{"meta":{"query_hash":"fb1763f2f70b","filters":{"venue":"Journal of Data Science"},"cohort_total":25,"direct_labels_cover":0,"predictions_cover":25,"exported":25,"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/fb1763f2f70b","api":"https://metacan.xera.ac/api/v1/cohort?venue=Journal+of+Data+Science"},"results":[{"id":"W2181556329","doi":"10.6339/jds.2004.02(1).135","title":"The Poisson Inverse Gaussian Regression Model in the Analysis of Clustered Counts Data","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":49,"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 Guelph; Western University","funders":"","keywords":"Poisson regression; Statistics; Mathematics; Covariate; Poisson distribution; Negative binomial distribution; Count data; Inverse Gaussian distribution; Regression analysis; Generalized linear model; Regression; Distribution (mathematics); Population","score_opus":0.14271075028666894,"score_gpt":0.3991228810900525,"score_spread":0.25641213080338354,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2181556329","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":[0.0050805924,0.00042445905,0.99275535,0.00032067622,0.00005275923,0.00007718103,0.00017365733,0.0002285473,0.0008867633],"genre_scores_gemma":[0.19196253,0.0019786665,0.7957805,0.00065582985,0.00034179832,0.0012233716,0.0010628623,0.00049333693,0.0065010367],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.975911,0.017660724,0.00057154614,0.0025113889,0.0027362998,0.0006090143],"domain_scores_gemma":[0.9609412,0.031961888,0.0020661973,0.0028495158,0.0018352041,0.00034597455],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.037862103,0.0010639247,0.0020403108,0.0022383651,0.0010145009,0.0023310836,0.00549582,0.0023710062,0.0034374031],"category_scores_gemma":[0.08831883,0.0013364963,0.0023916797,0.00534075,0.002759085,0.003372589,0.002410769,0.0035873658,0.0014448403],"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":[0.00020488446,0.000130774,0.015478273,0.0005286464,0.0006185742,0.000684121,0.0015428081,0.25615317,0.0012711495,0.6296114,0.008351541,0.08542468],"study_design_scores_gemma":[0.000059161695,0.00012804363,0.0034998006,0.00014878286,0.00013656325,0.00033355874,0.0001964893,0.7338882,0.00060023344,0.2495908,0.01131738,0.00010102506],"about_ca_topic_score_codex":0.031861804,"about_ca_topic_score_gemma":0.023768248,"teacher_disagreement_score":0.037862103,"about_ca_system_score_codex":0.0023069233,"about_ca_system_score_gemma":0.0043526134,"threshold_uncertainty_score":0.20023638},"labels":[],"label_agreement":null},{"id":"W2181591901","doi":"10.6339/jds.2008.06(3).508","title":"Evaluating Aortic Stenosis Using the Archimedean Copula Methodology","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":16,"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 Northern British Columbia","funders":"","keywords":"Copula (linguistics); Stenosis; China; Statistics; Econometrics; Actuarial science; Mathematics; Medicine; Cardiology; Internal medicine; Business; Geography","score_opus":0.8478728438504952,"score_gpt":0.6353580343979097,"score_spread":0.2125148094525855,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2181591901","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":[0.09643006,0.0042129424,0.89151776,0.0010174545,0.00036346418,0.0002490651,0.0008241134,0.0005603131,0.004824944],"genre_scores_gemma":[0.5784276,0.0029551554,0.41335616,0.00017617816,0.0003272296,0.00023349343,0.00092177565,0.00022075152,0.003381706],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99741507,0.0013109752,0.00019544197,0.00023766432,0.0007279835,0.00011271568],"domain_scores_gemma":[0.9902451,0.00616023,0.000829985,0.00054016244,0.0018618825,0.00036273315],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051162015,0.0010404924,0.001307817,0.002975925,0.00045304492,0.0018270323,0.00072087103,0.0006893609,0.0031743692],"category_scores_gemma":[0.026241994,0.000275387,0.0011066456,0.0020077836,0.00051600416,0.0010290575,0.00097326963,0.0011538677,0.00053237414],"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":[0.00067670265,0.00030495969,0.09931121,0.0008249703,0.0007594491,0.0019437864,0.0003051503,0.26432073,0.0062561417,0.029019294,0.015634455,0.58064324],"study_design_scores_gemma":[0.000036137528,0.00054372457,0.028026486,0.00013369093,0.00017351305,0.0012341843,0.0002359755,0.94165665,0.003944293,0.019613575,0.0043406053,0.00006112954],"about_ca_topic_score_codex":0.003916795,"about_ca_topic_score_gemma":0.003281225,"teacher_disagreement_score":0.0051162015,"about_ca_system_score_codex":0.00044793138,"about_ca_system_score_gemma":0.0015943424,"threshold_uncertainty_score":0.02705735},"labels":[],"label_agreement":null},{"id":"W2182961570","doi":"10.6339/jds.2004.02(1).142","title":"A Two-Stage Bayesian Model for Predicting Winners in Major League Baseball","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Sports Analytics and Performance","field":"Economics, Econometrics and Finance","cited_by":43,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada; Korea Science and Engineering Foundation","keywords":"League; Markov chain Monte Carlo; Bayesian probability; Bayesian inference; Computer science; Econometrics; Field (mathematics); Markov chain; Inference; Artificial intelligence; Operations research; Machine