10.1016/s0169-023x(02)00133-7","title":"Optimizing temporal queries: efficient handling of duplicates","year":2002,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; SQL; Query language; Query optimization; Query by Example; Relational database; Temporal database; Semantics (computer science); Expressive power; Context (archaeology); Sargable; Information retrieval; Programming language; Database; Web search query; Search engine","score_opus":0.038932829926481824,"score_gpt":0.24963580992641615,"score_spread":0.21070297999993431,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1983768816","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.12335644,0.003768716,0.8495,0.0013058537,0.000471572,0.0002952661,0.0016716305,0.012356844,0.007273614],"genre_scores_gemma":[0.42385888,0.0013905487,0.5616745,0.0005689126,0.0004344936,0.00021471365,0.003538725,0.0020316404,0.006287615],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.991804,0.0017881871,0.0009356915,0.0010807771,0.0036944286,0.0006969456],"domain_scores_gemma":[0.9838435,0.007553881,0.0010466847,0.00454368,0.0026548652,0.00035741914],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0053675915,0.0012552501,0.0023585754,0.0018615174,0.001633156,0.0039922534,0.00483478,0.0017365152,0.0031667405],"category_scores_gemma":[0.020735139,0.0011909095,0.0012032223,0.004802144,0.0013147503,0.008217559,0.0034673344,0.0015249257,0.001405733],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0039443593,0.000632574,0.0064471005,0.0010276762,0.00040683575,0.00097196875,0.0013367884,0.07599457,0.04906671,0.04932555,0.06717879,0.7436671],"study_design_scores_gemma":[0.0006721465,0.00043692297,0.0014538608,0.00008640213,0.00043631924,0.0013192982,0.0007997855,0.8034211,0.06544702,0.09697572,0.028797667,0.00015383071],"about_ca_topic_score_codex":0.005861953,"about_ca_topic_score_gemma":0.006236839,"teacher_disagreement_score":0.005861953,"about_ca_system_score_codex":0.0012579113,"about_ca_system_score_gemma":0.0030307958,"threshold_uncertainty_score":0.02838689},"labels":[],"label_agreement":null},{"id":"W1984778493","doi":"10.1016/j.datak.2010.01.005","title":"The consistency extractor system: Answer set programs for consistent query answering in databases","year":2010,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Logic, Reasoning, and Knowledge","field":"Computer Science","cited_by":82,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University","funders":"","keywords":"Computer science; Datalog; Answer set programming; Database theory; Consistency (knowledge bases); Extractor; Programming language; Relational database; Deductive database; Set (abstract data type); Information retrieval; Database; Artificial intelligence","score_opus":0.055622060533341455,"score_gpt":0.2879527423906523,"score_spread":0.23233068185731087,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1984778493","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00322614,0.00017481594,0.9799968,0.00045840518,0.000030370806,0.00017760006,0.00066135486,0.014592692,0.00068192923],"genre_scores_gemma":[0.075745046,0.00035356008,0.91466457,0.00066744536,0.00013718671,0.00049894326,0.0028572853,0.0028106878,0.0022653602],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9923626,0.0024571433,0.00076646794,0.0010325984,0.002982114,0.00039904678],"domain_scores_gemma":[0.98650163,0.009328169,0.000618442,0.0020974213,0.0012432856,0.00021106686],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010010423,0.0013638715,0.0022853294,0.0024639326,0.0012071696,0.004572512,0.005496748,0.0027415676,0.008407577],"category_scores_gemma":[0.0319539,0.0019208282,0.0021670319,0.0021616009,0.002805467,0.009835663,0.00582619,0.0046742735,0.0022660238],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016901261,0.0006278694,0.0043836236,0.0015299949,0.00043470095,0.0005523483,0.0014852867,0.041137077,0.017221224,0.21854225,0.06872012,0.64367545],"study_design_scores_gemma":[0.0007000509,0.00027331267,0.0009731199,0.00027715354,0.00048975984,0.00042243622,0.00035018005,0.60529876,0.054335535,0.2889946,0.04770012,0.0001849462],"about_ca_topic_score_codex":0.0031389706,"about_ca_topic_score_gemma":0.0036620395,"teacher_disagreement_score":0.010010423,"about_ca_system_score_codex":0.0010650404,"about_ca_system_score_gemma":0.003533918,"threshold_uncertainty_score":0.052940786},"labels":[],"label_agreement":null},{"id":"W2002413229","doi":"10.1016/j.datak.2012.09.005","title":"Comparing functionality of software systems: An ontological approach","year":2012,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Software Engineering Research","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Software system; Software engineering; Adaptation (eye); Function (biology); USable; Software development; Software; Software metric; Software quality; Point (geometry); Software construction; Function point; Quality (philosophy); Software sizing; World Wide Web; Programming language","score_opus":0.13962881508491679,"score_gpt":0.31436609711026914,"score_spread":0.17473728202535235,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2002413229","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.16646494,0.0005929579,0.8076092,0.0012604959,0.00007145385,0.00031424165,0.0009699265,0.0006111232,0.022105671],"genre_scores_gemma":[0.61139673,0.0005047121,0.3848816,0.00016661505,0.00003227603,0.00018695554,0.0016415454,0.00013221275,0.0010573148],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.992191,0.0030510798,0.0011860228,0.00064001424,0.0024701706,0.00046179522],"domain_scores_gemma":[0.98269147,0.0105126295,0.0010270475,0.0035612213,0.0017499741,0.0004577295],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007802832,0.0006332292,0.0006532074,0.00963484,0.0019320605,0.0070888223,0.002571203,0.001677986,0.00187896],"category_scores_gemma":[0.025253268,0.0005982196,0.0026994017,0.005578896,0.0046209213,0.011985433,0.0034268573,0.0013716918,0.00029469965],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016627609,0.00023236826,0.018761946,0.0006292909,0.00026612228,0.0005868811,0.007017516,0.014035983,0.012522172,0.7852103,0.0011157149,0.15945555],"study_design_scores_gemma":[0.000054570064,0.0003088634,0.017943965,0.00084850774,0.0009353033,0.00092948833,0.011061015,0.09883818,0.012604464,0.8239443,0.032386944,0.00014442587],"about_ca_topic_score_codex":0.0062020007,"about_ca_topic_score_gemma":0.0045564296,"teacher_disagreement_score":0.00963484,"about_ca_system_score_codex":0.0020271393,"about_ca_system_score_gemma":0.0030405226,"threshold_uncertainty_score":0.041265786},"labels":[],"label_agreement":null},{"id":"W2011939484","doi":"10.1016/j.datak.2014.09.004","title":"Privacy-preserving trajectory stream publishing","year":2014,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":46,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; Concordia University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Data publishing; Trajectory; Computer science; Publishing; Internet privacy; World Wide Web; Political science; Law; Physics","score_opus":0.04019297403125713,"score_gpt":0.2642967261602062,"score_spread":0.2241037521289491,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2011939484","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020933822,0.0005567865,0.96812695,0.0020843116,0.0003360628,0.00010468465,0.0013737808,0.0013893953,0.0050941464],"genre_scores_gemma":[0.8055511,0.0018386812,0.17433482,0.0006621237,0.000761645,0.00019958806,0.004008459,0.00035134103,0.012292321],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99560463,0.0010512455,0.000505386,0.0007157988,0.0016809369,0.00044208887],"domain_scores_gemma":[0.97956353,0.004852523,0.00094892795,0.01250774,0.00172861,0.00039871712],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0042969286,0.00057357154,0.0016729932,0.0014357037,0.0014657925,0.005317653,0.0028063743,0.0018287976,0.0034898159],"category_scores_gemma":[0.026856095,0.00061895285,0.0011187263,0.004742379,0.0014105164,0.0076931934,0.0046649673,0.0030372054,0.0021378237],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012826938,0.00024492794,0.0057503125,0.00032955845,0.00021543229,0.00093223853,0.00064558047,0.134086,0.008052793,0.5402718,0.021396061,0.2867926],"study_design_scores_gemma":[0.0001067665,0.00012179322,0.00064024964,0.000059607723,0.00008975974,0.001056358,0.00026487611,0.52753353,0.02071649,0.42785856,0.021510933,0.00004109312],"about_ca_topic_score_codex":0.0010660077,"about_ca_topic_score_gemma":0.000733883,"teacher_disagreement_score":0.005317653,"about_ca_system_score_codex":0.0010449861,"about_ca_system_score_gemma":0.003149205,"threshold_uncertainty_score":0.022724628},"labels":[],"label_agreement":null},{"id":"W2017416607","doi":"10.1016/j.datak.2009.01.003","title":"Formalizing visibility characteristics in hierarchical systems","year":2009,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Service-Oriented Architecture and Web Services","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Visibility; Hierarchy; Computer science; Theoretical computer science; Coherence (philosophical gambling strategy); Artificial intelligence; Mathematics; Physics; Statistics; Optics","score_opus":0.019145620733663993,"score_gpt":0.2609460380770489,"score_spread":0.24180041734338492,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2017416607","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06269003,0.000245188,0.92723674,0.0007433008,0.000051480434,0.00010016967,0.00011888253,0.00052489364,0.008289206],"genre_scores_gemma":[0.7933488,0.00027325444,0.20304583,0.0001383288,0.00011664362,0.00014770764,0.00029386897,0.00026433534,0.0023712306],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99525875,0.0014078743,0.0005742002,0.00067471876,0.0013489053,0.00073561183],"domain_scores_gemma":[0.97613436,0.015540352,0.002015589,0.0032965746,0.0022342263,0.0007788921],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005440859,0.0006482045,0.0007125832,0.0020970819,0.0014695297,0.0052262354,0.0020305358,0.002072178,0.0032273307],"category_scores_gemma":[0.02347123,0.0015085657,0.0018088493,0.0020352362,0.005808865,0.014328287,0.0047654086,0.0043213177,0.0003345872],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00002727973,0.000026048405,0.0010121485,0.00007010857,0.000015587268,0.00012383604,0.00077177736,0.00875184,0.001308215,0.9799472,0.00035640312,0.0075895493],"study_design_scores_gemma":[0.00003305785,0.00002771787,0.000491085,0.00006193609,0.00005345369,0.00012292185,0.00037692994,0.07781404,0.002434619,0.91383505,0.004721788,0.000027435473],"about_ca_topic_score_codex":0.00523557,"about_ca_topic_score_gemma":0.005407174,"teacher_disagreement_score":0.005440859,"about_ca_system_score_codex":0.002241911,"about_ca_system_score_gemma":0.002198125,"threshold_uncertainty_score":0.02877438},"labels":[],"label_agreement":null},{"id":"W2025155398","doi":"10.1016/j.datak.2009.10.001","title":"Skyline queries with constraints: