{"meta":{"query_hash":"a02b92407db7","filters":{"venue":"Data Science Journal"},"cohort_total":31,"direct_labels_cover":4,"predictions_cover":31,"exported":31,"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/a02b92407db7","api":"https://metacan.xera.ac/api/v1/cohort?venue=Data+Science+Journal"},"results":[{"id":"W1446508819","doi":"","title":"The Canadian Enhanced Polar Outflow Probe (e-POP) Mission: Current Status and Planned Observations and Data Distribution:Current Status and Planned Observations and Data Distribution","year":2009,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Methane Hydrates and Related Phenomena","field":"Environmental Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Current (fluid); Outflow; Distribution (mathematics); Polar; Environmental science; Meteorology; Geography; Physics; Engineering; Electrical engineering; Mathematics","score_opus":0.08234495579074003,"score_gpt":0.31894442929116634,"score_spread":0.23659947350042632,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1446508819","genre_codex":"dataset","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":[0.18012497,0.003922798,0.019847246,0.004012659,0.0005601712,0.003287516,0.6300001,0.010369401,0.1478751],"genre_scores_gemma":[0.34631625,0.004666182,0.07436365,0.0016795773,0.0003128536,0.001979341,0.5274842,0.0015774033,0.041620586],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9984816,0.000057007488,0.000032763975,0.00013883128,0.00090756186,0.00038232072],"domain_scores_gemma":[0.9946137,0.00022251309,0.00030663962,0.00036599854,0.0035337876,0.00095738703],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027821022,0.0011700435,0.0007633232,0.0038818978,0.0029396901,0.0014909592,0.002879068,0.0009693034,0.0101396395],"category_scores_gemma":[0.0020606637,0.0004597663,0.00040150285,0.006455013,0.000958647,0.0014257834,0.0014380572,0.0010186839,0.0036221102],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.004553875,0.0007478161,0.13170162,0.0017202183,0.00036189368,0.0003029008,0.0010217439,0.0068200664,0.06847171,0.0035110575,0.40773875,0.37304837],"study_design_scores_gemma":[0.00080674514,0.00040095058,0.33155313,0.0003321204,0.00026238683,0.00012386577,0.00075353775,0.010006101,0.032354295,0.0013175318,0.6218203,0.00026904215],"about_ca_topic_score_codex":0.85996073,"about_ca_topic_score_gemma":0.8964268,"teacher_disagreement_score":0.85996073,"about_ca_system_score_codex":0.009020347,"about_ca_system_score_gemma":0.03310571,"threshold_uncertainty_score":0.2817279},"labels":[],"label_agreement":null},{"id":"W1978333903","doi":"10.2481/dsj.1.45","title":"A Mexican case study on a centralised database from world natural history museums","year":2002,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":33,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Distribution (mathematics); Geography; Ornithology; Natural history; Species richness; Endemism; Database; Collections management; Library science; Ethnology; Ecology; Archaeology; History; Computer science; Biology; Southern Hemisphere","score_opus":0.14651411399662242,"score_gpt":0.3052122749726442,"score_spread":0.15869816097602177,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1978333903","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":[0.977797,0.0007697146,0.0017444241,0.0011531715,0.00001818289,0.0004822142,0.008052336,0.00006978461,0.009913264],"genre_scores_gemma":[0.9782681,0.00075343467,0.008134049,0.00017677236,0.00002658258,0.00043047892,0.009985401,0.000039774783,0.0021853468],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9947995,0.0023021055,0.0005350628,0.0006262479,0.001326,0.00041113366],"domain_scores_gemma":[0.98287225,0.007948549,0.002478328,0.0027617074,0.0033792,0.000560018],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0069222194,0.00025037312,0.00036306516,0.003965226,0.0022108802,0.0024727467,0.0013760633,0.0010111695,0.00415249],"category_scores_gemma":[0.020305382,0.0003505138,0.00034723524,0.01107062,0.00082420174,0.0024967454,0.0018956611,0.00069577596,0.00046432877],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00059211254,0.0008143357,0.8394696,0.00092546054,0.00014266271,0.009956315,0.02962001,0.001449011,0.0018937764,0.005357296,0.014220216,0.09555928],"study_design_scores_gemma":[0.00013116711,0.00036909417,0.85218525,0.0006360211,0.00018812415,0.006864709,0.054782297,0.0064069475,0.0018576115,0.0007663254,0.07572561,0.00008691803],"about_ca_topic_score_codex":0.08720831,"about_ca_topic_score_gemma":0.17799622,"teacher_disagreement_score":0.08720831,"about_ca_system_score_codex":0.0032972314,"about_ca_system_score_gemma":0.0024075005,"threshold_uncertainty_score":0.17340136},"labels":[],"label_agreement":null},{"id":"W1985018337","doi":"10.2481/dsj.4.21","title":"Rescuing and recovering lost or endangered data","year":2005,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Digital and Cyber Forensics","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dominion Astrophysical Observatory; Herzberg Institute of Astrophysics","funders":"","keywords":"Scope (computer science); Computer science; Transparency (behavior); Open data; Metadata; Usability; Reuse; Data science; Implementation; World Wide Web; Data curation; Open science; Engineering; Software engineering; Computer security","score_opus":0.1006452456693089,"score_gpt":0.31146841387309215,"score_spread":0.21082316820378327,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1985018337","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":[0.11878092,0.007384607,0.7044547,0.053696856,0.0068231868,0.0013487986,0.0031903246,0.009874385,0.09444623],"genre_scores_gemma":[0.55783564,0.004805648,0.32703403,0.00865272,0.0016342037,0.00054910587,0.0050860457,0.0026066182,0.09179595],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98341733,0.004969651,0.0014672709,0.0015999852,0.0073250816,0.0012207379],"domain_scores_gemma":[0.8813122,0.024025384,0.00888492,0.06468382,0.017841363,0.0032523575],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.015669877,0.0009272688,0.0011438134,0.006391687,0.005175385,0.007431432,0.0056672394,0.005158108,0.010680277],"category_scores_gemma":[0.11212153,0.0007494161,0.0013168544,0.0042488417,0.0041867443,0.01086217,0.012421403,0.00515339,0.008839803],"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":[0.0006518177,0.00034224943,0.016366525,0.0011264033,0.00023941195,0.0036030815,0.00691415,0.005678807,0.009325089,0.13702784,0.1307702,0.6879544],"study_design_scores_gemma":[0.00009675656,0.00047602295,0.0085017495,0.002903951,0.0003471412,0.008993546,0.009623928,0.038417988,0.043851305,0.24917962,0.637296,0.00031198256],"about_ca_topic_score_codex":0.002447749,"about_ca_topic_score_gemma":0.0023580324,"teacher_disagreement_score":0.015669877,"about_ca_system_score_codex":0.0012466849,"about_ca_system_score_gemma":0.008742338,"threshold_uncertainty_score":0.08287126},"labels":[],"label_agreement":null},{"id":"W2021867077","doi":"10.2481/dsj.8.27","title":"CISTI'S Activities in Support of Scientific Data Management in Canada 2008-2010","year":2009,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Scope (computer science); Computer science; Transparency (behavior); Usability; Metadata; Implementation; Open data; Reuse; Data management; Data science; World Wide Web; Publishing; Data publishing; Political science; Database; Engineering; Software engineering; Computer security","score_opus":0.16621443021793386,"score_gpt":0.38273437009272043,"score_spread":0.21651993987478657,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2021867077","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0636505,0.024088075,0.05265517,0.22056569,0.013041455,0.004648612,0.065772265,0.014304887,0.54127336],"genre_scores_gemma":[0.21559276,0.024375724,0.1311391,0.010576713,0.0026393498,0.0014601215,0.062268168,0.0028697082,0.54907835],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9847771,0.0012571913,0.00067476777,0.0011625051,0.010400846,0.0017276673],"domain_scores_gemma":[0.9127129,0.0040269177,0.0020595514,0.002768007,0.0653463,0.013086419],"candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.017506473,0.0009591047,0.00089942326,0.006863173,0.00767073,0.007392597,0.003481832,0.001331164,0.020764997],"category_scores_gemma":[0.039159846,0.00067788706,0.0005786565,0.018134031,0.0029142199,0.002846597,0.004792187,0.0024728796,0.007944568],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":true,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0004998741,0.00016165776,0.016210595,0.0007570591,0.00006699267,0.00067754474,0.0049846605,0.0016688029,0.0050404873,0.042307884,0.6491129,0.27851152],"study_design_scores_gemma":[0.000031165386,0.000027633787,0.0076129544,0.00013238171,0.000011808543,0.00009294243,0.001320962,0.0018822011,0.0019414903,0.0009758349,0.9859416,0.000028993672],"about_ca_topic_score_codex":0.9673039,"about_ca_topic_score_gemma":0.9491015,"teacher_disagreement_score":0.9965182,"about_ca_system_score_codex":0.08695627,"about_ca_system_score_gemma":0.37163037,"threshold_uncertainty_score":0.6309148},"labels":[],"label_agreement":null},{"id":"W2031989156","doi":"10.2481/dsj.2.25","title":"Data