learning; Mathematics","score_opus":0.09328473143994379,"score_gpt":0.3027483287623079,"score_spread":0.2094635973223641,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2182961570","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":[0.37687126,0.0011866102,0.6050221,0.0035327424,0.00016826644,0.0005146361,0.005658652,0.00093625806,0.0061095036],"genre_scores_gemma":[0.9054903,0.0008633287,0.07318306,0.00035119776,0.00024528828,0.00075753895,0.005772753,0.000107689906,0.0132288905],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9972161,0.0013964522,0.00013658719,0.00056745706,0.0002482443,0.00043520393],"domain_scores_gemma":[0.98322797,0.013559049,0.0012080404,0.00046465744,0.0009552582,0.0005849727],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011523505,0.0013372966,0.0022632375,0.0022295064,0.0009904173,0.00280014,0.0045788237,0.003568847,0.008001757],"category_scores_gemma":[0.020566154,0.0019523918,0.00146244,0.0022063125,0.001337952,0.0033052897,0.0015754497,0.0034860955,0.0016655062],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009508852,0.0004778069,0.030419663,0.00013048481,0.00026893764,0.00028194964,0.00045424345,0.87158716,0.0006595006,0.053015333,0.005577726,0.03617632],"study_design_scores_gemma":[0.000073451585,0.00005397902,0.0026770732,0.000019577985,0.000036903322,0.000027902299,0.000040437095,0.9821293,0.00007685696,0.014309437,0.0005257505,0.000029205072],"about_ca_topic_score_codex":0.057445146,"about_ca_topic_score_gemma":0.054716468,"teacher_disagreement_score":0.057445146,"about_ca_system_score_codex":0.0023381605,"about_ca_system_score_gemma":0.002514057,"threshold_uncertainty_score":0.11422151},"labels":[],"label_agreement":null},{"id":"W2185241920","doi":"10.6339/jds.201101_09(1).0005","title":"Estimating Transmissibility of Seasonal Influenza Virus by Surveillance Data","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Influenza Virus Research Studies","field":"Medicine","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Public Health Agency of Canada","funders":"Public Health Agency; Public Health Agency of Canada","keywords":"Transmissibility (structural dynamics); Outbreak; Estimation; Transmission (telecommunications); Flu season; Seasonal influenza; Virus; Vaccination; Influenza A virus; Virology; Statistics; Biology; Computer science; Medicine; Mathematics; Coronavirus disease 2019 (COVID-19); Disease; Engineering; Infectious disease (medical specialty)","score_opus":0.23217288171701067,"score_gpt":0.4759016103250826,"score_spread":0.24372872860807193,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2185241920","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":[0.8150938,0.0007995598,0.1806151,0.0002671766,0.000025394449,0.0000803129,0.0020127804,0.00015833057,0.00094763807],"genre_scores_gemma":[0.97800523,0.00042132186,0.019086367,0.000015286043,0.000030462847,0.000035551286,0.002197243,0.000010006818,0.00019852795],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99819475,0.00087456655,0.00020300344,0.00035629794,0.00028010493,0.0000912952],"domain_scores_gemma":[0.987788,0.00871477,0.0013156721,0.0013495724,0.00065465143,0.00017728758],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003666552,0.0004933465,0.00051931274,0.0030446998,0.0002310685,0.0009787313,0.00071454607,0.000590693,0.0003895201],"category_scores_gemma":[0.021142365,0.0004220651,0.0007488756,0.0017328754,0.00035810244,0.0021701592,0.00094311277,0.00054860994,0.00012417551],"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":[0.00049554004,0.00017410543,0.52634317,0.0003072997,0.00068604655,0.00046084184,0.0004404191,0.3307506,0.009442069,0.006207515,0.0006479062,0.12404449],"study_design_scores_gemma":[0.000013375145,0.0001906817,0.08100807,0.000044870158,0.00011137736,0.00029282275,0.00019933442,0.9065078,0.0034013235,0.0074492656,0.00074838253,0.00003268171],"about_ca_topic_score_codex":0.004289601,"about_ca_topic_score_gemma":0.0029188085,"teacher_disagreement_score":0.004289601,"about_ca_system_score_codex":0.0005388059,"about_ca_system_score_gemma":0.00034833368,"threshold_uncertainty_score":0.019390821},"labels":[],"label_agreement":null},{"id":"W2185363721","doi":"10.6339/jds.2010.08(1).585","title":"A Bayesian Estimator of the Intracluster Correlation Coefficient from Correlated Binary Responses","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Water Quality and Resources Studies","field":"Environmental 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":"Western University","funders":"","keywords":"Mathematics; Estimator; Statistics; Beta-binomial distribution; Negative binomial distribution; Poisson distribution","score_opus":0.05897463788431407,"score_gpt":0.2840856813768295,"score_spread":0.22511104349251543,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2185363721","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":[0.009621713,0.00059211405,0.9878051,0.00019192812,0.000041638836,0.00014050576,0.0002956678,0.00027441513,0.0010368824],"genre_scores_gemma":[0.2683797,0.0013195917,0.7241584,0.0004965596,0.00025867537,0.0012849711,0.0023428933,0.00018258716,0.0015767108],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9852107,0.009045621,0.0006871626,0.0021649925,0.0025280616,0.0003635262],"domain_scores_gemma":[0.92847115,0.058694586,0.0040061814,0.0034253062,0.0049380753,0.00046466492],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02198046,0.0009092565,0.0022912454,0.0038612373,0.00080067205,0.001637387,0.002753301,0.002147226,0.0037722513],"category_scores_gemma":[0.12370949,0.00065027276,0.0017905341,0.002984361,0.0016414062,0.0023994162,0.0018140933,0.003148533,0.0011109012],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000749093,0.00036136046,0.038964733,0.0014620641,0.0016444016,0.00052771287,0.0012525694,0.14239776,0.0067746337,0.1615492,0.011800593,0.6325159],"study_design_scores_gemma":[0.00026263733,0.00043039673,0.028767094,0.00080371276,0.0005608773,0.0010111931,0.00033263557,0.7344226,0.0040151933,0.21821432,0.010903008,0.00027629276],"about_ca_topic_score_codex":0.003149094,"about_ca_topic_score_gemma":0.0023482465,"teacher_disagreement_score":0.02198046,"about_ca_system_score_codex":0.00082455954,"about_ca_system_score_gemma":0.0021042589,"threshold_uncertainty_score":0.11624521},"labels":[],"label_agreement":null},{"id":"W2188439980","doi":"10.6339/jds.201404_12(2).0008","title":"SENSITIVITY ANALYSIS OF BAYES