Integrating skyline and traditional query operators","year":2009,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Skyline; Computer science; Pruning; Class (philosophy); Space (punctuation); Operator (biology); Database; Computation; Information retrieval; Data mining; Theoretical computer science; Algorithm; Artificial intelligence","score_opus":0.027096740140654725,"score_gpt":0.23442299360009852,"score_spread":0.2073262534594438,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2025155398","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.014840502,0.0010318325,0.9720415,0.00057678466,0.00007530432,0.00016460716,0.0013283106,0.0075051165,0.0024361562],"genre_scores_gemma":[0.16481128,0.0011451978,0.8260269,0.00039997316,0.00023163577,0.00018301037,0.0035682877,0.0017835874,0.0018501346],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9949195,0.0012761347,0.00050996226,0.0007555536,0.0022892116,0.00024965676],"domain_scores_gemma":[0.990529,0.0038902625,0.00074825925,0.0028303717,0.0015756412,0.00042641276],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051619466,0.0010475712,0.002088022,0.0020678905,0.0008346182,0.0037830002,0.0036125102,0.0011025743,0.00400952],"category_scores_gemma":[0.012761494,0.00073893607,0.00106857,0.0056675966,0.0011302949,0.010911664,0.0045952415,0.00148146,0.000847752],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0022727961,0.0006525393,0.009341892,0.0019556708,0.00057659263,0.0008272638,0.0014356626,0.057702146,0.023225473,0.11147872,0.082315505,0.7082158],"study_design_scores_gemma":[0.0003778154,0.0003532865,0.0017907625,0.00016659849,0.00026655933,0.0008587037,0.0006025313,0.78831375,0.018026875,0.11488999,0.07421883,0.0001343086],"about_ca_topic_score_codex":0.0042892382,"about_ca_topic_score_gemma":0.008845294,"teacher_disagreement_score":0.0051619466,"about_ca_system_score_codex":0.0007951144,"about_ca_system_score_gemma":0.0015747674,"threshold_uncertainty_score":0.027299285},"labels":[],"label_agreement":null},{"id":"W2036207056","doi":"10.1016/s0169-023x(01)00018-0","title":"Using information extraction and natural language generation to answer e-mail","year":2001,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Natural language generation; Question answering; Domain (mathematical analysis); Natural language; Information extraction; Text messaging; Natural language processing; Frequently asked questions; Information retrieval; World Wide Web; Artificial intelligence","score_opus":0.03129771263752603,"score_gpt":0.3164762397476485,"score_spread":0.2851785271101225,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2036207056","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03219078,0.0005265296,0.94171965,0.0012238275,0.00025296063,0.00069764117,0.0014444953,0.016069131,0.005875005],"genre_scores_gemma":[0.16154426,0.00032265333,0.82849073,0.00043246336,0.00016946235,0.0003953925,0.0045519015,0.0004470424,0.0036460576],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99683005,0.0014739912,0.00029262045,0.0005454548,0.0007138678,0.00014412963],"domain_scores_gemma":[0.9907872,0.0064966897,0.0004194024,0.00066234544,0.0015272783,0.00010701821],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027803844,0.0010549312,0.000976931,0.003698568,0.0010324247,0.0019781196,0.0014356142,0.0016543374,0.0058239703],"category_scores_gemma":[0.012890568,0.0005775355,0.0013109179,0.001964611,0.00068010046,0.0032500576,0.001367671,0.0012594027,0.0032265533],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00054631336,0.00072803,0.0039897375,0.0009534042,0.00016388588,0.0010564728,0.0011323817,0.0122111235,0.036254298,0.028645905,0.030619776,0.8836987],"study_design_scores_gemma":[0.00036747498,0.0004230545,0.0025948763,0.0002225447,0.00048681014,0.0013523008,0.0007958126,0.69349056,0.11925326,0.120469965,0.06039759,0.00014577518],"about_ca_topic_score_codex":0.0018805716,"about_ca_topic_score_gemma":0.0025017834,"teacher_disagreement_score":0.0058239703,"about_ca_system_score_codex":0.00070252817,"about_ca_system_score_gemma":0.0014547699,"threshold_uncertainty_score":0.019483149},"labels":[],"label_agreement":null},{"id":"W2043640532","doi":"10.1016/j.datak.2004.12.009","title":"Complexity and clarity in conceptual modeling: Comparison of mandatory and optional properties","year":2005,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":239,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia; Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"CLARITY; Computer science; Grammar; Cognitive psychology; Natural language processing; Rule-based machine translation; Domain (mathematical analysis); Domain model; Linguistics; Psychology; Artificial intelligence; Domain knowledge; Mathematics","score_opus":0.20680406000893842,"score_gpt":0.32137278644686057,"score_spread":0.11456872643792215,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2043640532","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.27978545,0.0017925475,0.691824,0.003411695,0.00015923358,0.00030022842,0.0005127535,0.00054487406,0.02166927],"genre_scores_gemma":[0.91300386,0.0005467187,0.08490016,0.00013762969,0.00009648815,0.0001423402,0.00044410504,0.00018911617,0.0005396481],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9672541,0.01738322,0.002984583,0.0014048365,0.009811774,0.001161466],"domain_scores_gemma":[0.64662063,0.29165712,0.016802039,0.026005566,0.01604778,0.0028668272],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.033827644,0.0006411945,0.001116196,0.005034317,0.0019702164,0.008770312,0.0023377659,0.002362881,0.0034733897],"category_scores_gemma":[0.19015947,0.0012704633,0.0028839002,0.0028487486,0.0062864185,0.031019215,0.0052356417,0.0031492198,0.0003575285],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00083827757,0.00017106756,0.017180722,0.0007463022,0.00023865925,0.0002174487,0.0068275672,0.022756698,0.0036507451,0.88019687,0.0018371355,0.06533858],"study_design_scores_gemma":[0.00010748838,0.00022146276,0.008904579,0.00042598372,0.00036297468,0.0003620112,0.0018703901,0.15590507,0.0059980825,0.8202013,0.005505945,0.00013476469],"about_ca_topic_score_codex":0.0019387964,"about_ca_topic_score_gemma":0.0018489452,"teacher_disagreement_score":0.033827644,"about_ca_system_score_codex":0.0027856736,"about_ca_system_score_gemma":0.0027262915,"threshold_uncertainty_score":0.17889982},"labels":[],"label_agreement":null},{"id":"W2052266824","doi":"10.1016/j.datak.2003.10.007","title":"Conceptual construction on incomplete survey data","year":2003,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Rough Sets and Fuzzy Logic","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Data mining; Knowledge extraction; Raw data; Data science; Conceptual model; Set (abstract data type); Information retrieval; Database","score_opus":0.15590708562556674,"score_gpt":0.294042282112188,"score_spread":0.1381351964866213,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2052266824","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.030097311,0.00048172288,0.9634358,0.0011793188,0.00003784703,0.00008933756,0.00079337624,0.00019630708,0.003688872],"genre_scores_gemma":[0.47482973,0.0010970904,0.5183493,0.0002480794,0.00007311373,0.00044578448,0.0033121712,0.00009957169,0.0015451902],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.98972845,0.0060570273,0.00080859155,0.0012557916,0.0017801956,0.0003699475],"domain_scores_gemma":[0.9586286,0.025428556,0.003037895,0.008625678,0.0036276672,0.0006515453],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01165211,0.00057429576,0.0011070543,0.007155336,0.0015574204,0.005359141,0.0025536413,0.0014687217,0.0042715655],"category_scores_gemma":[0.06188249,0.0009192445,0.0022162336,0.008474491,0.0031563127,0.010923265,0.00457343,0.0021324991,0.00053105084],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00004843199,0.00003874363,0.0035441713,0.0003437542,0.00009462567,0.00019808392,0.0018195733,0.014116709,0.0004341869,0.9074542,0.0024206517,0.06948688],"study_design_scores_gemma":[0.000016155962,0.000032517422,0.001993878,0.00021389454,0.00007101889,0.00024638794,0.0014165362,0.07735794,0.00079846056,0.9038005,0.0140163135,0.000036331756],"about_ca_topic_score_codex":0.0028116715,"about_ca_topic_score_gemma":0.0032128873,"teacher_disagreement_score":0.01165211,"about_ca_system_score_codex":0.0020662497,"about_ca_system_score_gemma":0.0022491778,"threshold_uncertainty_score":0.061622977},"labels":[],"label_agreement":null},{"id":"W2068851374","doi":"10.1016/j.datak.2010.01.001","title":"Towards an accurate functional size measurement procedure for conceptual models in an MDA environment","year":2010,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Service-Oriented Architecture and Web Services","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"École de Technologie Supérieure; Université du Québec à Montréal","funders":"","keywords":"Function point; Computer science; Measure (data warehouse); Function (biology); Software; Phase (matter); Data mining; Software development; Programming language; Physics","score_opus":0.061132141842537235,"score_gpt":0.2597888943664013,"score_spread":0.19865675252386406,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2068851374","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008620186,0.000026859107,0.9878357,0.000060772018,0.000011201431,0.00010411179,0.00008198532,0.002640856,0.00061836414],"genre_scores_gemma":[0.13774365,0.000060767466,0.86026126,0.000046026526,0.000011330431,0.00035637658,0.00039926483,0.00065735425,0.0004640566],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98418033,0.0050832457,0.0017661068,0.0011676104,0.0073214546,0.00048125043],"domain_scores_gemma":[0.9434285,0.019130528,0.004202194,0.015317422,0.017295908,0.0006253636],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014350399,0.0012192555,0.0010270968,0.004813447,0.0010969301,0.0046946546,0.0025249117,0.001626271,0.0017831043],"category_scores_gemma":[0.07634542,0.0011969489,0.0012380743,0.002031049,0.0013952883,0.005999883,0.0033498427,0.0027613752,0.001076231],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002808479,0.0004689683,0.019373631,0.00060820533,0.00023987028,0.00020380039,0.0023697107,0.07512468,0.088788964,0.13522388,0.0054516704,0.6718658],"study_design_scores_gemma":[0.00006186817,0.00039697601,0.009640508,0.00022189847,0.00017941288,0.00027534072,0.0008500298,0.81762016,0.09593045,0.062743805,0.011918462,0.00016103647],"about_ca_topic_score_codex":0.004065884,"about_ca_topic_score_gemma":0.004338401,"teacher_disagreement_score":0.014350399,"about_ca_system_score_codex":0.0017408826,"about_ca_system_score_gemma":0.0032493733,"threshold_uncertainty_score":0.075893044},"labels":[],"label_agreement":null},{"id":"W2073390485","doi":"10.1016/j.datak.2004.10.001","title":"Extracting conceptual relationships from specialized documents","year":2004,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Variety (cybernetics); Granularity; Data science; Conceptual