integration and knowledge discovery in biomedical databases. Reliable information from unreliable sources","year":2003,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Artificial Intelligence in Healthcare","field":"Health Professions","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"","keywords":"Scope (computer science); Computer science; Data science; Usability; Metadata; Transparency (behavior); Open data; Reuse; Implementation; World Wide Web; Data discovery; Data publishing; Publishing; Software engineering; Political science; Engineering","score_opus":0.2754938637313639,"score_gpt":0.5016500590355036,"score_spread":0.22615619530413972,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2031989156","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":[0.0067679877,0.076502904,0.85501075,0.034708902,0.0013043983,0.000595469,0.0027689547,0.0018635234,0.020477086],"genre_scores_gemma":[0.1494963,0.044748392,0.7807166,0.005439686,0.0024836578,0.00087977643,0.0073523987,0.00057208166,0.0083111115],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9604355,0.018429363,0.0045870477,0.0030945528,0.0127971675,0.0006563522],"domain_scores_gemma":[0.93059564,0.047174703,0.0044741128,0.011595683,0.00530367,0.0008560392],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.043434348,0.0009461007,0.0033378208,0.015827496,0.0026863944,0.017415568,0.003775336,0.003931697,0.004110878],"category_scores_gemma":[0.111411326,0.0019487641,0.002696335,0.026899697,0.0067494125,0.02651634,0.01253912,0.0036729975,0.0031759571],"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":[0.0002291806,0.00011085976,0.0044486756,0.0042726267,0.00059297966,0.00090350595,0.0031041272,0.008177101,0.0013139078,0.5193773,0.033527367,0.4239423],"study_design_scores_gemma":[0.000043894994,0.000046202484,0.0013132268,0.0016569921,0.0002440066,0.00089788163,0.0009914953,0.0338826,0.002417628,0.8319884,0.12643251,0.00008509575],"about_ca_topic_score_codex":0.0033346259,"about_ca_topic_score_gemma":0.0027853095,"teacher_disagreement_score":0.043434348,"about_ca_system_score_codex":0.0026365924,"about_ca_system_score_gemma":0.006708412,"threshold_uncertainty_score":0.22970557},"labels":[],"label_agreement":null},{"id":"W2032869846","doi":"10.2481/dsj.6.od26","title":"Canadian National Consultation on Access to Scientific Research Data","year":2007,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"National Research Council Canada; Université de Montréal","funders":"National Research Council Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Strong","keywords":"Scope (computer science); Computer science; Transparency (behavior); Usability; Metadata; Data science; Open data; Reuse; Implementation; World Wide Web; Engineering; Computer security; Software engineering","score_opus":0.6443846780180462,"score_gpt":0.5804227013872811,"score_spread":0.06396197663076508,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2032869846","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0010281061,0.02302877,0.0018115999,0.7502089,0.01955619,0.0010371138,0.037136003,0.00079944154,0.16539393],"genre_scores_gemma":[0.0193063,0.03059763,0.021726316,0.5477103,0.003638262,0.0027806757,0.03007909,0.0010288648,0.34313256],"study_design_codex":"not_applicable","study_design_gemma":"qualitative","domain_scores_codex":[0.8626081,0.015920015,0.012512413,0.0061051445,0.08654446,0.016309874],"domain_scores_gemma":[0.64353967,0.08876326,0.004989986,0.016699009,0.20630334,0.039704707],"candidate_categories":["metaresearch","scholarly_communication","open_science"],"consensus_categories":[],"category_scores_codex":[0.11025025,0.0016548889,0.0034403855,0.013381071,0.01695284,0.01550049,0.012167596,0.023436485,0.046190053],"category_scores_gemma":[0.19637789,0.0023133755,0.004270209,0.026353488,0.0065321536,0.0061430247,0.012361742,0.018411007,0.009258871],"study_design_candidate":"qualitative","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":true,"about_ca_system_consensus":true,"study_design_scores_codex":[0.00003205033,0.000008492242,0.0002555374,0.00035198644,0.000017400853,0.00007688293,0.00039071712,0.000059521386,0.00007239095,0.012778254,0.97928613,0.0066706263],"study_design_scores_gemma":[0.000027117014,0.000004768072,0.0021954144,0.0006693461,0.000026875725,0.000039971987,0.00033716447,0.00006756962,0.000066917826,0.0015220895,0.9949964,0.000046332116],"about_ca_topic_score_codex":0.9719106,"about_ca_topic_score_gemma":0.9758616,"teacher_disagreement_score":0.9878324,"about_ca_system_score_codex":0.16417052,"about_ca_system_score_gemma":0.6882334,"threshold_uncertainty_score":0.96944374},"labels":[],"label_agreement":null},{"id":"W2036014579","doi":"10.2481/dsj.5.178","title":"Geography Markup Language","year":2006,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Geographic Information Systems Studies","field":"Social Sciences","cited_by":105,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Galdos Systems (Canada)","funders":"","keywords":"Markup language; XML; Geospatial analysis; Geomatics; Computer science; Document type definition; World Wide Web; Interoperability; Information retrieval; Geography; Document Structure Description; Remote sensing","score_opus":0.024363899319102905,"score_gpt":0.3337672918235531,"score_spread":0.30940339250445015,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2036014579","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":[0.005304781,0.0021876306,0.2965349,0.008381772,0.0021635238,0.0014611051,0.08105385,0.035897054,0.5670153],"genre_scores_gemma":[0.06493755,0.0050293976,0.38215515,0.006716901,0.0009130063,0.0023651826,0.17983904,0.013294785,0.3447489],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9975782,0.0006349872,0.00041788566,0.00034630104,0.0008540907,0.00016857786],"domain_scores_gemma":[0.99557245,0.0012013642,0.00032086106,0.0012294538,0.0014298459,0.00024591797],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0023521918,0.00082173775,0.0006352651,0.0038898885,0.0014163838,0.004749239,0.002649446,0.0012910733,0.09463055],"category_scores_gemma":[0.010032222,0.00044173145,0.0005514633,0.0049635777,0.0010815355,0.0053529916,0.0036931129,0.0017894802,0.07719017],"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":[0.00009971897,0.00006787583,0.0010212498,0.0005879956,0.00002096056,0.0004972304,0.001698612,0.0015869311,0.0025650135,0.28755382,0.53163683,0.17266388],"study_design_scores_gemma":[0.000005517233,0.0000055453334,0.00016096218,0.000061703075,0.0000039645565,0.00018082853,0.00010832079,0.00033743642,0.0005495433,0.0122375125,0.98633426,0.000014520277],"about_ca_topic_score_codex":0.0048461407,"about_ca_topic_score_gemma":0.0033611294,"teacher_disagreement_score":0.09463055,"about_ca_system_score_codex":0.0013306538,"about_ca_system_score_gemma":0.0029538076,"threshold_uncertainty_score":0.3165707},"labels":[],"label_agreement":null},{"id":"W2059472642","doi":"10.2481/dsj.5.143","title":"A hashing technique using separate binary tree","year":2006,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Scope (computer science); Usability; Data science; Metadata; Transparency (behavior); Data publishing; Implementation; World Wide Web; Publishing; Software engineering; Computer security; Political science","score_opus":0.05802956779394889,"score_gpt":0.3299944953690478,"score_spread":0.2719649275750989,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2059472642","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":[0.02263773,0.0018433383,0.964668,0.0002318936,0.00080230174,0.00023851359,0.000349516,0.0025920018,0.0066367197],"genre_scores_gemma":[0.23105596,0.0012095388,0.7406788,0.0003651432,0.00039238192,0.00028674866,0.0018110988,0.00047710727,0.023723263],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984341,0.00018948464,0.0001510663,0.000327001,0.000734364,0.0001639809],"domain_scores_gemma":[0.99780625,0.00043680076,0.00013857287,0.0009493836,0.00056302216,0.000106074185],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00095758995,0.0006172356,0.0011931361,0.002315005,0.0012410335,0.0013968665,0.0018346974,0.0012627905,0.012704835],"category_scores_gemma":[0.0038465927,0.00050746,0.0008589944,0.003310687,0.0008217053,0.004128381,0.0030084061,0.0011440995,0.008978047],"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":[0.0007375075,0.00018701713,0.0012553982,0.00042727808,0.00008238549,0.00021904107,0.0003589464,0.008193961,0.05123429,0.027969398,0.015217384,0.8941174],"study_design_scores_gemma":[0.0004442316,0.0025388119,0.0055326084,0.00034470393,0.0003886396,0.0075619044,0.0009428406,0.5053044,0.22679394,0.078072555,0.1716384,0.00043707612],"about_ca_topic_score_codex":0.0010178941,"about_ca_topic_score_gemma":0.001035648,"teacher_disagreement_score":0.012704835,"about_ca_system_score_codex":0.0004346931,"about_ca_system_score_gemma":0.0009808681,"threshold_uncertainty_score":0.042501926},"labels":[],"label_agreement":null},{"id":"W2083696640","doi":"10.2481/dsj.009-026","title":"Data