FACTOR FOR CATEGORICAL DATA WITH EMPHASIS ON SPARSE MULTINOMIAL DATA","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Sensory Analysis and Statistical Methods","field":"Agricultural and Biological Sciences","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":"Government of Newfoundland and Labrador","funders":"","keywords":"Categorical variable; Multinomial distribution; Contingency table; Bayes factor; Prior probability; Dirichlet distribution; Computer science; Bayes' theorem; Bayesian probability; Sensitivity (control systems); Context (archaeology); Econometrics; Statistics; Consistency (knowledge bases); Factor analysis; Data mining; Mathematics; Artificial intelligence; Machine learning","score_opus":0.2991029132364069,"score_gpt":0.4089341009435077,"score_spread":0.10983118770710076,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2188439980","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":[0.08319892,0.0042374507,0.9067301,0.0012968713,0.00031481712,0.0006245307,0.00038800074,0.00037481007,0.002834431],"genre_scores_gemma":[0.8185487,0.001057798,0.17733552,0.0006171935,0.00028651065,0.0008235544,0.00044555968,0.0002058557,0.0006793582],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.8619335,0.11706657,0.003446499,0.008240194,0.008225226,0.0010880475],"domain_scores_gemma":[0.20539136,0.7663671,0.0073815235,0.014744241,0.005489826,0.0006258885],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.14779235,0.0013866092,0.0029459638,0.0037972503,0.0017628303,0.003571376,0.0023622024,0.0032029673,0.0031339223],"category_scores_gemma":[0.5391139,0.0007992598,0.0031996334,0.0024976197,0.003648782,0.004704412,0.0036425155,0.0043305377,0.0003171415],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0046742572,0.0004020953,0.07907245,0.0037791934,0.0048684264,0.0021606735,0.003304628,0.50483936,0.006767096,0.14553134,0.003977795,0.24062261],"study_design_scores_gemma":[0.00021240689,0.001334022,0.015477139,0.00094455533,0.0013087845,0.0022761533,0.0009814742,0.70171344,0.007331107,0.26216158,0.005924296,0.00033501332],"about_ca_topic_score_codex":0.0028683913,"about_ca_topic_score_gemma":0.0011798245,"teacher_disagreement_score":0.14779235,"about_ca_system_score_codex":0.0023952012,"about_ca_system_score_gemma":0.0021551668,"threshold_uncertainty_score":0.78161},"labels":[],"label_agreement":null},{"id":"W2189164006","doi":"10.6339/jds.2004.02(2).144","title":"Wavelet Analysis of Tide-affected Low Streamflows Series","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Image and Signal Denoising Methods","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Series (stratigraphy); Streamflow; Wavelet; Environmental science; Quantile; Flow (mathematics); Noise (video); Frequency domain; Wavelet transform; Geology; Mathematics; Statistics; Computer science; Geography; Mathematical analysis; Drainage basin","score_opus":0.04068391126091666,"score_gpt":0.33389003210052914,"score_spread":0.2932061208396125,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2189164006","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":[0.8501012,0.00012310786,0.14802153,0.000056380617,0.000026312331,0.000026660273,0.0002538838,0.00012491946,0.0012659288],"genre_scores_gemma":[0.97781926,0.00016805914,0.021066265,0.0000057075754,0.000010853585,0.000012681022,0.00029619664,0.000016790782,0.0006041941],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999864,0.00002830657,0.000015576627,0.000027069578,0.000044639637,0.000020297855],"domain_scores_gemma":[0.99958676,0.00018201472,0.00007712863,0.00004275348,0.000097522505,0.0000138920195],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006865721,0.000194768,0.00016887386,0.0007846166,0.00012798767,0.00032099464,0.00013720554,0.00012915483,0.00043002854],"category_scores_gemma":[0.0014175781,0.00005934562,0.00024401404,0.0009674444,0.00019405471,0.0002584343,0.0001562031,0.000254786,0.000115634284],"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":[0.0007633065,0.00029178016,0.056702264,0.00040256104,0.00016442369,0.0012375122,0.0012840164,0.11249885,0.29380503,0.010388271,0.0013837132,0.5210782],"study_design_scores_gemma":[0.000013569636,0.00021509387,0.29190114,0.000038144884,0.000090201276,0.00040158123,0.00055348146,0.6365548,0.06281772,0.0033954617,0.0039714407,0.00004734823],"about_ca_topic_score_codex":0.0010154147,"about_ca_topic_score_gemma":0.00095188065,"teacher_disagreement_score":0.0010154147,"about_ca_system_score_codex":0.00014997362,"about_ca_system_score_gemma":0.00014459879,"threshold_uncertainty_score":0.0036309958},"labels":[],"label_agreement":null},{"id":"W2190667587","doi":"10.6339/jds.2003.01(2).131","title":"Analysis of Unbalanced Microarray Data","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Microarray analysis techniques; China; Data science; Microarray databases; Library science; Data analysis; Data mining; Computer science; Statistics; Political science; Mathematics; Biology; Gene","score_opus":0.06929950204773849,"score_gpt":0.36920503824217027,"score_spread":0.2999055361944318,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2190667587","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":[0.049902223,0.0020302983,0.9076702,0.0017094152,0.0009671418,0.0004377561,0.031779665,0.002787235,0.00271608],"genre_scores_gemma":[0.31619263,0.0021301385,0.5758262,0.0008427283,0.0012322478,0.0028544553,0.09364902,0.0014699952,0.005802597],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98923665,0.0037872358,0.00085438654,0.0021377066,0.0035320045,0.00045198447],"domain_scores_gemma":[0.9616882,0.024430793,0.0021387022,0.0072620413,0.0040588304,0.00042150673],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011986963,0.000930177,0.0019162813,0.003560282,0.0007696779,0.0014743283,0.0010533516,0.000520326,0.009725481],"category_scores_gemma":[0.0558451,0.00051257317,0.0014176511,0.005044945,0.0007450304,0.0015410071,0.0011116687,0.0017093696,0.002122036],"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":[0.0020384581,0.0002330799,0.054518685,0.0017782784,0.0017596069,0.0007590309,0.00037972277,0.038887456,0.05127212,0.030189276,0.111733645,0.70645076],"study_design_scores_gemma":[0.00031493453,0.00076706364,0.12611261,0.0002929922,0.0008684893,0.0017731228,0.0005567023,0.4924067,0.027730832,0.20183347,0.14710511,0.00023797821],"about_ca_topic_score_codex":0.0009644263,"about_ca_topic_score_gemma":0.0015391279,"teacher_disagreement_score":0.011986963,"about_ca_system_score_codex":0.0008531461,"about_ca_system_score_gemma":0.0013641872,"threshold_uncertainty_score":0.06339383},"labels":[],"label_agreement":null},{"id":"W2514070870","doi":"10.6339/jds.201604_14(2).0002","title":"Comparing