model; Representation (politics); Management science; Scientific modelling; Conceptual framework; Knowledge management; Data mining; Artificial intelligence; Database; Engineering","score_opus":0.09260547872331132,"score_gpt":0.29443267366585935,"score_spread":0.20182719494254803,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2073390485","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.20150718,0.0038445953,0.7502148,0.0017238489,0.0003215769,0.0011387038,0.016191669,0.0075510573,0.017506488],"genre_scores_gemma":[0.23544286,0.0028568832,0.70901805,0.00023221699,0.00016042551,0.00031477082,0.04660063,0.0006619185,0.004712262],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99881136,0.00019479086,0.00020766845,0.00027486897,0.0004064497,0.00010484735],"domain_scores_gemma":[0.9951526,0.0022852712,0.00041399308,0.00097252487,0.0009807583,0.00019486349],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011627361,0.0011237769,0.0012922168,0.010751344,0.0014016312,0.0034246596,0.0015488585,0.0017413212,0.005136579],"category_scores_gemma":[0.009829368,0.0008987123,0.0017604931,0.011093687,0.0006845728,0.006271967,0.0022714227,0.0018121524,0.0031802831],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00037693337,0.0006635707,0.009350288,0.0023135485,0.00027803055,0.0032832988,0.002726171,0.005776055,0.046331298,0.05674441,0.026557723,0.8455987],"study_design_scores_gemma":[0.00036323786,0.00060742145,0.029302014,0.0018491498,0.0024725276,0.008845833,0.00867554,0.24691892,0.117809534,0.25668997,0.32616147,0.00030446457],"about_ca_topic_score_codex":0.0043747327,"about_ca_topic_score_gemma":0.0069154473,"teacher_disagreement_score":0.010751344,"about_ca_system_score_codex":0.0011288332,"about_ca_system_score_gemma":0.00308841,"threshold_uncertainty_score":0.017183542},"labels":[],"label_agreement":null},{"id":"W2073874433","doi":"10.1016/j.datak.2008.06.007","title":"Data privacy protection in multi-party clustering","year":2008,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"McMaster University","keywords":"Cluster analysis; Privacy protection; Internet privacy; Computer security; Data Protection Act 1998; Information privacy; Computer science; Business; Artificial intelligence","score_opus":0.1658117584570718,"score_gpt":0.31546151492775104,"score_spread":0.14964975647067924,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2073874433","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.036084294,0.00043772167,0.95788884,0.0016364446,0.000064595544,0.00008893322,0.00022858908,0.00029596357,0.0032746],"genre_scores_gemma":[0.8778353,0.0003266303,0.117422074,0.00034741344,0.0001137024,0.00013454958,0.00029435116,0.000088570945,0.0034374027],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98178583,0.007947638,0.0011125221,0.0025293638,0.0053633335,0.00126118],"domain_scores_gemma":[0.9476671,0.01914139,0.0028213905,0.027375612,0.0023382464,0.00065613113],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011195427,0.00053439557,0.001963206,0.0014658679,0.0036368798,0.005339757,0.0040305634,0.0037332312,0.0019370312],"category_scores_gemma":[0.040458873,0.0011032053,0.0018749306,0.0034459238,0.0035102575,0.009960263,0.008182185,0.004365894,0.0007699479],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012522526,0.00026859157,0.005275837,0.00036321866,0.00032697534,0.00063375913,0.0017911852,0.19445674,0.0103314845,0.6305198,0.0054558557,0.14932431],"study_design_scores_gemma":[0.000052421157,0.000077689205,0.000769203,0.00005707192,0.00007357935,0.00071354426,0.0003670599,0.5060968,0.015179058,0.47172123,0.0048467205,0.000045582965],"about_ca_topic_score_codex":0.0008215068,"about_ca_topic_score_gemma":0.0005722588,"teacher_disagreement_score":0.011195427,"about_ca_system_score_codex":0.0023071559,"about_ca_system_score_gemma":0.0023949633,"threshold_uncertainty_score":0.059207737},"labels":[],"label_agreement":null},{"id":"W2079690139","doi":"10.1016/j.datak.2006.11.002","title":"Load balancing and data placement for multi-tiered database systems","year":2006,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"IBM (Canada)","funders":"","keywords":"Computer science; Database; Materialized view; View; Database server; Workload; Online analytical processing; Database tuning; Server; Table (database); IBM; Database administrator; Distributed database; Database testing; Operating system; Application server; Load balancing (electrical power); Data warehouse; Database design","score_opus":0.057066989033939385,"score_gpt":0.3019583151155923,"score_spread":0.2448913260816529,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2079690139","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.22471386,0.0031204126,0.75596535,0.0015914012,0.0005190193,0.00025754134,0.0007640208,0.0060121026,0.00705633],"genre_scores_gemma":[0.8202351,0.0003839562,0.17345151,0.00018902942,0.00017669237,0.00009128208,0.0005671654,0.00031101122,0.004594261],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9973677,0.00065082253,0.00033592785,0.00040128033,0.0007551654,0.00048910483],"domain_scores_gemma":[0.99510324,0.0013307884,0.0003896233,0.0016500568,0.001040132,0.00048623624],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002174549,0.0008293857,0.0011339383,0.0012545185,0.002088974,0.0033514285,0.003729957,0.0010892267,0.0050645326],"category_scores_gemma":[0.007683186,0.0009684344,0.00044479952,0.003204073,0.0007915994,0.004922389,0.0024643089,0.0009357112,0.00088389695],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0033954685,0.00081460166,0.011480995,0.00050842855,0.00022672398,0.00045984675,0.00089061534,0.364609,0.052430127,0.027949829,0.03248292,0.5047514],"study_design_scores_gemma":[0.000103906095,0.00017617753,0.0021178958,0.000014929002,0.000069481255,0.00022324472,0.0003038031,0.962184,0.010323278,0.02082719,0.003613489,0.000042582542],"about_ca_topic_score_codex":0.0070429193,"about_ca_topic_score_gemma":0.014078478,"teacher_disagreement_score":0.0070429193,"about_ca_system_score_codex":0.0015766773,"about_ca_system_score_gemma":0.002043777,"threshold_uncertainty_score":0.01694262},"labels":[],"label_agreement":null},{"id":"W2079796778","doi":"10.1016/j.datak.2013.03.005","title":"Subject-based semantic document clustering for digital forensic investigations","year":2013,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Digital and Cyber Forensics","field":"Computer Science","cited_by":26,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Subject (documents); Computer science; Cluster analysis; Information retrieval; Document clustering; Natural language processing; Artificial intelligence; Data science; World Wide Web","score_opus":0.023005084943141487,"score_gpt":0.2321864369908625,"score_spread":0.209181352047721,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2079796778","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.13257721,0.002047897,0.8477777,0.0003175107,0.00023324993,0.00043626432,0.0043273163,0.0063859792,0.0058968575],"genre_scores_gemma":[0.42892355,0.0010370502,0.5560168,0.000083769584,0.0001452619,0.00022091277,0.009022624,0.000303366,0.0042465967],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99908495,0.00013804117,0.00011064229,0.00023603418,0.00032791108,0.00010243079],"domain_scores_gemma":[0.9987091,0.00022195795,0.00013303774,0.00031068048,0.0005376413,0.00008769842],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008448762,0.00058071484,0.00078636355,0.008139671,0.0013402184,0.0020468098,0.0009276006,0.0010518834,0.0016123678],"category_scores_gemma":[0.0024471716,0.0002290328,0.0011526275,0.006337168,0.0005721767,0.0017898508,0.001094313,0.0005333201,0.0016276737],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009586053,0.00055386353,0.011662283,0.00055208476,0.00020889728,0.00028759034,0.0007223309,0.023655852,0.06290718,0.016566463,0.013161776,0.8687631],"study_design_scores_gemma":[0.000093339324,0.0004197227,0.020967925,0.00019267638,0.00046678056,0.0012826774,0.0020386765,0.81207234,0.06550688,0.056180585,0.04064301,0.00013543142],"about_ca_topic_score_codex":0.006236954,"about_ca_topic_score_gemma":0.011103615,"teacher_disagreement_score":0.008139671,"about_ca_system_score_codex":0.00078625424,"about_ca_system_score_gemma":0.0023800873,"threshold_uncertainty_score":0.012401283},"labels":[],"label_agreement":null},{"id":"W2081024207","doi":"10.1016/j.datak.2007.05.006","title":"A tree-projection-based algorithm for multi-label recurrent-item associative-classification rule generation","year":2007,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Association rule learning; Associative property; Computer science; Data mining; Apriori algorithm; Artificial intelligence; Classifier (UML); Pattern recognition (psychology); Machine learning; Projection (relational algebra); Algorithm; Mathematics","score_opus":0.15268889256032656,"score_gpt":0.35595682947668167,"score_spread":0.2032679369163551,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2081024207","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.003464374,0.00018940135,0.9913722,0.00011511199,0.00006046235,0.00023038441,0.00020942761,0.0037585795,0.000600033],"genre_scores_gemma":[0.029622817,0.00010245468,0.9674339,0.00012718096,0.00003891646,0.00037235022,0.00068611366,0.00013933118,0.0014769164],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983784,0.00034891415,0.00017026998,0.0004370304,0.0005524334,0.000112820126],"domain_scores_gemma":[0.9967759,0.0015805507,0.000104758896,0.00033345117,0.0010901885,0.000115063696],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0023487625,0.0013643703,0.0024497234,0.0027624392,0.0012894174,0.001711335,0.0033288288,0.0019954,0.009301401],"category_scores_gemma":[0.0062562414,0.000819456,0.0012310112,0.0036500876,0.0007325072,0.0019654506,0.0021385867,0.0028420545,0.0046919812],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00015034695,0.0002081701,0.0005802533,0.000106909254,0.00009872235,0.00008923143,0.000078656696,0.014125084,0.0036020454,0.0033017774,0.00700027,0.9706585],"study_design_scores_gemma":[0.00010973221,0.00012458408,0.0005454697,0.000039356266,0.00008862577,0.00031061613,0.000058200523,0.97463834,0.007766826,0.012102611,0.004176777,0.00003882274],"about_ca_topic_score_codex":0.007556651,"about_ca_topic_score_gemma":0.012727309,"teacher_disagreement_score":0.009301401,"about_ca_system_score_codex":0.0007422748,"about_ca_system_score_gemma":0.0030993226,"threshold_uncertainty_score":0.031116247},"labels":[{"model":"gemma","categories":[],"domain":null,"study_design":"simulation_or_modeling","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2082651521","doi":"10.1016/j.datak.2006.03.005","title":"Compact access control labeling for efficient secure XML query evaluation","year":2006,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Access Control and Trust","field":"Social Sciences","cited_by":22,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; XML Signature; Access control; XML