Management Activities of Canada's National Science Library - 2010 Update and Prospective","year":2010,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Mandate; Political science; Transformational leadership; Parliament; Research council; Library science; Engineering management; Public relations; Management; Business; Engineering; Computer science; Government (linguistics)","score_opus":0.0732256112792575,"score_gpt":0.34985363900726973,"score_spread":0.2766280277280122,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2083696640","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020356232,0.0406983,0.006484218,0.12678123,0.0046404963,0.0019273562,0.27236816,0.0075442605,0.5191998],"genre_scores_gemma":[0.08134204,0.06074526,0.04540927,0.012502587,0.0009132732,0.00084002304,0.3120472,0.0022965039,0.48390386],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.97071505,0.0014896195,0.002000112,0.0010926012,0.022047663,0.002655011],"domain_scores_gemma":[0.861553,0.004218617,0.0033563927,0.005933209,0.11121614,0.0137226805],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.019753339,0.0011092146,0.0011958736,0.021626266,0.012056732,0.017611336,0.0048841494,0.0020489653,0.01806989],"category_scores_gemma":[0.03430227,0.001296034,0.0008349642,0.046395186,0.0028233698,0.0049806247,0.005064605,0.0025820534,0.010459137],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":true,"about_ca_system_consensus":true,"study_design_scores_codex":[0.00007149452,0.00010440435,0.009634683,0.00064495986,0.000026691509,0.00009591444,0.0016268897,0.00029404613,0.0007151058,0.007921405,0.83469915,0.14416528],"study_design_scores_gemma":[0.00000499206,0.0000067165834,0.008887796,0.00016157076,0.000010442398,0.00003522733,0.0008795902,0.0001470176,0.00057014293,0.00015143308,0.98911977,0.000025292004],"about_ca_topic_score_codex":0.99018025,"about_ca_topic_score_gemma":0.9899283,"teacher_disagreement_score":0.9823887,"about_ca_system_score_codex":0.15853982,"about_ca_system_score_gemma":0.51695055,"threshold_uncertainty_score":0.97597456},"labels":[],"label_agreement":null},{"id":"W2123356220","doi":"10.2481/dsj.5.174","title":"The Crystallographic Information File (CIF)","year":2006,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McMaster University; Brockhouse Institute for Materials Research","funders":"","keywords":"Computer science; Scope (computer science); Transparency (behavior); Usability; Open data; Metadata; World Wide Web; Data science; Reuse; Data publishing; Data curation; Implementation; Software; Publishing; Political science; Software engineering; Computer security","score_opus":0.014380177954742022,"score_gpt":0.22447322685658352,"score_spread":0.2100930489018415,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2123356220","genre_codex":"dataset","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":[0.00091846345,0.003206887,0.006812198,0.0034207525,0.0070345155,0.0014795391,0.8959472,0.006926744,0.07425359],"genre_scores_gemma":[0.0076143667,0.00526221,0.016383717,0.0015665531,0.0017166116,0.004091658,0.9338439,0.0031067461,0.026414268],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.993912,0.0009705888,0.0005158784,0.0010341998,0.002386665,0.0011807773],"domain_scores_gemma":[0.98968226,0.0036933818,0.0011037505,0.0013881124,0.0030933113,0.0010392696],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.008213235,0.0071522566,0.02777502,0.010325862,0.009478618,0.010144536,0.02089088,0.009696944,0.884328],"category_scores_gemma":[0.021256533,0.005102168,0.0044982797,0.037304755,0.003007604,0.012310795,0.006646013,0.012316038,0.5554547],"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":[0.00014878908,0.000078679295,0.00017232946,0.0027129347,0.00005517149,0.000074407355,0.00008504299,0.000515536,0.00060950604,0.0062825778,0.97981226,0.009452823],"study_design_scores_gemma":[0.0029513454,0.00022671142,0.001571309,0.001207082,0.00021452204,0.00043437298,0.0004254331,0.0028003214,0.0013621703,0.014870842,0.97357225,0.00036360306],"about_ca_topic_score_codex":0.02201495,"about_ca_topic_score_gemma":0.025176987,"teacher_disagreement_score":0.884328,"about_ca_system_score_codex":0.009308064,"about_ca_system_score_gemma":0.019339098,"threshold_uncertainty_score":0.16499197},"labels":[],"label_agreement":null},{"id":"W2125590838","doi":"10.2481/dsj.4.106","title":"The long-term preservation of accurate and authentic digital data: the INTERPARES project","year":2005,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":46,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Scope (computer science); Computer science; Usability; Transparency (behavior); Implementation; Metadata; Digital preservation; Open data; Data science; Reuse; World Wide Web; Data publishing; Electronic publishing; Publishing; Open science; Data curation; Political science; Software engineering; Engineering; The Internet; Computer security","score_opus":0.13836117726891836,"score_gpt":0.3153712717204733,"score_spread":0.17701009445155494,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2125590838","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":[0.11134002,0.012591074,0.47663233,0.07940268,0.0026259173,0.0011704791,0.0027616378,0.006886721,0.3065891],"genre_scores_gemma":[0.45698363,0.009956756,0.42703828,0.0043156384,0.0014733465,0.0009781152,0.0067422213,0.00331329,0.08919878],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9617088,0.017093424,0.0018143497,0.0026836286,0.015104462,0.0015953827],"domain_scores_gemma":[0.8547962,0.038429778,0.006405311,0.07489886,0.018991355,0.0064785746],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.06982692,0.0006299266,0.0005880071,0.0031525495,0.0058092014,0.0149232065,0.0045680543,0.0034995738,0.006327204],"category_scores_gemma":[0.08304152,0.0005338699,0.0007219851,0.005604105,0.015342216,0.026223125,0.03255798,0.004356535,0.002476266],"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":[0.0002130494,0.00022724087,0.0028346237,0.00044186012,0.000047720754,0.0003069922,0.0080894325,0.0016396161,0.001338333,0.60490847,0.044279117,0.33567363],"study_design_scores_gemma":[0.0000581473,0.00028602357,0.0033870542,0.0008596717,0.000060786933,0.0010770803,0.012409811,0.00635315,0.011108085,0.22791082,0.73639596,0.00009337153],"about_ca_topic_score_codex":0.004108653,"about_ca_topic_score_gemma":0.0020759187,"teacher_disagreement_score":0.9850768,"about_ca_system_score_codex":0.0031600613,"about_ca_system_score_gemma":0.01730598,"threshold_uncertainty_score":0.36928445},"labels":[],"label_agreement":null},{"id":"W2132929844","doi":"10.2481/dsj.008-011","title":"Structured Query Translation in Peer to Peer Database Sharing Systems","year":2009,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Computer science; Scope (computer science); Usability; Transparency (behavior); Data sharing; World Wide Web; Metadata; Implementation; Open data; Reuse; Data science; Database; Software engineering","score_opus":0.07014835883326626,"score_gpt":0.3407390784558296,"score_spread":0.27059071962256337,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2132929844","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":[0.040846523,0.0012815011,0.93204755,0.0025579683,0.0004146836,0.0018677783,0.0008395397,0.007727283,0.012417199],"genre_scores_gemma":[0.4809901,0.00107666,0.49483663,0.0011944426,0.0004633707,0.0016082728,0.003073304,0.0015056219,0.0152516],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9578718,0.020955551,0.0044270502,0.00302987,0.011724606,0.001991146],"domain_scores_gemma":[0.92454505,0.04162369,0.002044756,0.021114772,0.009082134,0.0015895489],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.032106943,0.00093287276,0.0025487733,0.0024391378,0.0041135517,0.008625732,0.0056417696,0.004660389,0.009147139],"category_scores_gemma":[0.076873995,0.0015701498,0.0015965216,0.0057975305,0.005284288,0.014838799,0.013361566,0.0022773934,0.0055611],"study_design_candidate":"bench_or_experimental","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":[0.0022756031,0.0009550753,0.0040587853,0.0017005326,0.0003815055,0.0024923421,0.010883783,0.10572647,0.012498215,0.47438464,0.04516045,0.33948255],"study_design_scores_gemma":[0.0005292504,0.00039862556,0.00052611297,0.00018855004,0.00010117942,0.0006262182,0.0024781043,0.3987891,0.011949966,0.53894687,0.045278605,0.00018745867],"about_ca_topic_score_codex":0.0043505225,"about_ca_topic_score_gemma":0.0025076342,"teacher_disagreement_score":0.032106943,"about_ca_system_score_codex":0.0020434454,"about_ca_system_score_gemma":0.0050867633,"threshold_uncertainty_score":0.1697998},"labels":[],"label_agreement":null},{"id":"W2133007173","doi":"10.2481/dsj.4.127","title":"Multi-sensor