Pre-Post Change Across Groups: Guidelines for Choosing between Difference Scores, ANCOVA, and Residual Change Scores","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":60,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Analysis of covariance; Residual; Statistics; Significant difference; Psychology; Mathematics","score_opus":0.732100748033184,"score_gpt":0.5505718425854673,"score_spread":0.18152890544771672,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2514070870","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0033405514,0.0064681587,0.9115558,0.0065684076,0.007176665,0.031600576,0.019626739,0.005497623,0.008165485],"genre_scores_gemma":[0.005084194,0.0013100739,0.93308514,0.0018164126,0.00041172543,0.051001295,0.0027314248,0.0017586808,0.0028009787],"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.73558253,0.19329822,0.04210356,0.005777583,0.02221792,0.0010202153],"domain_scores_gemma":[0.45448256,0.44665036,0.015782848,0.029067703,0.05170177,0.0023146884],"candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.21547356,0.004881492,0.007278097,0.012681763,0.002076542,0.0040135724,0.0073169954,0.0070860847,0.0268219],"category_scores_gemma":[0.5365975,0.0041824207,0.006591364,0.012926361,0.0048708827,0.0048685856,0.0032634854,0.013790033,0.0154905515],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0030881916,0.0014715728,0.005912216,0.011193919,0.0017523716,0.00087789877,0.0025254793,0.0031437327,0.0036406026,0.03207331,0.590018,0.3443028],"study_design_scores_gemma":[0.0057225716,0.0047497037,0.04369739,0.025648413,0.0021818248,0.0020371045,0.0039534764,0.029607149,0.016826553,0.22356327,0.64040315,0.0016094665],"about_ca_topic_score_codex":0.0040097926,"about_ca_topic_score_gemma":0.010289136,"teacher_disagreement_score":0.78452647,"about_ca_system_score_codex":0.0016585697,"about_ca_system_score_gemma":0.0057255025,"threshold_uncertainty_score":0.96746075},"labels":[],"label_agreement":null},{"id":"W2564423956","doi":"10.6339/jds.201407_12(3).0001","title":"Factorial ANOVA with unbalanced data: A fresh look at the types of sums of squares","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":51,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"","keywords":"Confusion; Factorial; Factorial analysis; Analysis of variance; Statistics; Type (biology); Econometrics; Mathematics; Factorial experiment; Explained sum of squares; Lack-of-fit sum of squares; Generalized least squares; Computer science; Arithmetic; Psychology; Non-linear least squares","score_opus":0.7074570963113356,"score_gpt":0.5968141994031163,"score_spread":0.11064289690821927,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2564423956","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":[0.0015475977,0.012025092,0.9663254,0.008249409,0.0042822985,0.00032950885,0.00033690064,0.00037542966,0.0065283403],"genre_scores_gemma":[0.04839767,0.01899037,0.90636194,0.0093308,0.008458851,0.0032424186,0.0005333236,0.0012424375,0.0034421715],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.7392431,0.17540938,0.018102566,0.012195022,0.05342496,0.0016250493],"domain_scores_gemma":[0.4868029,0.44301558,0.01594214,0.03261634,0.020299427,0.001323538],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1670331,0.0028743343,0.005202207,0.0072529274,0.002541747,0.009574843,0.006775101,0.0037607264,0.010711812],"category_scores_gemma":[0.37639207,0.001545253,0.005393759,0.011710905,0.022459863,0.015931835,0.0074443095,0.015161975,0.0029750313],"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":[0.0003950957,0.00009947736,0.0032020782,0.0035893274,0.00060259725,0.00051527884,0.002712924,0.0023614382,0.000857095,0.6458071,0.027150996,0.31270665],"study_design_scores_gemma":[0.000054155935,0.00032655706,0.0013882885,0.0016492096,0.00012938316,0.00072649843,0.0006243025,0.0056691077,0.0010012621,0.90910804,0.07916654,0.0001566888],"about_ca_topic_score_codex":0.001434687,"about_ca_topic_score_gemma":0.0016704942,"teacher_disagreement_score":0.1670331,"about_ca_system_score_codex":0.004210243,"about_ca_system_score_gemma":0.005045721,"threshold_uncertainty_score":0.88336605},"labels":[],"label_agreement":null},{"id":"W3047907238","doi":"10.6339/jds.202007_18(3).0015","title":"COVID-19 Fatality: A Cross-Sectional Study using Adaptive Lasso Penalized Sliced Inverse Regression","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Case fatality rate; Lasso (programming language); Coronavirus disease 2019 (COVID-19); Outbreak; Feature selection; Regression; Pandemic; Regression analysis; Population; Statistics; Medicine; Computer science; Disease; Mathematics; Infectious disease (medical specialty); Virology; Artificial intelligence; Environmental health; Internal