Encryption; XML database; Efficient XML Interchange; Streaming XML; XML framework; Granularity; Data access; Database; XML validation; XML; Computer network; World Wide Web; Programming language","score_opus":0.06344754137285828,"score_gpt":0.3806975017838266,"score_spread":0.3172499604109683,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2082651521","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.034848154,0.00017234165,0.956647,0.00041648472,0.000046384197,0.0002806105,0.00047385236,0.005571128,0.0015439894],"genre_scores_gemma":[0.5743365,0.00010898762,0.42040348,0.00022247556,0.00008071128,0.00031036578,0.0010774526,0.00058737665,0.002872645],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98981184,0.0029941064,0.0011083467,0.0011835771,0.003949195,0.0009529585],"domain_scores_gemma":[0.9732084,0.012196381,0.001521747,0.009185916,0.00340661,0.00048096367],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0057006315,0.00071816903,0.0019056539,0.0023539248,0.00173452,0.0050421217,0.0016597828,0.0014339909,0.0047055953],"category_scores_gemma":[0.023081224,0.00084668555,0.0010718916,0.0021229086,0.0021522043,0.009105924,0.003999448,0.0025334407,0.0011331587],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025553883,0.00055332226,0.004541896,0.00041295556,0.00012253293,0.0003644167,0.0013895722,0.06886094,0.04773636,0.28893778,0.014321782,0.570203],"study_design_scores_gemma":[0.00012436393,0.0001796859,0.00060664903,0.000046350928,0.0000742618,0.0002192751,0.00024539686,0.74848974,0.04406741,0.20029886,0.0055855536,0.00006239448],"about_ca_topic_score_codex":0.0048142257,"about_ca_topic_score_gemma":0.0053323316,"teacher_disagreement_score":0.0057006315,"about_ca_system_score_codex":0.0034232847,"about_ca_system_score_gemma":0.0038676048,"threshold_uncertainty_score":0.030148208},"labels":[],"label_agreement":null},{"id":"W2090992497","doi":"10.1016/j.datak.2004.06.005","title":"eMarketplaces for enterprise and cross enterprise integration","year":2004,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Auction Theory and Applications","field":"Decision Sciences","cited_by":23,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"National Research Council Canada; Western University","funders":"","keywords":"Enterprise integration; Business; Process management; Enterprise software","score_opus":0.08516768738760731,"score_gpt":0.4081140141277604,"score_spread":0.32294632674015306,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2090992497","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.053782836,0.000805264,0.9002603,0.0031040676,0.0002832355,0.00018852536,0.00023016214,0.0010726817,0.040272843],"genre_scores_gemma":[0.68145597,0.00050514017,0.29291862,0.00033193125,0.0001842287,0.00023114002,0.00038362006,0.00016379406,0.02382548],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9966724,0.0014314086,0.00023491694,0.0005011425,0.0007444678,0.00041571044],"domain_scores_gemma":[0.99346393,0.0025586118,0.00057161547,0.0020449848,0.0007551119,0.00060565624],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006712499,0.00047290605,0.0009978034,0.0014836018,0.0024196303,0.006630281,0.0031051284,0.003641261,0.024261417],"category_scores_gemma":[0.014515471,0.00066571403,0.0011580129,0.0025085313,0.0021403967,0.017153613,0.0068376483,0.0031385797,0.0023161077],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018784094,0.00019080575,0.00068190007,0.000049307488,0.000020415244,0.00008820323,0.0001249404,0.008280039,0.00041688606,0.923918,0.0025188606,0.06352283],"study_design_scores_gemma":[0.00007321477,0.000083740524,0.0002446525,0.000048936057,0.000022383587,0.00017044401,0.00024907538,0.08041192,0.0011082075,0.90278685,0.014773669,0.00002687204],"about_ca_topic_score_codex":0.0014680498,"about_ca_topic_score_gemma":0.0018011335,"teacher_disagreement_score":0.024261417,"about_ca_system_score_codex":0.0014591778,"about_ca_system_score_gemma":0.0019449495,"threshold_uncertainty_score":0.08116251},"labels":[],"label_agreement":null},{"id":"W2094060849","doi":"10.1016/j.datak.2010.03.007","title":"Ranking bias in deep web size estimation using capture recapture method","year":2010,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada; State Key Laboratory of Novel Software Technology","keywords":"Ranking (information retrieval); Computer science; Estimation; Sampling (signal processing); Data mining; Process (computing); Matching (statistics); Mark and recapture; Rank (graph theory); Limit (mathematics); Statistics; Information retrieval; Mathematics","score_opus":0.04137654040511619,"score_gpt":0.3090537414834847,"score_spread":0.2676772010783685,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2094060849","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.12112279,0.00043932727,0.87661743,0.00018591573,0.000058894846,0.00003705051,0.0001817816,0.0005524368,0.000804336],"genre_scores_gemma":[0.8827783,0.00019595302,0.11281955,0.00023021144,0.00011076278,0.00010232664,0.0006749718,0.00015513922,0.002932728],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9954774,0.0023979014,0.00016951055,0.0009810714,0.00062389224,0.00035040063],"domain_scores_gemma":[0.93984675,0.049281977,0.0029360321,0.0054532024,0.0020614092,0.00042060268],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.016332176,0.0006481474,0.0013342382,0.0026301013,0.0007127601,0.0013767312,0.0034134812,0.0017189778,0.0013951575],"category_scores_gemma":[0.046589427,0.0008022934,0.0011061846,0.0019508564,0.0011135978,0.002739721,0.0014044784,0.001309363,0.00046845264],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00049397536,0.0002090281,0.15134063,0.0003782659,0.0013050429,0.0005498679,0.00038657465,0.4513527,0.0070024063,0.06371681,0.006325785,0.3169389],"study_design_scores_gemma":[0.000016742431,0.00005036124,0.0109271,0.000018540773,0.00009993631,0.00012445166,0.000035843816,0.9651141,0.0026106953,0.020254418,0.00071280106,0.00003503654],"about_ca_topic_score_codex":0.004529726,"about_ca_topic_score_gemma":0.0051086154,"teacher_disagreement_score":0.016332176,"about_ca_system_score_codex":0.00093922764,"about_ca_system_score_gemma":0.0006469178,"threshold_uncertainty_score":0.086373866},"labels":[],"label_agreement":null},{"id":"W2114928940","doi":"10.1016/j.datak.2013.02.001","title":"Assessing the quality factors found in in-line documentation written in natural language: The JavadocMiner","year":2013,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Software Engineering Research","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"U.S. Department of Defense","keywords":"Documentation; Natural (archaeology); Quality (philosophy); Line (geometry); Natural language processing; Computer science; Linguistics; Psychology; History; Mathematics; Programming language; Archaeology; Philosophy","score_opus":0.059091997112096124,"score_gpt":0.39128965147035494,"score_spread":0.3321976543582588,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2114928940","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.98226905,0.00027697804,0.012053192,0.00014868383,0.000016325901,0.00006660742,0.0011350025,0.0028184643,0.0012156296],"genre_scores_gemma":[0.91683245,0.00022007029,0.074730635,0.00007109977,0.000014660435,0.00007035029,0.0053570233,0.0010028703,0.0017008967],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9932473,0.0017380164,0.0012451132,0.00086395716,0.002699762,0.0002058676],"domain_scores_gemma":[0.8842165,0.0609452,0.018629316,0.010188616,0.024601066,0.0014192596],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008280525,0.00045649408,0.0005453289,0.0036784825,0.0004767217,0.0021369196,0.00080319954,0.00086806825,0.0009938509],"category_scores_gemma":[0.06522698,0.0003589376,0.00037473039,0.0026500318,0.00036615558,0.0016488519,0.0011349887,0.0006043764,0.0005257657],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0023562724,0.001920891,0.3902852,0.0025673616,0.0004928734,0.00082413305,0.0065176887,0.0063798446,0.073204316,0.0010415744,0.008980397,0.50542957],"study_design_scores_gemma":[0.00062999857,0.0038094698,0.6573953,0.00084212964,0.0010794943,0.0026403018,0.0052062497,0.12965006,0.1717581,0.0019633737,0.024640746,0.00038470884],"about_ca_topic_score_codex":0.0028108067,"about_ca_topic_score_gemma":0.0049648476,"teacher_disagreement_score":0.008280525,"about_ca_system_score_codex":0.00059099926,"about_ca_system_score_gemma":0.0015801025,"threshold_uncertainty_score":0.04379213},"labels":[],"label_agreement":null},{"id":"W2134227416","doi":"10.1016/s0169-023x(00)00044-6","title":"Selecting and materializing horizontally partitioned warehouse views","year":2001,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Data warehouse; Cube (algebra); Aggregate (composite); Computer science; Table (database); Warehouse; Dimension (graph theory); Data cube; Materialized view; Dimensional modeling; Database; Constraint (computer-aided design); Online analytical processing; Data mining; Information retrieval; Mathematics; View; Geography; Database design; Combinatorics","score_opus":0.057702484360440495,"score_gpt":0.2827958738468403,"score_spread":0.2250933894863998,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2134227416","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.312275,0.0020895377,0.6432935,0.0011905372,0.000335059,0.0011892748,0.0033481247,0.02242979,0.013849152],"genre_scores_gemma":[0.26880255,0.00087417866,0.7168713,0.00019622434,0.0000997905,0.00020177834,0.0065871687,0.0020818207,0.004285245],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99796176,0.00026248064,0.00018182422,0.00031964388,0.0009266374,0.0003477612],"domain_scores_gemma":[0.99611974,0.0016367408,0.0002374127,0.00084203837,0.00086793024,0.0002962083],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017548007,0.0012782387,0.0017967777,0.00345595,0.0012724354,0.006027469,0.002075428,0.0012466861,0.0076176473],"category_scores_gemma":[0.0091049345,0.0016518995,0.0021176564,0.0050217384,0.0008252127,0.005474849,0.0037061505,0.001475121,0.001513354],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0022181154,0.0005762768,0.013838572,0.0009454864,0.00040971139,0.0017341713,0.0025088869,0.021190401,0.12503438,0.015042483,0.038500264,0.77800125],"study_design_scores_gemma":[0.00061411643,0.0015476609,0.017663332,0.0005439307,0.0013750866,0.002125824,0.011528413,0.5587093,0.27050745,0.060375642,0.07462111,0.00038824405],"about_ca_topic_score_codex":0.0038177967,"about_ca_topic_score_gemma":0.007632033,"teacher_disagreement_score":0.0076176473,"about_ca_system_score_codex":0.0009373035,"about_ca_system_score_gemma":0.0018467578,"threshold_uncertainty_score":0.025483608},"labels":[],"label_agreement":null},{"id":"W2144326136","doi":"10.1016/j.datak.2008.12.001","title":"Privacy-preserving data publishing for cluster analysis","year":2008,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":71,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ontario Institute of Technology; Simon Fraser