data fusion for land vehicle attitude estimation using a fuzzy expert system","year":2005,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Inertial Sensor and Navigation","field":"Engineering","cited_by":31,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Gyroscope; Accelerometer; Kalman filter; Sensor fusion; Compass; Computer science; Inertial measurement unit; Fuzzy logic; Dead reckoning; Vibrating structure gyroscope; Artificial intelligence; Position (finance); Field (mathematics); Inertial navigation system; Global Positioning System; Computer vision; Control theory (sociology); Real-time computing; Inertial frame of reference; Engineering; Mathematics; Geography; Telecommunications; Aerospace engineering","score_opus":0.10433213707563253,"score_gpt":0.35159105032011034,"score_spread":0.2472589132444778,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2133007173","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":[0.06330845,0.00020120128,0.9343819,0.00011139667,0.00003492998,0.000060288534,0.000033643613,0.00045165504,0.0014166037],"genre_scores_gemma":[0.7812287,0.00012251668,0.21718186,0.00008032825,0.0000255424,0.000080534206,0.00005369536,0.0000102695985,0.0012165075],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99981123,0.00003863182,0.000019524161,0.000040333427,0.00007368004,0.000016593567],"domain_scores_gemma":[0.99970955,0.000099403565,0.000024584826,0.000024050341,0.00013035981,0.0000119906035],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005313078,0.0002696709,0.00040917858,0.00029362048,0.00031920997,0.00039934786,0.0003760149,0.0005474477,0.00086974027],"category_scores_gemma":[0.001118199,0.00014599894,0.00032896604,0.00018314694,0.00017141963,0.00047498985,0.0002704673,0.00035806338,0.00021711114],"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":[0.00047309085,0.00028228312,0.0025661904,0.00020509542,0.00013710556,0.0003131059,0.00030396474,0.26071274,0.09140826,0.0037788064,0.0018968073,0.6379225],"study_design_scores_gemma":[0.00002448701,0.00010978052,0.00098484,0.00000850573,0.00003066053,0.000048637983,0.000023233833,0.9870422,0.009745273,0.0010835285,0.0008868171,0.0000121114435],"about_ca_topic_score_codex":0.0035343056,"about_ca_topic_score_gemma":0.005396499,"teacher_disagreement_score":0.0035343056,"about_ca_system_score_codex":0.00032970766,"about_ca_system_score_gemma":0.00040386926,"threshold_uncertainty_score":0.007027507},"labels":[],"label_agreement":null},{"id":"W2141093030","doi":"10.2481/dsj.3.202","title":"Appraising digital records for long-term preservation","year":2004,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":18,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Scope (computer science); Computer science; Digital preservation; Usability; Transparency (behavior); Metadata; Implementation; Open data; World Wide Web; Data science; Reuse; Data curation; Data publishing; Electronic publishing; Publishing; Political science; Software engineering; The Internet; Computer security; Engineering","score_opus":0.13525818568589779,"score_gpt":0.30130213518249066,"score_spread":0.16604394949659287,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2141093030","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":[0.12489561,0.064722724,0.06938496,0.24438249,0.013786945,0.002310936,0.004905284,0.0012183202,0.47439274],"genre_scores_gemma":[0.83980215,0.029434834,0.066999026,0.009179274,0.0054630963,0.0008016879,0.0038499683,0.00083638675,0.04363348],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.8960173,0.046379954,0.012756541,0.002073284,0.04062507,0.002147928],"domain_scores_gemma":[0.40029258,0.28482875,0.03540327,0.07445246,0.19699447,0.008028423],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.11375282,0.00047440614,0.000787747,0.018479923,0.0056363493,0.0369664,0.003496414,0.003820792,0.017007649],"category_scores_gemma":[0.4708805,0.00045580184,0.00067257043,0.029013596,0.009744972,0.039056722,0.010817597,0.0033501273,0.003787803],"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":[0.0006033457,0.00012741445,0.024849627,0.0040373043,0.0001565838,0.0005220695,0.031757213,0.0009380378,0.00063114683,0.36572888,0.09898904,0.47165936],"study_design_scores_gemma":[0.000080678874,0.000386479,0.021002134,0.013303347,0.00029186686,0.0006851436,0.07437645,0.0020414628,0.0024226261,0.10241479,0.78283894,0.00015602467],"about_ca_topic_score_codex":0.0060216775,"about_ca_topic_score_gemma":0.007641024,"teacher_disagreement_score":0.11375282,"about_ca_system_score_codex":0.008637528,"about_ca_system_score_gemma":0.016317695,"threshold_uncertainty_score":0.60158956},"labels":[],"label_agreement":null},{"id":"W2154198926","doi":"10.2481/dsj.14-047","title":"Data-PE: A Framework for Evaluating Data Publication Policies at Scholarly Journals","year":2015,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"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; Scope (computer science); Usability; Transparency (behavior); Metadata; Open data; Data science; Data publishing; Implementation; Reuse; Publishing; World Wide Web; Political science; Software engineering; Engineering","score_opus":0.7646680752783171,"score_gpt":0.5850843913237225,"score_spread":0.17958368395459468,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2154198926","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"reporting","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"reporting","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.025918365,0.001571449,0.8178897,0.006866645,0.0008274465,0.010930698,0.03187674,0.01211628,0.09200258],"genre_scores_gemma":[0.11486608,0.00066723936,0.85026395,0.0008029277,0.00032721044,0.008598509,0.01833673,0.0012024784,0.004934924],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.7016701,0.16279696,0.048705503,0.016015736,0.063376404,0.0074352347],"domain_scores_gemma":[0.39113012,0.39291006,0.046017475,0.060888138,0.09610736,0.012946748],"candidate_categories":["metaresearch","open_science"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.31220078,0.0027562052,0.004003277,0.0652151,0.007161181,0.03232554,0.006991662,0.005364609,0.01639744],"category_scores_gemma":[0.52773124,0.0021303662,0.004241704,0.05426259,0.007134519,0.044925638,0.016467063,0.0052084713,0.0065966803],"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":[0.0012319383,0.0006810657,0.059361283,0.003358299,0.00089214725,0.00028980357,0.00653403,0.010529802,0.0012090983,0.58495206,0.059295226,0.27166525],"study_design_scores_gemma":[0.00052115845,0.00085238874,0.023688138,0.0040594554,0.00061706867,0.0005680083,0.011487069,0.08725765,0.007459357,0.5048236,0.35797873,0.00068737334],"about_ca_topic_score_codex":0.016158117,"about_ca_topic_score_gemma":0.011548233,"teacher_disagreement_score":0.9930083,"about_ca_system_score_codex":0.010661706,"about_ca_system_score_gemma":0.03589238,"threshold_uncertainty_score":0.84817886},"labels":[],"label_agreement":null},{"id":"W2585472482","doi":"10.5334/dsj-2017-003","title":"Legal and Ethical Issues around Incorporating Traditional Knowledge in Polar Data Infrastructures","year":2017,"lang":"en","type":"article","venue":"Data Science Journal","topic":"International Maritime Law Issues","field":"Environmental Science","cited_by":21,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University; University of Ottawa","funders":"","keywords":"Traditional knowledge; Acknowledgement; Knowledge management; Interoperability; Context (archaeology); Sociology of scientific knowledge; Knowledge sharing; Inclusion (mineral); Body of knowledge; Computer science; Engineering ethics; Indigenous; Sociology; Social science; World Wide Web; Engineering; Computer