medicine","score_opus":0.723200934496076,"score_gpt":0.58094962354606,"score_spread":0.14225131095001609,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3047907238","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":[0.97629255,0.00035971164,0.02179391,0.00021451339,0.000051494313,0.00009618027,0.00091199146,0.000033887904,0.00024580577],"genre_scores_gemma":[0.97796255,0.00016942984,0.01815025,0.00009254562,0.00006549681,0.00016476295,0.0030156332,0.000018973424,0.00036034037],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.997003,0.0020268743,0.00016573294,0.00051370345,0.00017432251,0.000116342206],"domain_scores_gemma":[0.9859226,0.008256771,0.0021668656,0.002063248,0.0011103279,0.00048012479],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010371685,0.0006290504,0.00064898777,0.00087888906,0.0004781065,0.00058565824,0.0007977328,0.00083662703,0.0011245995],"category_scores_gemma":[0.014438405,0.00038227206,0.0013364855,0.000671226,0.0004728389,0.00044332666,0.0008694871,0.0015433157,0.00017424818],"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":[0.0012353265,0.0010993918,0.9562931,0.00015019912,0.0015033276,0.00075295236,0.00047903674,0.013145844,0.0019080823,0.0011387444,0.0022167128,0.0200773],"study_design_scores_gemma":[0.00019736715,0.0027333049,0.6971407,0.000113998474,0.0011105477,0.0016846437,0.0012676957,0.28833073,0.0026635446,0.0014647224,0.003159209,0.00013363115],"about_ca_topic_score_codex":0.0042126747,"about_ca_topic_score_gemma":0.0036300113,"teacher_disagreement_score":0.010371685,"about_ca_system_score_codex":0.00027189346,"about_ca_system_score_gemma":0.00056703214,"threshold_uncertainty_score":0.054851353},"labels":[],"label_agreement":null},{"id":"W3053629049","doi":"10.6339/jds.202007_18(3).0018","title":"Data Visualization and Descriptive Analysis for Understanding Epidemiological Characteristics of COVID-19: A Case Study of a Dataset from January 22, 2020 to March 29, 2020","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Case fatality rate; Epidemiology; Coronavirus disease 2019 (COVID-19); Coronavirus; Incubation period; Medicine; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); 2019-20 coronavirus outbreak; Nonparametric statistics; Disease; Internal medicine; Pathology; Statistics; Psychology; Incubation; Infectious disease (medical specialty); Outbreak","score_opus":0.6960204092600549,"score_gpt":0.5511004728700827,"score_spread":0.14491993638997225,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3053629049","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":[0.85972655,0.0018432986,0.017276611,0.0038894578,0.00015966351,0.0001944561,0.113106355,0.0010818236,0.0027218917],"genre_scores_gemma":[0.87013215,0.0004990499,0.039544336,0.00014005842,0.00008685103,0.00019733478,0.08856386,0.00007156636,0.00076482614],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9988287,0.0004012554,0.00020887912,0.0002004918,0.00025288167,0.00010782046],"domain_scores_gemma":[0.9892916,0.006818557,0.0013610875,0.00092593534,0.001213797,0.00038907086],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029160662,0.000466183,0.00039110126,0.0051650302,0.00049674994,0.0011286975,0.00058150274,0.0007286182,0.0009666566],"category_scores_gemma":[0.010235019,0.00013401176,0.0007893953,0.003753086,0.00031754572,0.0008277163,0.0008241082,0.0007047968,0.0002944873],"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":[0.00094136817,0.0005844994,0.8462569,0.000643437,0.00037333008,0.0027315472,0.0016823364,0.018364849,0.0038464034,0.0042382223,0.03728856,0.08304857],"study_design_scores_gemma":[0.00009494071,0.00035048573,0.7524094,0.0003116393,0.00017097463,0.0029761584,0.006671122,0.16430132,0.005102398,0.0074312557,0.06005315,0.000127171],"about_ca_topic_score_codex":0.011789187,"about_ca_topic_score_gemma":0.018507672,"teacher_disagreement_score":0.011789187,"about_ca_system_score_codex":0.0008344476,"about_ca_system_score_gemma":0.00093196135,"threshold_uncertainty_score":0.023441076},"labels":[],"label_agreement":null},{"id":"W3127126615","doi":"10.6339/21-jds997","title":"A Vine Copula Model for Climate Trend Analysis using Canadian Temperature Data","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Horticultural and Viticultural Research","field":"Agricultural and Biological Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Western University; Actua; University of Waterloo","funders":"","keywords":"Copula (linguistics); Vine copula; Multivariate statistics; Econometrics; Computer science; Climate change; Climate model; Data mining; Mathematics; Machine learning; Ecology","score_opus":0.24212649655141386,"score_gpt":0.3988594280744594,"score_spread":0.15673293152304554,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3127126615","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":[0.10655345,0.0010462063,0.87394536,0.0013359151,0.00017126199,0.00033266557,0.0075490284,0.0014950328,0.007571051],"genre_scores_gemma":[0.72501385,0.0013781304,0.24228273,0.00035041606,0.000104141734,0.0007204111,0.011111318,0.00073143456,0.018307626],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99907136,0.0003076109,0.000035207624,0.00025070336,0.00014482021,0.00019023987],"domain_scores_gemma":[0.9975721,0.0012339756,0.00019472126,0.00020748083,0.00066106865,0.00013072156],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003255009,0.0010338032,0.0009954922,0.0014096831,0.001102683,0.0017282318,0.0033868174,0.0010824858,0.0054663448],"category_scores_gemma":[0.009968679,0.00067035976,0.0013817996,0.002706829,0.00084609794,0.0013162005,0.0011262863,0.0022515922,0.0011708351],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00008702703,0.000044439734,0.007149036,0.000060756283,0.00012404463,0.0001297836,0.00014783636,0.924081,0.00043652632,0.033645302,0.007253861,0.026840435],"study_design_scores_gemma":[0.0000067070405,0.000007007367,0.00090578955,0.0000058589185,0.000009412121,0.000010869402,0.000017534989,0.99440706,0.000058106107,0.0031960357,0.0013656056,0.000009993042],"about_ca_topic_score_codex":0.5440559,"about_ca_topic_score_gemma":0.44644952,"teacher_disagreement_score":0.45594412,"about_ca_system_score_codex":0.004404006,"about_ca_system_score_gemma":0.005801886,"threshold_uncertainty_score":0.9172585},"labels":[],"label_agreement":null},{"id":"W3135587851","doi":"10.6339/jds.201807_16(3).0002","title":"Minimum