University; Concordia University","funders":"","keywords":"Masking (illustration); Data publishing; Computer science; Cluster (spacecraft); ENCODE; Data mining; Process (computing); Publishing; Focus (optics); Data anonymization; Class (philosophy); Data quality; Information retrieval; Information privacy; Artificial intelligence; Computer security; Engineering","score_opus":0.10371521504469686,"score_gpt":0.30616418974752596,"score_spread":0.2024489747028291,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2144326136","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.017657412,0.0004554803,0.97689897,0.0011348918,0.0001432694,0.0001020426,0.0006301852,0.0008840122,0.0020938022],"genre_scores_gemma":[0.6327802,0.0009427636,0.35541457,0.00057490607,0.00047935304,0.00030637966,0.0019633716,0.00038654573,0.0071518985],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9911855,0.002892425,0.0008234037,0.0013122966,0.0031092213,0.0006770661],"domain_scores_gemma":[0.9556216,0.010762128,0.0016903663,0.02847267,0.002801351,0.0006518385],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006926941,0.00052633806,0.0018545708,0.0017924886,0.002728708,0.0050569265,0.0030762586,0.001532212,0.002324322],"category_scores_gemma":[0.03288099,0.0008451137,0.0015952702,0.0053322734,0.002297281,0.007467228,0.0050823535,0.0031825835,0.0013055023],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0013384564,0.00036395577,0.006240323,0.00041257605,0.00035592774,0.00051070383,0.0013138123,0.100911066,0.013056225,0.5428522,0.017808944,0.31483585],"study_design_scores_gemma":[0.00008015711,0.00012509411,0.0007616991,0.000052213967,0.00011865005,0.00083783024,0.00036444116,0.33802596,0.03172371,0.61124367,0.016613524,0.00005301265],"about_ca_topic_score_codex":0.001013557,"about_ca_topic_score_gemma":0.0010511065,"teacher_disagreement_score":0.006926941,"about_ca_system_score_codex":0.0015037089,"about_ca_system_score_gemma":0.0038118965,"threshold_uncertainty_score":0.03663361},"labels":[],"label_agreement":null},{"id":"W2145399176","doi":"10.1016/j.datak.2006.06.001","title":"Combined mining of Web server logs and web contents for classifying user navigation patterns and predicting users’ future requests","year":2006,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":205,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"","keywords":"Web mining; Computer science; Web page; Web server; Web navigation; Personalization; Static web page; World Wide Web; Web log analysis software; Information retrieval; Data Web; Web API; Database; Data mining; The Internet","score_opus":0.030135819550464835,"score_gpt":0.25834862715059415,"score_spread":0.2282128076001293,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2145399176","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.75301707,0.0015093941,0.2298942,0.00040868475,0.00012137897,0.00031483415,0.006520971,0.0049702465,0.003243297],"genre_scores_gemma":[0.87398857,0.00046815752,0.11667613,0.000080752274,0.00009493824,0.00015917794,0.0064684153,0.00008656913,0.0019772842],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99871397,0.00030391396,0.00015137705,0.00025657666,0.00048282143,0.000091449634],"domain_scores_gemma":[0.99511045,0.0024618707,0.00044454727,0.0006334841,0.0011205758,0.00022912268],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000991649,0.00079310464,0.0015253208,0.006811072,0.00056969473,0.0013805125,0.0007824204,0.0009524498,0.000631808],"category_scores_gemma":[0.0047638486,0.00040735252,0.0012498052,0.005124612,0.0001676625,0.0015075096,0.000465096,0.0008033989,0.0009312504],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007183869,0.0028889736,0.31366864,0.0005949721,0.0011345743,0.0003856215,0.0006213582,0.015365554,0.046803407,0.00073210883,0.005879508,0.6112069],"study_design_scores_gemma":[0.0000555029,0.00063178485,0.15824057,0.000070887676,0.0007350749,0.0007850746,0.0004334004,0.8120016,0.020951962,0.0024071638,0.0035957375,0.00009131998],"about_ca_topic_score_codex":0.009285218,"about_ca_topic_score_gemma":0.027915977,"teacher_disagreement_score":0.009285218,"about_ca_system_score_codex":0.00033585887,"about_ca_system_score_gemma":0.00094808417,"threshold_uncertainty_score":0.01846236},"labels":[],"label_agreement":null},{"id":"W3037550867","doi":"10.1016/j.datak.2013.01.001","title":"Diamond dicing","year":2013,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université TÉLUQ; University of New Brunswick","funders":"","keywords":"Wafer dicing; Materials science; Nanotechnology; Wafer","score_opus":0.030581411111876202,"score_gpt":0.24660457927354787,"score_spread":0.21602316816167166,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3037550867","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009290629,0.0015022578,0.19704567,0.0026010347,0.005889241,0.0004514878,0.0025633178,0.0074154837,0.77324086],"genre_scores_gemma":[0.12299487,0.0017167914,0.29232666,0.002030158,0.0008671508,0.00040565035,0.005145384,0.003731426,0.57078195],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9988838,0.00006633622,0.000054007556,0.00021962398,0.00067019963,0.00010598598],"domain_scores_gemma":[0.9986926,0.00017315696,0.00004374784,0.00042160874,0.00055084884,0.00011808302],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0006719115,0.00084911095,0.0008076594,0.0029982275,0.0020398253,0.0021260483,0.001552524,0.0013744248,0.20764938],"category_scores_gemma":[0.0025494013,0.00053173036,0.0007107699,0.001831517,0.0008522142,0.0017469758,0.0025143141,0.002190732,0.05262961],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00027884232,0.00006525256,0.0005593763,0.00043760665,0.000025720625,0.0005022361,0.0002091069,0.002620682,0.015561314,0.1736902,0.31908306,0.48696664],"study_design_scores_gemma":[0.000042517353,0.000058465328,0.00030320985,0.00007672168,0.000013328223,0.00067018456,0.00012596334,0.0045292736,0.011158597,0.02757488,0.9554117,0.000035184108],"about_ca_topic_score_codex":0.0015458968,"about_ca_topic_score_gemma":0.002550814,"teacher_disagreement_score":0.79235065,"about_ca_system_score_codex":0.0007051152,"about_ca_system_score_gemma":0.0013349742,"threshold_uncertainty_score":0.6946564},"labels":[],"label_agreement":null},{"id":"W3131740359","doi":"10.1016/j.datak.2021.101876","title":"Generation of Gaussian sets for clustering methods assessment","year":2021,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Advanced Clustering Algorithms Research","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Thales (Canada)","funders":"","keywords":"Cluster analysis; Computer science; Generator (circuit theory); Fuzzy clustering; Curse of dimensionality; Rand index; Mixture model; Data mining; Homogeneity (statistics); Artificial intelligence; Gaussian; Pattern recognition (psychology); Set (abstract data type); Hierarchical clustering; Maximization; Sensitivity (control systems); Algorithm; Machine learning; Mathematics; Mathematical optimization; Power (physics); Engineering","score_opus":0.1936835263463012,"score_gpt":0.47320929907024467,"score_spread":0.27952577272394347,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3131740359","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.016625797,0.0001336606,0.9788477,0.00014305922,0.000080497215,0.0004186926,0.00037643992,0.0015848668,0.0017893169],"genre_scores_gemma":[0.17255045,0.00014815663,0.82220834,0.00010005046,0.000035749392,0.0007899234,0.0017777812,0.000709928,0.0016794789],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9928108,0.0027005489,0.00040443847,0.00087575923,0.0027693128,0.00043914656],"domain_scores_gemma":[0.9839677,0.0058430005,0.00062133174,0.0023440474,0.0068748957,0.00034909305],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009924402,0.0015130701,0.0012830726,0.0052485643,0.0019547737,0.0029163668,0.002567967,0.0021444652,0.005694978],"category_scores_gemma":[0.040438406,0.0007541541,0.0024856238,0.0029356705,0.0010445595,0.0019335966,0.0039442056,0.002505636,0.0020064723],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000797116,0.00047841354,0.008731701,0.0006865494,0.00037454863,0.00026348693,0.0013956546,0.22840476,0.01591539,0.1112095,0.018864147,0.61287886],"study_design_scores_gemma":[0.00007129453,0.00018278747,0.0022630072,0.00012661524,0.00007467827,0.00013980227,0.00022280274,0.9240623,0.015146071,0.04927241,0.008371458,0.0000666709],"about_ca_topic_score_codex":0.005646474,"about_ca_topic_score_gemma":0.0058634263,"teacher_disagreement_score":0.009924402,"about_ca_system_score_codex":0.0021200378,"about_ca_system_score_gemma":0.0037608177,"threshold_uncertainty_score":0.052485883},"labels":[],"label_agreement":null},{"id":"W3197885731","doi":"10.1016/j.datak.2021.101926","title":"Exploring the research landscape of data warehousing and mining based on DaWaK Conference full-text articles","year":2021,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Big Data and Business Intelligence","field":"Business, Management and Accounting","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Ministry of Science and ICT, South Korea; National Research Foundation of Korea; Conselho Nacional de Desenvolvimento Científico e Tecnológico; National Natural Science Foundation of China; National Science Foundation","keywords":"Data science; Data warehouse; Computer science; Information retrieval; Data mining","score_opus":0.5264044121500385,"score_gpt":0.3663522096925166,"score_spread":0.16005220245752194,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3197885731","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07899578,0.20970838,0.2383525,0.14869058,0.00523655,0.00048680842,0.0020032905,0.0026642082,0.31386194],"genre_scores_gemma":[0.4027365,0.19210742,0.3458502,0.007938489,0.0027398276,0.00034813522,0.0024961252,0.0011308371,0.044652473],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99391574,0.0024187982,0.000497245,0.0007255723,0.0019926988,0.00044998367],"domain_scores_gemma":[0.9667066,0.01976,0.0011844803,0.0026789687,0.0076930975,0.0019768684],"candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.017628793,0.00061165355,0.0010050151,0.014381166,0.0028418556,0.025635775,0.0018732453,0.0013378017,0.009752742],"category_scores_gemma":[0.019206785,0.0007649448,0.00073162635,0.019766677,0.004490066,0.023482598,0.0047519556,0.0032175898,0.0025195132],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00022937759,0.00022684151,0.0048830044,0.002399714,0.000071709896,0.00024263706,0.0018861861,0.0022572528,0.0036510644,0.4844913,0.028351167,0.47130975],"study_design_scores_gemma":[0.000039741488,0.00016108295,0.005182808,0.0030456942,0.00007520243,0.0004528134,0.009362879,0.016977275,0.0066056447,0.3334544,0.6244963,0.0001461961],"about_ca_topic_score_codex":0.004213723,"about_ca_topic_score_gemma":0.0066683376,"teacher_disagreement_score":0.9856188,"about_ca_system_score_codex":0.0043866206,"about_ca_system_score_gemma":0.0087915575,"threshold_uncertainty_score":0.09323108},"labels":[],"label_agreement":null},{"id":"W3197889602","doi":"10.1016/j.datak.2021.101924","title":"Mining