security; Geography","score_opus":0.07870823258626229,"score_gpt":0.3662956398552014,"score_spread":0.2875874072689391,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2585472482","genre_codex":"commentary","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":[0.07668497,0.0053921216,0.18964382,0.39023674,0.0017789761,0.0007778162,0.00022450626,0.00020356134,0.33505753],"genre_scores_gemma":[0.85329026,0.0029192679,0.07999864,0.042379286,0.0010461996,0.0011177114,0.00016555711,0.00018829157,0.01889469],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.818396,0.11711069,0.01205617,0.009376273,0.035928104,0.0071328743],"domain_scores_gemma":[0.73886126,0.18621215,0.012582999,0.037394814,0.020561947,0.004386812],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.13515507,0.0004725307,0.0010191911,0.0026258905,0.016374307,0.027982024,0.005717576,0.017502945,0.002687],"category_scores_gemma":[0.16230218,0.0011313667,0.001228649,0.003921771,0.094896175,0.03418604,0.020876594,0.01973644,0.0009816034],"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":[0.000012584806,0.000021232276,0.0006847049,0.00007019992,0.000007974475,0.00034452343,0.032416347,0.00027811513,0.00025027653,0.95566547,0.0024907717,0.0077577974],"study_design_scores_gemma":[0.00001945485,0.00003205749,0.0009921301,0.0013708499,0.00002302431,0.0006907563,0.035065923,0.0009419386,0.0010233641,0.8084858,0.15127285,0.00008178957],"about_ca_topic_score_codex":0.011654538,"about_ca_topic_score_gemma":0.01093738,"teacher_disagreement_score":0.9942824,"about_ca_system_score_codex":0.011597564,"about_ca_system_score_gemma":0.02612342,"threshold_uncertainty_score":0.7147769},"labels":[],"label_agreement":null},{"id":"W2610869427","doi":"10.5334/dsj-2017-024","title":"All or Nothing: The False Promise of Anonymity","year":2017,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Juvenile Diabetes Research Foundation","funders":"","keywords":"Anonymity; Data sharing; Computer science; Identification (biology); Data anonymization; Process (computing); Internet privacy; State (computer science); Computer security; Information privacy; Medicine","score_opus":0.8656067754571136,"score_gpt":0.6914985035282208,"score_spread":0.17410827192889278,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2610869427","genre_codex":"commentary","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":[0.0037974308,0.004345627,0.019963888,0.89375114,0.0069475216,0.000060691862,0.00013417547,0.00008827296,0.07091116],"genre_scores_gemma":[0.34051046,0.0067599257,0.015390139,0.59162015,0.018401077,0.0006559378,0.00017609917,0.00027597335,0.026210248],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.7982061,0.13966554,0.006279104,0.01061763,0.038675178,0.0065564364],"domain_scores_gemma":[0.7424797,0.20086367,0.008596747,0.031165177,0.011387352,0.005507454],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.10798475,0.0006178404,0.0017008381,0.0012414373,0.011434773,0.026428722,0.004096669,0.02369952,0.0047869203],"category_scores_gemma":[0.20521936,0.00088803034,0.0015706732,0.0015606374,0.095969066,0.037439384,0.014741109,0.036857996,0.002712166],"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":[0.000029315135,0.000008918996,0.00015690984,0.0000650198,0.000013773286,0.00009749985,0.0035930534,0.0000990805,0.000055446817,0.9409947,0.04601479,0.008871526],"study_design_scores_gemma":[0.000031661344,0.000034678073,0.00016645936,0.00041897173,0.000015736978,0.00030522162,0.0017805798,0.0002816788,0.00022444545,0.79699326,0.19969523,0.00005206818],"about_ca_topic_score_codex":0.002301231,"about_ca_topic_score_gemma":0.0014694828,"teacher_disagreement_score":0.10798475,"about_ca_system_score_codex":0.0062180976,"about_ca_system_score_gemma":0.017629072,"threshold_uncertainty_score":0.57108474},"labels":[],"label_agreement":null},{"id":"W2765696161","doi":"10.5334/dsj-2017-048","title":"The Northern Voice: Listening to Indigenous and Northern Perspectives on Management of Data in Canada","year":2017,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Indigenous Studies and Ecology","field":"Health Professions","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Indigenous; Interoperability; Active listening; Class (philosophy); The Internet; Geography; World Wide Web; Computer science; Sociology; Ecology","score_opus":0.08398072553422269,"score_gpt":0.4148932033926937,"score_spread":0.330912477858471,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2765696161","genre_codex":"commentary","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":[0.27169222,0.019434921,0.002666008,0.44801405,0.0059154094,0.00023832325,0.0010869835,0.00027232908,0.25067982],"genre_scores_gemma":[0.8486395,0.010545193,0.002180251,0.060015507,0.0005109308,0.00014025002,0.00035021186,0.00037308343,0.077245064],"study_design_codex":"qualitative","study_design_gemma":"qualitative","domain_scores_codex":[0.9809094,0.004634552,0.00046636752,0.0014307386,0.0051880246,0.007370921],"domain_scores_gemma":[0.97482806,0.007399664,0.000694151,0.00050725386,0.0066037197,0.009967269],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.013665298,0.00080016744,0.0009757855,0.002320538,0.08171867,0.0233549,0.0043744086,0.005318858,0.009804279],"category_scores_gemma":[0.017335199,0.0007760563,0.0007659493,0.0055806637,0.027005369,0.0071109575,0.013208166,0.00972347,0.0010257023],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":true,"about_ca_system_consensus":true,"study_design_scores_codex":[0.00005172187,0.000023357092,0.0029973888,0.00015553828,0.000018720684,0.001176006,0.90434366,0.0001166491,0.0005041998,0.009531697,0.06583686,0.015244208],"study_design_scores_gemma":[0.0000035033424,0.0000068175264,0.0022225797,0.00016316702,0.000009175235,0.000104727675,0.84315914,0.00005820452,0.000084214,0.0007794496,0.15336342,0.00004566278],"about_ca_topic_score_codex":0.99367297,"about_ca_topic_score_gemma":0.9961344,"teacher_disagreement_score":0.9956256,"about_ca_system_score_codex":0.17091264,"about_ca_system_score_gemma":0.28603423,"threshold_uncertainty_score":0.96162385},"labels":[],"label_agreement":null},{"id":"W2785284360","doi":"10.5334/dsj-2018-001","title":"Science Metadata Management, Interoperability and Data Citations of the National Institute of Polar Research, Japan","year":2018,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Metadata; Data management plan; Interoperability; Directory; Stewardship (theology); Data center; Meta Data Services; Open science; Data management; Data mapping; Library science; Data as a service; Research center; World Wide Web; Metadata management; Computer science; Metadata repository; Database; Political science; Business; Service (business)","score_opus":0.478091744716791,"score_gpt":0.5185561253619065,"score_spread":0.04046438064511554,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2785284360","genre_codex":"other","genre_gemma":"methods","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06762857,0.026039727,0.2507673,0.047784593,0.006219449,0.002427851,0.049395658,0.010411789,0.539325],"genre_scores_gemma":[0.32324794,0.022121716,0.36465463,0.0028807167,0.00215165,0.002834029,0.1087932,0.0032069702,0.17010918],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98862326,0.0025702338,0.0026415717,0.0014734087,0.004186978,0.00050457753],"domain_scores_gemma":[0.9654808,0.0031512263,0.0030076297,0.009505793,0.017050859,0.0018036942],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.017409295,0.00046950902,0.00073080213,0.013341904,0.0041752574,0.009539172,0.0017778752,0.00094646314,0.011495886],"category_scores_gemma":[0.036958933,0.000575735,0.00057296717,0.025821364,0.0017817104,0.011168985,0.007209508,0.0013309676,0.0064014033],"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":[0.00013948035,0.00011514875,0.026916638,0.0007879495,0.000091471535,0.00021426543,0.004551994,0.00094818475,0.0029121998,0.16576621,0.2037809,0.5937756],"study_design_scores_gemma":[0.000025296218,0.000028814073,0.016217828,0.00037042718,0.00010124405,0.00013283375,0.002134665,0.0026787058,0.0033401425,0.0313839,0.9435221,0.00006412128],"about_ca_topic_score_codex":0.059827145,"about_ca_topic_score_gemma":0.039780095,"teacher_disagreement_score":0.9904608,"about_ca_system_score_codex":0.0057391045,"about_ca_system_score_gemma":0.024727039,"threshold_uncertainty_score":0.11895782},"labels":[],"label_agreement":null},{"id":"W2891686592","doi":"10.5334/dsj-2018-020","title":"Using Classified and Unclassified Land Cover Data to Estimate the Footprint of Human Settlement","year":2018,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Impact of Light on Environment and Health","field":"Environmental Science","cited_by":45,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Esri (Canada)","funders":"","keywords":"Geography; Population; Footprint; Land cover; Cartography; Settlement (finance); Orthophoto; Ancillary data; Georeference; Census; Remote sensing; Land use; Computer science; Physical geography; Archaeology; Ecology; World Wide