Profile Hellinger Distance Estimation for A Two-Sample Location-Shifted Model","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Statistical Methods and Inference","field":"Mathematics","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 Calgary","funders":"","keywords":"Hellinger distance; Nonparametric statistics; Parametric statistics; Robustness (evolution); Mathematics; Asymptotic distribution; Parametric model; Statistics; Semiparametric regression; Sample size determination; Consistency (knowledge bases); Estimation theory; Applied mathematics; Semiparametric model; Estimator","score_opus":0.24353646812470423,"score_gpt":0.46553765814242265,"score_spread":0.22200119001771842,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3135587851","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":[0.009965167,0.00018018011,0.989335,0.00013526963,0.000012408843,0.000034512723,0.00005521106,0.000044302364,0.00023802223],"genre_scores_gemma":[0.45734084,0.00085931335,0.5373046,0.00024265696,0.00008839252,0.0004328426,0.00078703644,0.00007974533,0.0028646293],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99561834,0.0025497414,0.00024405199,0.0008744052,0.00054574886,0.00016782137],"domain_scores_gemma":[0.9808865,0.015446542,0.001083621,0.0013851697,0.00095408777,0.00024417718],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0078026582,0.0009567336,0.0017928735,0.0012972578,0.00055552204,0.001449123,0.002199463,0.0016804697,0.0013806544],"category_scores_gemma":[0.029683486,0.0006303834,0.0012783435,0.0011631083,0.0015478875,0.003280503,0.0028313838,0.0024685243,0.00036687398],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00034212464,0.0001195515,0.006581127,0.0004476995,0.00032599593,0.00040280106,0.00043328982,0.73538035,0.0041525536,0.14119126,0.0016441335,0.10897913],"study_design_scores_gemma":[0.000016367145,0.00006485269,0.0007858182,0.000027564049,0.000026160935,0.000085856074,0.0000359753,0.9498555,0.0011034324,0.047261085,0.000704463,0.000032873268],"about_ca_topic_score_codex":0.0019856337,"about_ca_topic_score_gemma":0.0013213414,"teacher_disagreement_score":0.0078026582,"about_ca_system_score_codex":0.0013711903,"about_ca_system_score_gemma":0.001835923,"threshold_uncertainty_score":0.04126495},"labels":[],"label_agreement":null},{"id":"W3136368678","doi":"10.6339/jds.201407_12(3).0007","title":"Copulas Applications in Estimating Value-at-Risk (VaR): Iranian Crude Oil Prices","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Financial Risk and Volatility Modeling","field":"Economics, Econometrics and Finance","cited_by":0,"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 Northern British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Copula (linguistics); Economics; Econometrics; Crude oil; Normality; Risk management; Petroleum; Value at risk; Profit (economics); Multivariate statistics; Financial economics; Finance; Statistics; Microeconomics; Mathematics; Engineering","score_opus":0.07526248261515965,"score_gpt":0.3051591706552883,"score_spread":0.22989668804012864,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3136368678","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":[0.010484022,0.0018601618,0.9826528,0.00038477927,0.00011715128,0.00007623033,0.00037689443,0.00031653416,0.00373157],"genre_scores_gemma":[0.5673917,0.012092718,0.41064337,0.00039227438,0.00056644494,0.0006029145,0.0015127058,0.000441643,0.006356133],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99830127,0.00091474596,0.00011238447,0.00026205534,0.00030250594,0.00010699895],"domain_scores_gemma":[0.9949266,0.0035940153,0.00060252147,0.00023518407,0.0005755291,0.000066155444],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029534078,0.001570852,0.001096635,0.0018582491,0.00048230137,0.0018962574,0.0011751928,0.0012918118,0.0031591242],"category_scores_gemma":[0.01584836,0.00071268814,0.0016459493,0.0027327647,0.0006610819,0.0016065178,0.0015431591,0.0025812774,0.00097278855],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00006859786,0.00008666712,0.009475432,0.00042347144,0.0003942282,0.0010148762,0.0003234808,0.77307814,0.0017349053,0.08453643,0.0067052064,0.12215855],"study_design_scores_gemma":[0.00000989221,0.00004610686,0.0028763714,0.00010194456,0.00008733596,0.0003099608,0.000103593906,0.9509968,0.00084211497,0.039294936,0.005281519,0.00004939887],"about_ca_topic_score_codex":0.009812504,"about_ca_topic_score_gemma":0.0042606657,"teacher_disagreement_score":0.009812504,"about_ca_system_score_codex":0.0007676138,"about_ca_system_score_gemma":0.0012052085,"threshold_uncertainty_score":0.019510806},"labels":[],"label_agreement":null},{"id":"W3138069980","doi":"10.6339/jds.201207_10(3).0006","title":"A Data Mining Approach to Comparing American and Canadian Grade 10 Students’ PISA Science Test Performance","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Online Learning and Analytics","field":"Computer Science","cited_by":20,"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":"Test (biology); Academic achievement; Mathematics education; Achievement test; Psychology; Political science; Standardized