high utility patterns in interval-based event sequences","year":2021,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Regina","funders":"","keywords":"Pruning; Interval (graph theory); Computer science; Event (particle physics); Data mining; Range (aeronautics); Point (geometry); Sequential Pattern Mining; Algorithm; Mathematics; Engineering","score_opus":0.041472155210488644,"score_gpt":0.29062862660563255,"score_spread":0.2491564713951439,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3197889602","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.55986434,0.0014130869,0.43046483,0.00041700792,0.000104078186,0.00020642643,0.0040340764,0.0011168288,0.0023793085],"genre_scores_gemma":[0.90741086,0.00045303477,0.088227205,0.000049946797,0.00007891272,0.000077922436,0.0027999254,0.000050159455,0.0008520084],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9986712,0.000177996,0.00016648191,0.00027628362,0.0005861606,0.00012190303],"domain_scores_gemma":[0.99434376,0.00329743,0.00096351746,0.0004074449,0.00073136133,0.0002565007],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009808433,0.0005776577,0.0005798145,0.003808027,0.0004176025,0.0012057895,0.00087601726,0.0005731884,0.0014996456],"category_scores_gemma":[0.008804132,0.00027448582,0.0005805909,0.003324162,0.00036714357,0.0016416368,0.0006820198,0.00078604853,0.00046777082],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0027874,0.0010915126,0.21862954,0.0013438554,0.00054781896,0.005843346,0.001157056,0.10110104,0.036640137,0.029247327,0.0056080925,0.5960029],"study_design_scores_gemma":[0.00013726343,0.0008414179,0.055689126,0.00026087646,0.00034550962,0.0040095816,0.00087118195,0.822716,0.022714559,0.08441624,0.007916087,0.000082157574],"about_ca_topic_score_codex":0.0011731973,"about_ca_topic_score_gemma":0.0016930784,"teacher_disagreement_score":0.003808027,"about_ca_system_score_codex":0.00031300948,"about_ca_system_score_gemma":0.00052421645,"threshold_uncertainty_score":0.0051872134},"labels":[],"label_agreement":null},{"id":"W4327724188","doi":"10.1016/j.datak.2023.102177","title":"Preface","year":2023,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Geography","score_opus":0.1044293387069667,"score_gpt":0.33191004888109904,"score_spread":0.22748071017413235,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4327724188","genre_codex":"other","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0013534053,0.007532759,0.022442017,0.037600525,0.2295315,0.0011648932,0.026759336,0.0038468451,0.6697687],"genre_scores_gemma":[0.004360482,0.003530684,0.0063620964,0.0063259676,0.02492838,0.0004107586,0.015071882,0.0016946329,0.93731505],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99904853,0.0001399642,0.00006236473,0.0001400148,0.00054150174,0.00006753358],"domain_scores_gemma":[0.98793566,0.0018991041,0.0003285655,0.0011020602,0.007667259,0.0010673316],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0020340628,0.00097559305,0.00064736005,0.0039960304,0.002413294,0.003709936,0.0011843318,0.00084431615,0.4839282],"category_scores_gemma":[0.019719787,0.00033813008,0.0005571226,0.0027608816,0.0005981335,0.002957699,0.0021597194,0.0029906984,0.35254207],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000016037426,0.000012796142,0.000059909577,0.00006301624,0.000001128537,0.000019625326,0.000031989668,0.000047829853,0.000090736634,0.0032926898,0.9652375,0.031126643],"study_design_scores_gemma":[0.0000033167858,0.000008015373,0.00016293509,0.000095490635,0.0000014935047,0.000025493073,0.000049459166,0.000034628938,0.00010032742,0.0030292736,0.9964851,0.000004401165],"about_ca_topic_score_codex":0.0057926243,"about_ca_topic_score_gemma":0.0065365983,"teacher_disagreement_score":0.5160718,"about_ca_system_score_codex":0.0018567124,"about_ca_system_score_gemma":0.0025059923,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W4386987509","doi":"10.1016/j.datak.2023.102229","title":"Robotic process automation using process mining — A systematic literature review","year":2023,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Robotic Process Automation Applications","field":"Engineering","cited_by":20,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada; University of Ottawa","keywords":"Computer science; Automation; Process (computing); Preprocessor; Process mining; Event (particle physics); Software engineering; Domain (mathematical analysis); Software; Data science; Construct (python library); Robot; Data mining; Work in process; Business process; Artificial intelligence; Engineering; Business process modeling","score_opus":0.036817580297915056,"score_gpt":0.3163027075077396,"score_spread":0.27948512720982455,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4386987509","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.001699176,0.99510765,0.0012500026,0.00033672142,0.0001155184,0.00054624176,0.00046392233,0.000017841361,0.0004628584],"genre_scores_gemma":[0.015006638,0.97949135,0.0036061604,0.00057543686,0.00008393114,0.000650743,0.00042794718,0.000012208654,0.00014554332],"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","domain_scores_codex":[0.99287236,0.0018142727,0.0030225727,0.0006874543,0.0014199693,0.00018335739],"domain_scores_gemma":[0.9607456,0.02831145,0.0058844457,0.0009363991,0.0037366091,0.00038543745],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01232549,0.0017779574,0.0057245987,0.015539649,0.0009765257,0.0035652642,0.002547091,0.0020166298,0.0028960772],"category_scores_gemma":[0.043320198,0.0010425826,0.006300724,0.012217221,0.0012102718,0.003776319,0.0023056814,0.0013714567,0.00045245924],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00020768533,0.00010783547,0.0018195753,0.8268255,0.0057226685,0.00022683259,0.00049486145,0.00038019332,0.00045088597,0.0007914446,0.00260258,0.16036992],"study_design_scores_gemma":[0.00019315138,0.0003379876,0.0042781946,0.89893085,0.047430072,0.000796601,0.00079679204,0.00038885223,0.0006374351,0.0012987077,0.04481803,0.0000932314],"about_ca_topic_score_codex":0.0050258785,"about_ca_topic_score_gemma":0.015495077,"teacher_disagreement_score":0.015539649,"about_ca_system_score_codex":0.0022525426,"about_ca_system_score_gemma":0.017831344,"threshold_uncertainty_score":0.065184236},"labels":[],"label_agreement":null},{"id":"W4387842356","doi":"10.1016/j.datak.2023.102238","title":"User-generated short-text classification using cograph editing-based network clustering with an application in invoice categorization","year":2023,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McMaster University","funders":"Ontario Ministry of Research and Innovation; Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada; Mitacs; Canada Foundation for Innovation","keywords":"Computer science; Categorization; Cluster analysis; Invoice; Information retrieval; Product (mathematics); Natural language processing; Data mining; Artificial intelligence; World Wide Web","score_opus":0.04509888621521016,"score_gpt":0.2994695044326596,"score_spread":0.25437061821744944,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4387842356","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.23800173,0.0008916503,0.50752,0.000854451,0.001035988,0.0010707848,0.054063104,0.18145958,0.015102697],"genre_scores_gemma":[0.31692824,0.0002639541,0.6100978,0.00014513337,0.00016250525,0.0005513955,0.05507698,0.0034191592,0.013354925],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99912494,0.00014854183,0.00007511684,0.0003069835,0.00027987547,0.00006449297],"domain_scores_gemma":[0.99654645,0.0015548904,0.00017877581,0.0005191553,0.001039292,0.00016151187],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009462829,0.0010135826,0.0006067774,0.0059462986,0.0010315778,0.0012313208,0.0010993175,0.0009140779,0.008767792],"category_scores_gemma":[0.005891817,0.00026044357,0.00065841904,0.0040371264,0.00029299493,0.0015169386,0.0010151187,0.0006202015,0.004681647],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0013024902,0.0006644445,0.012130493,0.0011567244,0.00018131455,0.0009288308,0.0014085581,0.02093931,0.036885876,0.00630237,0.11310949,0.80499005],"study_design_scores_gemma":[0.00011169899,0.00017612817,0.012508776,0.00009296265,0.0001058004,0.00049114623,0.0008016844,0.856596,0.043264244,0.013475572,0.072243854,0.00013207908],"about_ca_topic_score_codex":0.006521841,"about_ca_topic_score_gemma":0.01584067,"teacher_disagreement_score":0.008767792,"about_ca_system_score_codex":0.00072052586,"about_ca_system_score_gemma":0.0009890049,"threshold_uncertainty_score":0.029331148},"labels":[],"label_agreement":null},{"id":"W4388090033","doi":"10.1016/j.datak.2023.102240","title":"Hierarchical framework for interpretable and specialized deep reinforcement learning-based predictive maintenance","year":2023,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Reliability and Maintenance Optimization","field":"Engineering","cited_by":22,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"H2020 Marie Skłodowska-Curie Actions; Canadian Institute of Steel Construction; Horizon 2020 Framework Programme; Österreichische Forschungsförderungsgesellschaft; Science Foundation Ireland; Horizon 2020","keywords":"Interpretability; Reinforcement learning; Machine learning; Computer science; Artificial intelligence; Probabilistic logic; Black box; Context (archaeology); Markov decision process; Risk analysis (engineering); Markov process","score_opus":0.015179626367081898,"score_gpt":0.25194097941082105,"score_spread":0.23676135304373916,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4388090033","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010300418,0.0003218107,0.98549485,0.00016712604,0.000042300875,0.000030973137,0.00022451016,0.001593722,0.0018242949],"genre_scores_gemma":[0.7552218,0.00029599495,0.23792507,0.0002112266,0.000086434615,0.0001721658,0.0007470979,0.00024075214,0.005099514],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996811,0.000053819815,0.000017216586,0.00010405977,0.00008196046,0.00006183515],"domain_scores_gemma":[0.99944085,0.00022293784,0.00005810237,0.00008880058,0.00014730515,0.000041939944],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00066042354,0.0007725386,0.0009458678,0.0005279849,0.0003220009,0.0008779212,0.0025200478,0.0011693347,0.0048991274],"category_scores_gemma":[0.0020494666,0.00045966372,0.0006757112,0.00044515755,0.0005899517,0.0011040244,0.0011521006,0.0016766326,0.000789146],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011221818,0.000097507254,0.0006341479,0.0000809222,0.00004457537,0.00008462937,0.00005986927,0.8639094,0.003155795,0.02021226,0.0033433074,0.10826527],"study_design_scores_gemma":[0.000003140622,0.000008387973,0.00004199818,0.0000029598614,0.0000035450146,0.0000043131736,0.0000017220862,0.9956185,0.0002360683,0.0038871877,0.00019008823,0.0000019767933],"about_ca_topic_score_codex":0.010754869,"about_ca_topic_score_gemma":0.017191842,"teacher_disagreement_score":0.010754869,"about_ca_system_score_codex":0.001137878,"about_ca_system_score_gemma":0.0016728514,"threshold_uncertainty_score":0.021384537},"labels":[],"label_agreement":null},{"id":"W4390245620","doi":"10.1016/j.datak.2023.102274","title":"Mining