Web; Demography","score_opus":0.23874262984261077,"score_gpt":0.43696381670121914,"score_spread":0.19822118685860837,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2891686592","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":[0.6065686,0.0011127392,0.21050413,0.00055443076,0.0002506842,0.00090819126,0.093119696,0.0027920902,0.08418944],"genre_scores_gemma":[0.78467095,0.00065279414,0.153462,0.00013935009,0.00006264909,0.00057596026,0.05471718,0.00029394095,0.0054252455],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9990262,0.00018520765,0.00006891646,0.00016839113,0.00046892266,0.000082386214],"domain_scores_gemma":[0.99826354,0.00038367324,0.00033737143,0.0003356274,0.0005959752,0.000083886895],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000993353,0.00072416884,0.00045827118,0.006529471,0.0004195073,0.0017146122,0.0006832057,0.0004578173,0.004073131],"category_scores_gemma":[0.004426804,0.00024969774,0.00079023,0.0077879718,0.00032909372,0.0012790931,0.00096264767,0.00036632753,0.0025949627],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0001527748,0.0003890194,0.45316222,0.00042466717,0.0005081232,0.00035068186,0.0012658648,0.08538615,0.0034315898,0.007926283,0.023306947,0.42369574],"study_design_scores_gemma":[0.000055237422,0.00014920975,0.6148163,0.00042674827,0.00014126989,0.00053301634,0.0038215204,0.29140326,0.0052292603,0.016322387,0.06693761,0.00016413402],"about_ca_topic_score_codex":0.045528855,"about_ca_topic_score_gemma":0.07940274,"teacher_disagreement_score":0.045528855,"about_ca_system_score_codex":0.0012360835,"about_ca_system_score_gemma":0.0008966416,"threshold_uncertainty_score":0.09052771},"labels":[],"label_agreement":null},{"id":"W2908272686","doi":"10.5334/dsj-2019-001","title":"Understanding Human Mobility Patterns in a Developing Country Using Mobile Phone Data","year":2019,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":47,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"TRIPS architecture; Mobile phone; Phone; Computer science; Destinations; Centroid; Entropy (arrow of time); Statistics; Geography; Mathematics; Telecommunications","score_opus":0.33013340903309646,"score_gpt":0.4317011090950576,"score_spread":0.10156770006196114,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2908272686","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":[0.9835955,0.00017723162,0.009527948,0.00019198464,0.0000063040416,0.000035810255,0.0047961874,0.000042331427,0.0016267055],"genre_scores_gemma":[0.9909603,0.00027295633,0.0058758403,0.000019815043,0.000005504312,0.000027373591,0.0026373814,0.000003974871,0.00019668127],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99979573,0.00006920911,0.000020508887,0.00005353286,0.000029970448,0.000031077878],"domain_scores_gemma":[0.99936694,0.0003049817,0.0001286738,0.00007469594,0.00008900053,0.00003576292],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003274026,0.00028783348,0.0002570508,0.002065933,0.0002715841,0.0007188908,0.00024364275,0.00028291336,0.0005542535],"category_scores_gemma":[0.0021075627,0.0001537341,0.00030894688,0.0035014518,0.00023998116,0.0008188784,0.00055070163,0.0003419852,0.00024388911],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009762185,0.00009675542,0.9055762,0.00012278106,0.00014242594,0.0006312665,0.00097571913,0.03376721,0.002357573,0.001904626,0.0012242241,0.05310364],"study_design_scores_gemma":[0.000009187259,0.00011443007,0.84861827,0.00007422709,0.00008989269,0.00043667058,0.005079732,0.13730983,0.001280597,0.0021867955,0.004761817,0.00003849239],"about_ca_topic_score_codex":0.036184676,"about_ca_topic_score_gemma":0.038965516,"teacher_disagreement_score":0.036184676,"about_ca_system_score_codex":0.0004339497,"about_ca_system_score_gemma":0.00039149742,"threshold_uncertainty_score":0.07194811},"labels":[],"label_agreement":null},{"id":"W2955083888","doi":"10.5334/dsj-2019-027","title":"Developing a Model Guidelines Addressing Legal Impediments to Open Access to Publicly Funded Research Data in Malaysia","year":2019,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Data access; Computer science; Business; Knowledge management; Database","score_opus":0.8726138537959646,"score_gpt":0.6560074045530535,"score_spread":0.21660644924291106,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2955083888","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":[0.12226425,0.0024426836,0.55601853,0.07100996,0.0005218898,0.009929707,0.0035889496,0.0011747254,0.23304935],"genre_scores_gemma":[0.3761988,0.0024551034,0.59543157,0.002745833,0.000073404335,0.0060992385,0.0020886667,0.00012363515,0.014783679],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9555452,0.019769104,0.0060824393,0.0025940654,0.013284896,0.0027243008],"domain_scores_gemma":[0.9080643,0.04199493,0.0106866695,0.0058242385,0.030667767,0.0027621726],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.045470092,0.00078736426,0.00082003407,0.006250395,0.0030731163,0.012542127,0.003387281,0.004101526,0.0036621492],"category_scores_gemma":[0.081729814,0.0010253057,0.0015187517,0.0044178455,0.0034861944,0.01091832,0.0059532165,0.0042735576,0.0010234636],"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":[0.00008924114,0.0004767223,0.015494873,0.001191053,0.000062020255,0.0009010655,0.007178247,0.036765147,0.0011798664,0.8386798,0.017846864,0.080135114],"study_design_scores_gemma":[0.00024942736,0.0005073796,0.013274772,0.0056460034,0.00023172685,0.00097051705,0.02014584,0.11905024,0.0061087282,0.5163679,0.31701678,0.00043073477],"about_ca_topic_score_codex":0.037212554,"about_ca_topic_score_gemma":0.049561106,"teacher_disagreement_score":0.9966127,"about_ca_system_score_codex":0.014231072,"about_ca_system_score_gemma":0.0602402,"threshold_uncertainty_score":0.24047166},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science","scholarly_communication"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W3048186973","doi":"10.5334/dsj-2020-043","title":"The CARE Principles for Indigenous Data Governance","year":2020,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":1206,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Laurentian University","funders":"Rural Development Administration; European Commission","keywords":"Indigenous; Sovereignty; Data governance; Data sharing; Corporate governance; Stewardship (theology); Political science; Public administration; Public relations; Sociology; Law; Business; Medicine; Data quality","score_opus":0.3437787918939087,"score_gpt":0.4254753768765994,"score_spread":0.08169658498269072,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3048186973","genre_codex":"other","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":[0.013140596,0.0047161616,0.31671986,0.23557326,0.0038429173,0.00111544,0.0004670209,0.0004498579,0.423975],"genre_scores_gemma":[0.55001074,0.00512363,0.21382388,0.08621753,0.003742975,0.003482447,0.0006747054,0.00062801776,0.13629606],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9263034,0.034229714,0.006044667,0.008649219,0.0195547,0.0052183005],"domain_scores_gemma":[0.9402927,0.030963928,0.0027141778,0.012024037,0.011638252,0.0023669067],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.055874646,0.00063659274,0.0010010565,0.0019439823,0.012138624,0.017503316,0.0031032634,0.011444367,0.004959771],"category_scores_gemma":[0.08514049,0.00092095305,0.0016067437,0.0021621955,0.055989813,0.016334595,0.014263109,0.016202051,0.0020104724],"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":[0.0000032545913,0.000007987849,0.00011098944,0.000038777915,0.0000038699127,0.00003533555,0.004605488,0.00008951381,0.00004789036,0.984363,0.006181141,0.004512767],"study_design_scores_gemma":[0.000015304831,0.000017331833,0.00029652377,0.00039072754,0.00001608252,0.00017896244,0.0019394634,0.00040681055,0.00030051905,0.6414441,0.3549539,0.000040296876],"about_ca_topic_score_codex":0.020352785,"about_ca_topic_score_gemma":0.015809685,"teacher_disagreement_score":0.99689674,"about_ca_system_score_codex":0.009121288,"about_ca_system_score_gemma":0.032686558,"threshold_uncertainty_score":0.29549688},"labels":[],"label_agreement":null},{"id":"W3111376437","doi":"10.5334/dsj-2020-047","title":"39 Hints to Facilitate the Use of Semantics for Data on Agriculture and Nutrition","year":2020,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Agriculture and Agri-Food Canada; Rural Development Administration; Agence Nationale de la Recherche; Ministry of Agriculture of the People's Republic of China; Department for International Development; Bill and Melinda Gates Foundation","keywords":"Computer science; Semantic interoperability; Interoperability; Conceptualization; Linked data; Knowledge management; Data science; Semantics (computer science); World Wide Web; Data sharing; Standardization; Semantic