test","score_opus":0.10987572929409287,"score_gpt":0.36575051924651825,"score_spread":0.2558747899524254,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3138069980","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":[0.9323123,0.00034978878,0.015087542,0.0006982412,0.00009246008,0.001799525,0.035053972,0.00053840195,0.014067664],"genre_scores_gemma":[0.9503489,0.00023959532,0.031107202,0.0000993879,0.00001672191,0.001207415,0.01394662,0.000036115263,0.002998038],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9966583,0.00045031446,0.00035777988,0.00046505226,0.0016754859,0.00039306164],"domain_scores_gemma":[0.9895488,0.003159348,0.0012250185,0.00045693753,0.00505081,0.0005590643],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003953649,0.00092239195,0.0009794482,0.014358514,0.0027374392,0.0023642497,0.0021022263,0.0005432669,0.0019377179],"category_scores_gemma":[0.014598389,0.00030544383,0.0011501822,0.017979423,0.0007365387,0.000637143,0.0012253205,0.001298329,0.0004933893],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00058333646,0.0010053581,0.8490982,0.00041604662,0.0006989811,0.0004558622,0.0071079973,0.0027311724,0.0031887337,0.00209324,0.008569917,0.12405118],"study_design_scores_gemma":[0.000035612517,0.0004066226,0.95535386,0.00013599358,0.00016734708,0.0002041498,0.014993375,0.01517406,0.0027322064,0.0007564807,0.009959345,0.00008099418],"about_ca_topic_score_codex":0.81478965,"about_ca_topic_score_gemma":0.8174941,"teacher_disagreement_score":0.18521035,"about_ca_system_score_codex":0.006591027,"about_ca_system_score_gemma":0.013962458,"threshold_uncertainty_score":0.37260216},"labels":[],"label_agreement":null},{"id":"W3157762174","doi":"10.6339/21-jds993","title":"Testing for COVID-19: Some Statistical Issues","year":2021,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"SARS-CoV-2 detection and testing","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Coronavirus disease 2019 (COVID-19); Coronavirus; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); 2019-20 coronavirus outbreak; Test (biology); Computer science; Statistical hypothesis testing; Statistical analysis; Statistics; Virology; Medicine; Mathematics; Outbreak; Biology","score_opus":0.27624816157876786,"score_gpt":0.47801977626905207,"score_spread":0.2017716146902842,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3157762174","genre_codex":"methods","genre_gemma":"methods","domain_codex":"methods","domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":"methods","prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.025659144,0.010325657,0.81369704,0.1287225,0.0065789195,0.0022818572,0.0009812039,0.0011838524,0.010569903],"genre_scores_gemma":[0.5285983,0.0022940105,0.39704287,0.053945642,0.010203937,0.004230318,0.0006170902,0.00058660755,0.0024812531],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.2888082,0.61973953,0.02453483,0.017448634,0.0470318,0.0024371431],"domain_scores_gemma":[0.056438968,0.9021471,0.01032326,0.020506019,0.009606428,0.0009781964],"candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.59026015,0.0021912095,0.004809714,0.005550388,0.0038978707,0.007877523,0.008329727,0.007953352,0.0033241715],"category_scores_gemma":[0.812265,0.0014626655,0.0058469023,0.007038286,0.021951038,0.009439767,0.0060522733,0.015652955,0.00063640863],"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":[0.0036109409,0.00044782195,0.12645932,0.003880308,0.0079282615,0.0049027437,0.0050211325,0.024588598,0.001535313,0.42386222,0.06286481,0.33489844],"study_design_scores_gemma":[0.0009927754,0.0020226617,0.02675818,0.0031502293,0.0020747157,0.0037475976,0.0019370925,0.21904601,0.0038390676,0.6831936,0.05277184,0.00046623952],"about_ca_topic_score_codex":0.006041039,"about_ca_topic_score_gemma":0.0037770816,"teacher_disagreement_score":0.40973985,"about_ca_system_score_codex":0.0062030703,"about_ca_system_score_gemma":0.005458553,"threshold_uncertainty_score":0.5052822},"labels":[],"label_agreement":null},{"id":"W3215602957","doi":"10.6339/22-jds1047","title":"A Review on Graph Neural Network Methods in Financial Applications","year":2022,"lang":"en","type":"review","venue":"Journal of Data Science","topic":"Advanced Graph Neural Networks","field":"Computer Science","cited_by":122,"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":"Institute for Catastrophic Loss Reduction","keywords":"Computer science; Graph; Categorization; Artificial neural network; Financial modeling; Data mining; Data science; Finance; Artificial intelligence; Theoretical computer science; Business","score_opus":0.19864865051033517,"score_gpt":0.48307997417449006,"score_spread":0.2844313236641549,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3215602957","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00090124743,0.95589626,0.028819215,0.0018086153,0.0012674071,0.00004691788,0.00020175408,0.00014029861,0.010918305],"genre_scores_gemma":[0.007899849,0.97111356,0.015506378,0.00056229916,0.0010675602,0.000053859596,0.00029766012,0.00004436542,0.003454514],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9996111,0.0000788808,0.00005155753,0.00008673637,0.00014779772,0.000023951521],"domain_scores_gemma":[0.9986551,0.0008524285,0.00006955398,0.00004755303,0.00033985707,0.000035452777],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010544579,0.0013507195,0.0010543274,0.002923572,0.000405476,0.0012728276,0.0015183476,0.0013708541,0.0065217544],"category_scores_gemma":[0.0031432137,0.00053139013,0.0008002613,0.005228167,0.0005888001,0.0026385465,0.00070129836,0.0014894642,0.003061124],"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":[0.000031932086,0.00008004721,0.0006500353,0.008689305,0.00011619897,0.00013665638,0.000075052456,0.0076260697,0.0006482947,0.034048274,0.04630078,0.9015974],"study_design_scores_gemma":[0.000016172935,0.00011025855,0.0016746031,0.0052241846,0.0002164082,0.00076782075,0.000105184765,0.019927762,0.0012198361,0.04925986,0.92138726,0.000090721645],"about_ca_topic_score_codex":0.003963791,"about_ca_topic_score_gemma":0.0043921005,"teacher_disagreement_score":0.0065217544,"about_ca_system_score_codex":0.00091074227,"about_ca_system_score_gemma":0.0014490929,"threshold_uncertainty_score":0.021817446},"labels":[],"label_agreement":null},{"id":"W4210580103","doi":"10.6339/22-jds1034","title":"Computational Challenges of t and Related