Keys for Graphs","year":2023,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McMaster University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Graph; Data deduplication; Theoretical computer science; Identification (biology); Data mining; Pruning; Database","score_opus":0.3915863678661295,"score_gpt":0.45612684361496436,"score_spread":0.06454047574883487,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4390245620","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.09354493,0.0030094627,0.83772933,0.0040127584,0.0005295525,0.0011661778,0.037019804,0.008719591,0.0142684085],"genre_scores_gemma":[0.44767943,0.0031447937,0.49160752,0.00062716956,0.00035862147,0.0008765376,0.042828046,0.0013494783,0.011528304],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99655247,0.00044452306,0.00040673272,0.0013027628,0.0009276579,0.0003657693],"domain_scores_gemma":[0.9858899,0.007058039,0.0012451652,0.0034759147,0.0018204562,0.0005104181],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0018084558,0.0013885954,0.0010672362,0.008389687,0.001729383,0.0043906854,0.0018696488,0.0014154318,0.009535979],"category_scores_gemma":[0.025124095,0.0010563949,0.002478394,0.00659231,0.0013871203,0.01894497,0.003869871,0.0022894377,0.004343549],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00074778585,0.0003087108,0.019171566,0.0023731734,0.00041084783,0.0010839801,0.0017954289,0.016518196,0.016705157,0.34317178,0.074952275,0.5227611],"study_design_scores_gemma":[0.00007423409,0.0001504275,0.0030830756,0.00044079366,0.00031734933,0.0012657889,0.00095772726,0.05755316,0.012168627,0.8686191,0.055295654,0.00007416354],"about_ca_topic_score_codex":0.002070138,"about_ca_topic_score_gemma":0.0035402486,"teacher_disagreement_score":0.009535979,"about_ca_system_score_codex":0.0013043737,"about_ca_system_score_gemma":0.0021736678,"threshold_uncertainty_score":0.031901002},"labels":[],"label_agreement":null},{"id":"W4394987794","doi":"10.1016/j.datak.2024.102306","title":"Effective text classification using BERT, MTM LSTM, and DT","year":2024,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":56,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada); University of Calgary","funders":"","keywords":"Security token; Computer science; Artificial intelligence; Encoder; Binary classification; Transformer; Recall; Deep learning; Long short term memory; Natural language processing; Artificial neural network; Machine learning; Recurrent neural network; Support vector machine","score_opus":0.04931219236959951,"score_gpt":0.3094819476933262,"score_spread":0.2601697553237267,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394987794","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.049094092,0.0022570896,0.9276817,0.0012633511,0.0011431879,0.00016206835,0.0017180869,0.009320917,0.007359538],"genre_scores_gemma":[0.46537128,0.0012101938,0.5054748,0.0007186317,0.00077770895,0.00022373091,0.0041403924,0.00042960208,0.021653691],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995552,0.00007035082,0.000045274086,0.00013954632,0.00012682218,0.000062814295],"domain_scores_gemma":[0.99908495,0.00032543222,0.00007557225,0.00012442042,0.00032669966,0.000062820916],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00062062964,0.0008233437,0.00072448223,0.0018119869,0.0006673686,0.0010940748,0.0010243567,0.0012921399,0.005158818],"category_scores_gemma":[0.0024480983,0.00022915447,0.00060308794,0.0018133991,0.00030550736,0.0026564065,0.0007310713,0.0012833177,0.0033267254],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002427495,0.0001341855,0.0007072202,0.00014614622,0.00004037334,0.000099432364,0.00004500004,0.020095931,0.02194989,0.0048454525,0.0138783045,0.9378154],"study_design_scores_gemma":[0.000018067498,0.00007866747,0.000664972,0.000024120145,0.000038737275,0.00011809509,0.000039396935,0.9720643,0.013423923,0.008044233,0.0054685655,0.00001688484],"about_ca_topic_score_codex":0.007583996,"about_ca_topic_score_gemma":0.00974217,"teacher_disagreement_score":0.007583996,"about_ca_system_score_codex":0.0009054172,"about_ca_system_score_gemma":0.0015728936,"threshold_uncertainty_score":0.017257929},"labels":[],"label_agreement":null},{"id":"W4398243240","doi":"10.1016/j.datak.2024.102324","title":"Large language models: Expectations for semantics-driven systems engineering","year":2024,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Systems Engineering Methodologies and Applications","field":"Engineering","cited_by":49,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Foundation for Research and Technology-Hellas","keywords":"Computer science; Semantics (computer science); Programming language; Modeling language; Operational semantics; Software engineering","score_opus":0.07354628162855453,"score_gpt":0.31576459170632243,"score_spread":0.2422183100777679,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4398243240","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.014611867,0.00095488963,0.9553972,0.015923033,0.00032096464,0.00020004116,0.0005269573,0.0010436075,0.011021362],"genre_scores_gemma":[0.49547338,0.0017908473,0.48738733,0.0047245333,0.0014884697,0.0013675016,0.00203047,0.0012254379,0.0045121103],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.98432297,0.007637258,0.00076247874,0.0016338507,0.005026264,0.0006171371],"domain_scores_gemma":[0.8151162,0.14738862,0.003599877,0.018679122,0.012117001,0.0030991768],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0244144,0.0013019689,0.0027155806,0.0019965891,0.0019560822,0.009456662,0.0047329287,0.005453031,0.0068023433],"category_scores_gemma":[0.15589733,0.001811753,0.0025501444,0.0013573108,0.008140259,0.026429256,0.004648992,0.00978821,0.0023394232],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017468125,0.0001374414,0.0009831304,0.00024168829,0.00007099189,0.0000779393,0.0004978883,0.026638823,0.000860552,0.9520926,0.0037480968,0.014476069],"study_design_scores_gemma":[0.000033997643,0.000024937215,0.00010030031,0.000042436423,0.000019201982,0.00003431596,0.00008745705,0.07095242,0.0005204681,0.9251095,0.003055513,0.000019486892],"about_ca_topic_score_codex":0.004967204,"about_ca_topic_score_gemma":0.0033555147,"teacher_disagreement_score":0.0244144,"about_ca_system_score_codex":0.0039206627,"about_ca_system_score_gemma":0.0042056832,"threshold_uncertainty_score":0.12911725},"labels":[],"label_agreement":null},{"id":"W4405632789","doi":"10.1016/j.datak.2024.102393","title":"Coupling MDL and Markov chain Monte Carlo to sample diverse pattern sets","year":2024,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Image Retrieval and Classification Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Markov chain Monte Carlo; Coupling (piping); Statistical physics; Monte Carlo method; Markov chain; Computer science; Sample (material); Algorithm; Physics; Mathematics; Statistics; Machine learning; Materials science; Thermodynamics","score_opus":0.033426819862946294,"score_gpt":0.28473330654046375,"score_spread":0.2513064866775174,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405632789","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.013604059,0.00014711157,0.9843806,0.00020385499,0.000038963884,0.00007034868,0.000070657166,0.0006515755,0.00083277875],"genre_scores_gemma":[0.57350945,0.00021141766,0.42199975,0.00041729223,0.00012849252,0.00051610894,0.0005823686,0.00039796796,0.0022371945],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983468,0.00074974966,0.000100200166,0.0002410619,0.00042428676,0.00013790894],"domain_scores_gemma":[0.973656,0.022141987,0.00084027083,0.0015855298,0.0012988058,0.00047744255],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051906314,0.0006738356,0.0021026114,0.0018815857,0.00079582736,0.0020806678,0.0026692918,0.0022385314,0.0038445122],"category_scores_gemma":[0.026014239,0.0014090375,0.0011539474,0.0017834998,0.0016672013,0.0025936423,0.0021574106,0.0026298899,0.0007915583],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000050977786,0.00006587956,0.0007122814,0.000042809697,0.000044934866,0.000025344754,0.00002618874,0.97079057,0.00023603698,0.012502472,0.00035416672,0.015148362],"study_design_scores_gemma":[0.0000038269823,0.0000028578183,0.0000192402,0.0000018202094,0.0000018170944,0.000002580197,0.000001258577,0.99532604,0.000053573807,0.0045302035,0.00005477308,0.0000021422984],"about_ca_topic_score_codex":0.013033279,"about_ca_topic_score_gemma":0.014624813,"teacher_disagreement_score":0.013033279,"about_ca_system_score_codex":0.0022307097,"about_ca_system_score_gemma":0.0025467314,"threshold_uncertainty_score":0.027450979},"labels":[],"label_agreement":null},{"id":"W4408912911","doi":"10.1016/j.datak.2025.102440","title":"Customized long short-term memory architecture for multi-document summarization with improved text feature set","year":2025,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Topic Modeling","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Automatic summarization; Computer science; Term (time); Feature (linguistics); Set (abstract data type); Information retrieval; Architecture; Natural language processing; Artificial intelligence; Programming language; History; Linguistics","score_opus":0.023473699877713024,"score_gpt":0.2815724607136532,"score_spread":0.2580987608359402,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408912911","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07622569,0.003327904,0.8527089,0.0003597753,0.00069491565,0.0003199404,0.0039363685,0.058898322,0.0035283226],"genre_scores_gemma":[0.38012818,0.0008823655,0.5864038,0.000524704,0.00043704634,0.00074454985,0.013350858,0.0008394371,0.016689075],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99971324,0.00002908471,0.00004603561,0.000094105315,0.00007271098,0.0000448117],"domain_scores_gemma":[0.99932754,0.00014338367,0.000050497794,0.00013440034,0.00029810626,0.000046153087],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00041364605,0.0009054534,0.00085620984,0.0017896343,0.00058967876,0.0010729632,0.0016668164,0.0006282655,0.0087756235],"category_scores_gemma":[0.0010196218,0.00030578446,0.00059703144,0.002269811,0.00015159238,0.0014936435,0.0007630839,0.0006535944,0.0043267175],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011295596,0.0003265448,0.0010697739,0.00030370487,0.0001864673,0.00017631822,0.00015546278,0.008972655,0.08531071,0.0011286136,0.025212532,0.8760276],"study_design_scores_gemma":[0.00028481727,0.0013084314,0.0047110203,0.00006506864,0.0005268432,0.00041754145,0.0002599387,0.76520306,0.18772696,0.0046844687,0.034664642,0.00014715375],"about_ca_topic_score_codex":0.0056710136,"about_ca_topic_score_gemma":0.009160771,"teacher_disagreement_score":0.0087756235,"about_ca_system_score_codex":0.0005435116,"about_ca_system_score_gemma":0.0010327755,"threshold_uncertainty_score":0.029357374},"labels":[],"label_agreement":null},{"id":"W4411507372","doi":"10.1016/j.datak.2025.102480","title":"Large