Web","score_opus":0.29656193485514376,"score_gpt":0.34992526355676146,"score_spread":0.053363328701617696,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3111376437","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":[0.02093625,0.0012818517,0.90299493,0.030079639,0.00034743815,0.0017061528,0.005110662,0.0055394187,0.032003712],"genre_scores_gemma":[0.05460829,0.000724679,0.93190116,0.0024699501,0.000074507225,0.0006977443,0.0060099866,0.0009165869,0.0025971239],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9369188,0.03146348,0.014995774,0.0032604032,0.011978676,0.0013828344],"domain_scores_gemma":[0.8023643,0.121781625,0.0071572065,0.04042386,0.02545759,0.0028153826],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.084525734,0.001227408,0.0011606202,0.008987781,0.0025436287,0.010833569,0.0032143309,0.004687839,0.0049406476],"category_scores_gemma":[0.12998454,0.0017901405,0.0027164696,0.00771523,0.0058262013,0.024211813,0.014625782,0.0053029456,0.0029748136],"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":[0.00027551217,0.0001721869,0.008395811,0.002604551,0.00010542783,0.001401869,0.023798136,0.0045676497,0.016004065,0.7778568,0.025721394,0.13909657],"study_design_scores_gemma":[0.00003892743,0.000090799724,0.0024094526,0.0021670593,0.000075187,0.0009278917,0.0051509123,0.007037683,0.0091742,0.20733798,0.7654444,0.00014543238],"about_ca_topic_score_codex":0.008073497,"about_ca_topic_score_gemma":0.011843875,"teacher_disagreement_score":0.084525734,"about_ca_system_score_codex":0.0041249283,"about_ca_system_score_gemma":0.009599124,"threshold_uncertainty_score":0.44702017},"labels":[],"label_agreement":null},{"id":"W3127642832","doi":"10.5334/dsj-2021-007","title":"Stewardship Maturity Assessment Tools for Modernization of Climate Data Management","year":2021,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Climate variability and models","field":"Environmental Science","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Environment and Climate Change Canada","funders":"Environment and Climate Change Canada; Centrum fÖr Personcentrerad Vård; National Centers for Environmental Information; National Oceanic and Atmospheric Administration; Grains Research and Development Corporation; National Aeronautics and Space Administration","keywords":"Stewardship (theology); Data management; Maturity (psychological); Scope (computer science); Data quality; Computer science; Process (computing); Quality (philosophy); Usability; Process management; Environmental resource management; Business; Data science; Database; Environmental science; Political science","score_opus":0.13297788638193198,"score_gpt":0.36953920862532325,"score_spread":0.23656132224339127,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3127642832","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":[0.034217972,0.0009922758,0.88164425,0.00739603,0.00031818243,0.0048186495,0.0053321137,0.016056284,0.049224153],"genre_scores_gemma":[0.075011455,0.00047774462,0.9125686,0.00031413115,0.00005650412,0.00229364,0.005752932,0.00078360835,0.0027413538],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9556321,0.015340823,0.0096280845,0.0029153787,0.015080843,0.0014027589],"domain_scores_gemma":[0.7991746,0.087541096,0.022386327,0.02474849,0.06061643,0.005533036],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.067028105,0.0018181284,0.0008809016,0.019496879,0.00256715,0.011780934,0.0027273707,0.0015046771,0.008000716],"category_scores_gemma":[0.19331266,0.00108657,0.0023101203,0.011396367,0.0015706179,0.016482176,0.009145082,0.004929029,0.0028829954],"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":[0.0001805907,0.00082485325,0.03693908,0.0020727676,0.00023122344,0.00025073087,0.013606386,0.010577372,0.004264905,0.14372325,0.0655937,0.7217353],"study_design_scores_gemma":[0.00016556216,0.0006451487,0.04811697,0.0053580813,0.00025592343,0.00080747827,0.017707234,0.119171955,0.015906883,0.2615226,0.5296992,0.00064296235],"about_ca_topic_score_codex":0.0061201206,"about_ca_topic_score_gemma":0.008141815,"teacher_disagreement_score":0.067028105,"about_ca_system_score_codex":0.006069065,"about_ca_system_score_gemma":0.0122671025,"threshold_uncertainty_score":0.3544827},"labels":[],"label_agreement":null},{"id":"W4301396488","doi":"10.5334/dsj-2022-017","title":"A Survey on Publicly Available Open Datasets Derived From Electronic Health Records (EHRs) of Patients with Neuroblastoma","year":2022,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Neuroblastoma Research and Treatments","field":"Medicine","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Neuroblastoma; Computer science; Health records; Open science; License; The Internet; Electronic health record; Open data; World Wide Web; Internet privacy; Medicine; Information retrieval; Data mining; Political science; Mathematics; Statistics; Health care","score_opus":0.07305559528067233,"score_gpt":0.3504970068041415,"score_spread":0.27744141152346913,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4301396488","genre_codex":"dataset","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":[0.10244699,0.017764082,0.005774887,0.010402432,0.00051519886,0.0003713639,0.8513907,0.0009353296,0.010399079],"genre_scores_gemma":[0.08552543,0.008559856,0.0073492927,0.003592142,0.00027250784,0.0008958807,0.89253575,0.00022177181,0.001047476],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.9803949,0.005603066,0.004827285,0.0022522719,0.0060142432,0.0009082492],"domain_scores_gemma":[0.9168954,0.048961394,0.012480886,0.006739492,0.012790139,0.0021327105],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.010214086,0.0006292441,0.0009444067,0.008501466,0.000736537,0.0022419856,0.0014861313,0.0013646354,0.0031232387],"category_scores_gemma":[0.06579199,0.0003146983,0.0010362383,0.015663758,0.00060855603,0.0030394024,0.0037584193,0.00096511486,0.0016585814],"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":[0.00097261486,0.00020320005,0.3698104,0.010269303,0.0007011017,0.00043078617,0.0019072355,0.0013101127,0.0019689866,0.0065938234,0.43834662,0.16748573],"study_design_scores_gemma":[0.00008956951,0.00011690805,0.42255384,0.0047224453,0.00035440028,0.0014055448,0.0036305594,0.001526093,0.0020329498,0.0035194948,0.5598889,0.00015931293],"about_ca_topic_score_codex":0.0060037975,"about_ca_topic_score_gemma":0.008588018,"teacher_disagreement_score":0.9985139,"about_ca_system_score_codex":0.0012749704,"about_ca_system_score_gemma":0.0030222752,"threshold_uncertainty_score":0.0540179},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4380538161","doi":"10.5334/dsj-2023-015","title":"Legal Regulation of State Electronic Services: Relevant Issues and Ways of Improvement","year":2023,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Corruption and Economic Development","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Government (linguistics); Business; Politics; Public relations; Public sector; Quality (philosophy); State (computer science); Order (exchange); Service (business); Public administration; The Republic; Political science; Marketing; Computer science; Law; Finance","score_opus":0.05167142070269107,"score_gpt":0.3291735194967818,"score_spread":0.2775020987940907,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4380538161","genre_codex":"commentary","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":[0.0065396614,0.0455323,0.031844422,0.85101205,0.0027089384,0.00046621202,0.00015167816,0.00035940108,0.06138536],"genre_scores_gemma":[0.6041532,0.083332136,0.13496318,0.14088011,0.0062484182,0.0020517965,0.0004983486,0.00035825316,0.027514538],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.92660296,0.04503669,0.004664769,0.004023843,0.012788344,0.0068833535],"domain_scores_gemma":[0.81860834,0.10353035,0.010000056,0.017578736,0.042692747,0.0075896843],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.09442486,0.0013005196,0.0020179593,0.007858914,0.010390717,0.041803323,0.0077422243,0.022231687,0.010531416],"category_scores_gemma":[0.13269648,0.0010115355,0.0022458099,0.010162902,0.04900141,0.059137728,0.009824604,0.021301707,0.0018326732],"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":[0.000031411608,0.00025153876,0.0019857734,0.0008546576,0.000035913436,0.00020116095,0.006993507,0.0010291502,0.00017638311,0.90562856,0.023908593,0.05890341],"study_design_scores_gemma":[0.000072116374,0.00018007985,0.0041269385,0.0072881524,0.00008435627,0.0003737759,0.04525118,0.0029718152,0.000732157,0.5639488,0.37473372,0.00023689907],"about_ca_topic_score_codex":0.034038976,"about_ca_topic_score_gemma":0.021833371,"teacher_disagreement_score":0.09442486,"about_ca_system_score_codex":0.024828458,"about_ca_system_score_gemma":0.08330019,"threshold_uncertainty_score":0.49937236},"labels":[],"label_agreement":null},{"id":"W4382141301","doi":"10.5334/dsj-2023-018","title":"Polar Data Forum IV – An Ocean of Opportunities","year":2023,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University; Inuvialuit Regional Corporation; Nunavut Research Institute; University of