language models for conceptual modeling: Assessment and application potential","year":2025,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"Fundação para a Ciência e a Tecnologia; Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science","score_opus":0.029500594423617237,"score_gpt":0.3180202898758975,"score_spread":0.28851969545228023,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4411507372","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.05721329,0.0023980243,0.925136,0.0039836755,0.00016556942,0.00030113707,0.0019163804,0.0027552433,0.0061305785],"genre_scores_gemma":[0.44115838,0.0022546959,0.54873806,0.00048273354,0.0002037658,0.00075873564,0.003551555,0.0007204637,0.002131542],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99339175,0.004382949,0.00026802652,0.00044549283,0.0013990315,0.00011285117],"domain_scores_gemma":[0.92195374,0.065008074,0.0013238094,0.006596714,0.0042820782,0.00083565776],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012134491,0.00089211564,0.0011335274,0.003050379,0.0012175994,0.005642711,0.0029533822,0.0018483388,0.0055613844],"category_scores_gemma":[0.062229194,0.0008553507,0.0018076116,0.003184738,0.0014637747,0.012814323,0.0021998116,0.002262287,0.0010592544],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009196377,0.0008963728,0.009566668,0.0009416882,0.0007076706,0.00024773495,0.001305461,0.24860243,0.0025782161,0.50263196,0.014084072,0.21751808],"study_design_scores_gemma":[0.00007220844,0.00008645376,0.0006325508,0.00013156087,0.00014826811,0.000080627346,0.00026103124,0.7609238,0.0010507433,0.22845645,0.008116593,0.000039728788],"about_ca_topic_score_codex":0.010133788,"about_ca_topic_score_gemma":0.012803627,"teacher_disagreement_score":0.012134491,"about_ca_system_score_codex":0.0033687684,"about_ca_system_score_gemma":0.0034646762,"threshold_uncertainty_score":0.064174116},"labels":[],"label_agreement":null},{"id":"W4411540031","doi":"10.1016/j.datak.2025.102482","title":"Domain knowledge in artificial intelligence: Using conceptual modeling to increase machine learning accuracy and explainability","year":2025,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Artificial intelligence; Computer science; Domain (mathematical analysis); Machine learning; Domain knowledge; Artificial Intelligence System; Mathematics","score_opus":0.06691927417424057,"score_gpt":0.33926893247535406,"score_spread":0.27234965830111346,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4411540031","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.08609538,0.0015484204,0.90261316,0.003088485,0.00013748878,0.00014963913,0.0003135532,0.00078979717,0.0052639917],"genre_scores_gemma":[0.73426217,0.0008528435,0.26289275,0.00031239382,0.000114189,0.00022857246,0.00044415845,0.0001433318,0.00074963184],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9951349,0.0029488693,0.00029244673,0.0006531387,0.0008702002,0.00010053912],"domain_scores_gemma":[0.9426749,0.04404849,0.0023807425,0.007825116,0.002715843,0.0003548739],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0073386705,0.0007347451,0.0006889307,0.0025051548,0.0006811179,0.003533021,0.0016487263,0.0014509347,0.0019061747],"category_scores_gemma":[0.060694814,0.00049943244,0.0011765457,0.0021638626,0.0015612325,0.010764025,0.002748304,0.002611614,0.00032529002],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00053612154,0.0007695143,0.025536878,0.00071617396,0.0005934539,0.00020343794,0.0018778947,0.18726629,0.004803581,0.27587038,0.0043444177,0.49748188],"study_design_scores_gemma":[0.00004374451,0.00007296571,0.00213981,0.00012674903,0.00018415092,0.00006311284,0.00015973572,0.7046068,0.0028846143,0.28671876,0.0029620202,0.0000374906],"about_ca_topic_score_codex":0.0031033398,"about_ca_topic_score_gemma":0.0025668289,"teacher_disagreement_score":0.0073386705,"about_ca_system_score_codex":0.0014038379,"about_ca_system_score_gemma":0.001498784,"threshold_uncertainty_score":0.038811088},"labels":[],"label_agreement":null},{"id":"W4412980931","doi":"10.1016/j.datak.2025.102499","title":"Corrigendum to “Large Language Models for Conceptual Modeling: Assessment and Application Potential” [Knowledge and Data Engineering Volume 160, November 2025, 102480]","year":2025,"lang":"en","type":"erratum","venue":"Data & Knowledge Engineering","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Volume (thermodynamics); Computer science; Conceptual model; Data science; Database; Physics","score_opus":0.04627239330600777,"score_gpt":0.3163658736183723,"score_spread":0.2700934803123646,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4412980931","genre_codex":"editorial","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00022045914,0.0017668651,0.004826576,0.10807558,0.83411396,0.00013338917,0.004434102,0.0020546748,0.0443744],"genre_scores_gemma":[0.0040309755,0.0035380505,0.0068503157,0.063892834,0.09925779,0.00030240865,0.008664257,0.0022690278,0.8111943],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9929202,0.0011700835,0.00068645034,0.0008516067,0.00386456,0.00050719414],"domain_scores_gemma":[0.9663382,0.0053836,0.000722807,0.001839002,0.02476387,0.0009524017],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0049090427,0.0024690328,0.0023804677,0.006057404,0.0059609218,0.0088021895,0.0037076895,0.0079813665,0.21374533],"category_scores_gemma":[0.041025594,0.0013573879,0.002527323,0.004284673,0.0017751093,0.0045255115,0.0029044582,0.007878931,0.1587043],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000007958095,0.000003677651,0.000011350639,0.000018875699,0.0000022005268,0.000014127528,0.0000038442413,0.00002608441,0.000017864711,0.0005266033,0.9973673,0.002000152],"study_design_scores_gemma":[0.000024976105,0.00001932901,0.00046932453,0.00015210902,0.000026757885,0.00005684229,0.000050604478,0.00080986705,0.00028241175,0.0036418273,0.9944285,0.000037373837],"about_ca_topic_score_codex":0.08813502,"about_ca_topic_score_gemma":0.13421507,"teacher_disagreement_score":0.21374533,"about_ca_system_score_codex":0.008395775,"about_ca_system_score_gemma":0.006837992,"threshold_uncertainty_score":0.7150494},"labels":[],"label_agreement":null},{"id":"W4413052970","doi":"10.1016/j.datak.2025.102500","title":"Corrigendum to “Domain knowledge in artificial intelligence: Using conceptual modeling to increase machine learning accuracy and explainability” [Knowledge and Data Engineering Volume 160, November 2025, 102482]","year":2025,"lang":"en","type":"erratum","venue":"Data & Knowledge Engineering","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Computer science; Domain knowledge; Artificial intelligence; Domain (mathematical analysis); Volume (thermodynamics); Machine learning; Knowledge engineering; Mathematics","score_opus":0.08647812660475818,"score_gpt":0.3363630339300379,"score_spread":0.24988490732527974,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413052970","genre_codex":"editorial","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00013155787,0.001344219,0.00082034286,0.088835694,0.8841364,0.000083099214,0.0019373412,0.00068433856,0.02202701],"genre_scores_gemma":[0.004733262,0.00362777,0.0028665257,0.15787514,0.15417552,0.00031424087,0.004436363,0.0011825495,0.6707886],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99206233,0.0009714846,0.0009115919,0.0010614805,0.004258444,0.0007347799],"domain_scores_gemma":[0.9704066,0.003961705,0.00068500487,0.0014347087,0.022613049,0.00089901814],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004259335,0.002642672,0.0030433002,0.00586689,0.007568601,0.008489485,0.0044912747,0.012374864,0.14082858],"category_scores_gemma":[0.039910495,0.001332977,0.003144266,0.0038762873,0.0021452704,0.0032196334,0.0032371108,0.010183068,0.11079195],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000009688247,0.0000042461747,0.000015066264,0.000022153597,0.0000031565608,0.000020876163,0.0000044349304,0.00001690521,0.000013832768,0.00038539243,0.9979533,0.0015509232],"study_design_scores_gemma":[0.00003598976,0.00002441195,0.0010165687,0.00019269747,0.000037836842,0.000059202666,0.000063418476,0.00048956735,0.00022740827,0.002662021,0.99514705,0.00004383875],"about_ca_topic_score_codex":0.09917548,"about_ca_topic_score_gemma":0.16132478,"teacher_disagreement_score":0.14082858,"about_ca_system_score_codex":0.010529633,"about_ca_system_score_gemma":0.0080237165,"threshold_uncertainty_score":0.4711185},"labels":[],"label_agreement":null},{"id":"W4416703099","doi":"10.1016/j.datak.2025.102536","title":"Enhancing clustering stability, compactness, and separation in multimodal data environments","year":2025,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Customer churn and segmentation","field":"Business, Management and Accounting","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Ministério da Ciência, Tecnologia, Inovações e Comunicações","keywords":"Cluster analysis; Benchmark (surveying); Stability (learning theory); Rand index; Hierarchical clustering; Benchmarking; Consensus clustering; Embedding; Focus (optics); Core (optical fiber)","score_opus":0.04420924631857913,"score_gpt":0.29964580499908583,"score_spread":0.2554365586805067,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4416703099","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.30202004,0.0004089859,0.6946422,0.00033303408,0.00003014343,0.00006477889,0.00014008729,0.00091492326,0.0014457426],"genre_scores_gemma":[0.86820436,0.0001547877,0.12943609,0.000072715025,0.00005533324,0.000060625913,0.00049509725,0.00022408637,0.0012968804],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9977767,0.0007557273,0.00014957844,0.00048597332,0.00056913076,0.000262857],"domain_scores_gemma":[0.99130416,0.005426342,0.00068670535,0.0007415349,0.0014816668,0.00035968176],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003780584,0.0011308593,0.0013267796,0.0026047272,0.0013180526,0.0025176432,0.0014018206,0.0013667276,0.00094161916],"category_scores_gemma":[0.015340299,0.00065395114,0.0007854956,0.0023760037,0.0010679375,0.003566838,0.0037269157,0.001101693,0.0004215062],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.001974484,0.00072075654,0.02238409,0.00027189511,0.00039382343,0.00025813855,0.0019968404,0.47921276,0.03693017,0.010381032,0.0029171938,0.44255883],"study_design_scores_gemma":[0.000017487422,0.00010901989,0.003296635,0.000011261961,0.000044221433,0.000069539296,0.00034798923,0.9817245,0.008051221,0.0058744755,0.00043201103,0.000021599211],"about_ca_topic_score_codex":0.005427242,"about_ca_topic_score_gemma":0.0070869634,"teacher_disagreement_score":0.005427242,"about_ca_system_score_codex":0.0012622721,"about_ca_system_score_gemma":0.001227678,"threshold_uncertainty_score":0.019993901},"labels":[],"label_agreement":null}]}