Calgary","funders":"LifeWatch – Niclas Öberg Foundation; Belgian Federal Science Policy Office","keywords":"Interoperability; Polar; Data management; Computer science; Library science; World Wide Web; Data science; Database","score_opus":0.5412294111931125,"score_gpt":0.46077708226501984,"score_spread":0.08045232892809268,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4382141301","genre_codex":"commentary","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":[0.01414086,0.021050038,0.02324745,0.69722384,0.056267265,0.00044888907,0.0018250598,0.0008919664,0.18490471],"genre_scores_gemma":[0.2634888,0.028927004,0.0447496,0.23312332,0.041124877,0.0022800302,0.0078751445,0.004471741,0.37395954],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.95449185,0.019730944,0.0014190798,0.0029276786,0.014490207,0.0069402447],"domain_scores_gemma":[0.88486445,0.05135402,0.0046039172,0.012395915,0.012431354,0.03435029],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.09554195,0.0007804437,0.0008488545,0.0040288484,0.014267468,0.039402977,0.0026408087,0.012179982,0.034403395],"category_scores_gemma":[0.08819683,0.000963646,0.0011182507,0.0054231486,0.010489314,0.034042355,0.036711078,0.016032984,0.008753772],"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":[0.00019161077,0.0001245194,0.0017687895,0.00047425818,0.000015376041,0.00068289036,0.011862302,0.0003232724,0.0009414383,0.42964944,0.45839557,0.09557054],"study_design_scores_gemma":[0.0000072668477,0.000014799197,0.00026676015,0.00021250478,0.0000010191462,0.00006493997,0.0025301669,0.000039418293,0.00010679095,0.013697833,0.98304015,0.000018411347],"about_ca_topic_score_codex":0.004687643,"about_ca_topic_score_gemma":0.0043494366,"teacher_disagreement_score":0.09554195,"about_ca_system_score_codex":0.0072667534,"about_ca_system_score_gemma":0.020402767,"threshold_uncertainty_score":0.50528014},"labels":[{"model":"gemma","categories":["scholarly_communication"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"not_applicable","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4386830764","doi":"10.5334/dsj-2023-035","title":"Umbrella Data Management Plans to Integrate FAIR Data: Lessons From the ISIDORe and BY-COVID Consortia for Pandemic Preparedness","year":2023,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer","funders":"Biotechnology and Biological Sciences Research Council","keywords":"Preparedness; Data management; Computer science; Multidisciplinary approach; Plan (archaeology); Download; Process (computing); Data management plan; Data science; Process management; Knowledge management; Engineering management; World Wide Web; Business; Engineering; Political science; Geography; Database","score_opus":0.43034568833391706,"score_gpt":0.48834636250226554,"score_spread":0.058000674168348476,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4386830764","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.013593497,0.004080885,0.55167115,0.36067122,0.0039092596,0.002617746,0.0017507434,0.0032384417,0.05846701],"genre_scores_gemma":[0.05936563,0.0019945945,0.8888759,0.022858659,0.0010506524,0.0023331714,0.004062475,0.0019117341,0.017547186],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","domain_scores_codex":[0.76529217,0.12479608,0.026450597,0.014907302,0.058572818,0.009981086],"domain_scores_gemma":[0.504484,0.18143174,0.012480418,0.19348618,0.080146484,0.02797123],"candidate_categories":["metaresearch","open_science"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.48550367,0.0016429588,0.0018081985,0.007506551,0.015122446,0.060581878,0.01693589,0.015645757,0.0070758355],"category_scores_gemma":[0.35521564,0.0026188493,0.0043473593,0.012535607,0.03126967,0.06660323,0.03338017,0.026089603,0.003295065],"study_design_candidate":"qualitative","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":[0.00009111515,0.00022393055,0.002087382,0.00038560922,0.00006440939,0.00028199158,0.009332332,0.0029758597,0.00071278005,0.87653524,0.052421864,0.05488738],"study_design_scores_gemma":[0.000104203515,0.0001451129,0.0016060306,0.0016447548,0.000043169104,0.0003428887,0.0060958597,0.0056121503,0.0015523643,0.38656014,0.59606844,0.00022483438],"about_ca_topic_score_codex":0.040897354,"about_ca_topic_score_gemma":0.020341108,"teacher_disagreement_score":0.9830641,"about_ca_system_score_codex":0.02448364,"about_ca_system_score_gemma":0.11329206,"threshold_uncertainty_score":0.6344656},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W4401463404","doi":"10.5334/dsj-2024-042","title":"Decentralised Semantics: A Semantic Engine User Perspective","year":2024,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"Canada First Research Excellence Fund","keywords":"Computer science; Perspective (graphical); Semantics (computer science); Information retrieval; Semantic computing; World Wide Web; Semantic Web; Programming language; Artificial intelligence","score_opus":0.040953870980605905,"score_gpt":0.3302790571581251,"score_spread":0.28932518617751923,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401463404","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":[0.013509474,0.0006711522,0.9216243,0.007822931,0.00025132316,0.00019702945,0.0004220857,0.0047689616,0.050732736],"genre_scores_gemma":[0.38332123,0.0014641403,0.58282024,0.0028905945,0.00043464924,0.00041360568,0.0013683479,0.0046613645,0.022625774],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","domain_scores_codex":[0.98461354,0.006330896,0.0017654863,0.001914406,0.00430879,0.001066829],"domain_scores_gemma":[0.9778541,0.008614488,0.0006126818,0.00856657,0.003326021,0.0010261547],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.034224972,0.000986331,0.0018263785,0.0029558996,0.0022913096,0.01966344,0.0049026506,0.0050690663,0.0075570657],"category_scores_gemma":[0.028262032,0.0016021392,0.002687666,0.0031901319,0.012114923,0.04015477,0.011834325,0.007336187,0.0033670606],"study_design_candidate":"qualitative","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":[0.00009627085,0.00004728219,0.0006321993,0.00014510633,0.000035636625,0.0002544081,0.005409517,0.0028203155,0.0015583568,0.9679331,0.00400163,0.017066045],"study_design_scores_gemma":[0.000052349173,0.000066241344,0.00027146054,0.00026764127,0.00006668923,0.0005744269,0.0024212082,0.032831077,0.004960718,0.71036655,0.24804537,0.000076366305],"about_ca_topic_score_codex":0.0057954118,"about_ca_topic_score_gemma":0.0042613014,"teacher_disagreement_score":0.034224972,"about_ca_system_score_codex":0.004605157,"about_ca_system_score_gemma":0.005365734,"threshold_uncertainty_score":0.18100113},"labels":[],"label_agreement":null},{"id":"W4402883940","doi":"10.5334/dsj-2024-046","title":"Knowledge Infrastructures Are Growing Up: The Case for Institutional (Data) Repositories 10 Years After the Holdren Memo","year":2024,"lang":"en","type":"article","venue":"Data Science Journal","topic":"Research Data Management Practices","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"University at Buffalo; University of Minnesota","keywords":"Computer science; Business; World Wide Web; Data science","score_opus":0.13277524692396442,"score_gpt":0.407377605795062,"score_spread":0.2746023588710976,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4402883940","genre_codex":"commentary","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":[0.25665483,0.008104803,0.009502914,0.5450062,0.0011231906,0.00012677802,0.0004955369,0.00042615327,0.17855954],"genre_scores_gemma":[0.93411636,0.0050050635,0.009057254,0.019091014,0.0006283389,0.00013888616,0.00035121062,0.00041528448,0.03119655],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.98072594,0.008452728,0.00088583824,0.0019915935,0.004401703,0.0035421639],"domain_scores_gemma":[0.9321728,0.034061335,0.0064285356,0.010446281,0.009386616,0.0075044893],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.029384075,0.00032246386,0.000439766,0.0040045097,0.020300005,0.038124237,0.0037218449,0.010247103,0.009593083],"category_scores_gemma":[0.057016533,0.0009032578,0.0010206681,0.0072936886,0.019677553,0.069380306,0.020440556,0.00915975,0.0013736901],"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":[0.0001120831,0.000100684156,0.016927717,0.0001815749,0.000034151875,0.0031931929,0.1057165,0.00050438143,0.00039725655,0.7415113,0.05372396,0.07759727],"study_design_scores_gemma":[0.000028744022,0.00006115268,0.008210943,0.0009725318,0.000038061855,0.0014169862,0.12438136,0.001062231,0.0007994581,0.07780516,0.7851158,0.00010755766],"about_ca_topic_score_codex":0.061359785,"about_ca_topic_score_gemma":0.07739386,"teacher_disagreement_score":0.96187574,"about_ca_system_score_codex":0.018475369,"about_ca_system_score_gemma":0.028907124,"threshold_uncertainty_score":0.15539968},"labels":[],"label_agreement":null}]}