Copulas","year":2022,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Financial Risk and Volatility Modeling","field":"Economics, Econometrics and Finance","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 Waterloo","funders":"","keywords":"Copula (linguistics); Skew; Quantile; Computer science; Marginal distribution; Applied mathematics; Mathematics; Econometrics; Random variable; Statistics","score_opus":0.12877524547984,"score_gpt":0.29539751068358244,"score_spread":0.16662226520374243,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4210580103","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":[0.014848553,0.0012344482,0.973539,0.00442528,0.00023946734,0.00008094893,0.00039832265,0.0007113503,0.004522552],"genre_scores_gemma":[0.21271507,0.0017593104,0.77795315,0.0012351254,0.00068364316,0.00042691673,0.00068684074,0.00087278005,0.0036672277],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9931854,0.004741619,0.0003592409,0.0007564407,0.0007004067,0.00025683158],"domain_scores_gemma":[0.9290556,0.06071236,0.0019487902,0.0052525434,0.0020142738,0.0010164703],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.015702521,0.0011269512,0.0026677295,0.0014602504,0.0015403294,0.0051010447,0.0033972128,0.0022924184,0.011457158],"category_scores_gemma":[0.07227075,0.0012303888,0.0027667787,0.0024057855,0.0023118365,0.0068554166,0.0049328334,0.0052495603,0.0022683137],"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":[0.00019333624,0.00012118476,0.0034383049,0.00044789776,0.00025877936,0.00055521127,0.0005224703,0.40667617,0.00044402265,0.483908,0.016335469,0.08709913],"study_design_scores_gemma":[0.000045755754,0.000021547958,0.00025662943,0.000052528616,0.000017067929,0.00012845644,0.00013817503,0.61328727,0.00017022259,0.38190323,0.0039534196,0.000025657206],"about_ca_topic_score_codex":0.009800604,"about_ca_topic_score_gemma":0.0086345645,"teacher_disagreement_score":0.015702521,"about_ca_system_score_codex":0.0014796794,"about_ca_system_score_gemma":0.0026826526,"threshold_uncertainty_score":0.08304381},"labels":[],"label_agreement":null},{"id":"W4392752805","doi":"10.6339/24-jds1121","title":"Testing Perceptual Accuracy in a U.S. General Population Survey Using Stacked Bar Charts","year":2024,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Data Visualization and Analytics","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":"Bar chart; Visualization; Perception; Computer science; Bar (unit); Data visualization; Information visualization; Pie chart; Population; Key (lock); Data science; Survey data collection; Human–computer interaction; Data mining; Psychology; Statistics; Mathematics; Geography","score_opus":0.24662321875917856,"score_gpt":0.4300851825046271,"score_spread":0.18346196374544851,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4392752805","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9927644,0.00006098879,0.003422395,0.00028395702,0.00004241732,0.00009378878,0.0009103972,0.000080026024,0.0023417408],"genre_scores_gemma":[0.9965557,0.000047944523,0.0021032866,0.00013180188,0.000016387881,0.00013440003,0.0007566457,0.000012673455,0.00024120382],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.98595285,0.008045066,0.0013176001,0.0016729934,0.002356689,0.00065479946],"domain_scores_gemma":[0.9104167,0.05635109,0.011910113,0.008478437,0.010891792,0.0019518997],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.017495526,0.0004185706,0.00043502531,0.0018927391,0.00067027356,0.0017166359,0.000976623,0.0012017837,0.0032071918],"category_scores_gemma":[0.102006,0.0003640951,0.0011098566,0.002226591,0.0011605672,0.001995681,0.0019413058,0.0011078117,0.00069123047],"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":[0.00033100197,0.00035560702,0.97400576,0.00008113788,0.0001786514,0.00007226918,0.004626767,0.0012391668,0.0004011232,0.000513787,0.0023234666,0.01587127],"study_design_scores_gemma":[0.000039269886,0.0009935485,0.9801958,0.00007132401,0.00008024866,0.0001182716,0.0066882116,0.008097414,0.00063766795,0.0008950904,0.0021363327,0.000046866247],"about_ca_topic_score_codex":0.01337785,"about_ca_topic_score_gemma":0.012106474,"teacher_disagreement_score":0.9825045,"about_ca_system_score_codex":0.00054791884,"about_ca_system_score_gemma":0.0006374692,"threshold_uncertainty_score":0.09252632},"labels":[],"label_agreement":null},{"id":"W4394949264","doi":"10.6339/24-jds1122","title":"BIE: Binary Image Encoding for the Classification of Tabular Data","year":2024,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Adversarial Robustness in Machine Learning","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":"Simon Fraser University","funders":"","keywords":"Computer science; Artificial intelligence; Encoding (memory); Pattern recognition (psychology); Convolutional neural network; Contextual image classification; Benchmark (surveying); Image (mathematics); Deep learning; Feature (linguistics); Field (mathematics); Binary image; Image processing; Mathematics","score_opus":0.1344438530442302,"score_gpt":0.40138674335794516,"score_spread":0.266942890313715,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394949264","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":[0.06598583,0.0028332972,0.89124894,0.00096855976,0.00069349154,0.00044613506,0.0070872903,0.021611176,0.009125199],"genre_scores_gemma":[0.2745471,0.0016267254,0.6986895,0.0007443694,0.00016683596,0.0005527565,0.014584676,0.0010098267,0.008078139],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99958354,0.000050615126,0.00004147239,0.00008449152,0.00017457933,0.000065377586],"domain_scores_gemma":[0.99907804,0.00022933637,0.00015516816,0.00024409562,0.00022968174,0.000063764266],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000584412,0.0010475981,0.0006222587,0.001958009,0.00025206315,0.0011724008,0.0013152392,0.0007127872,0.0062041637],"category_scores_gemma":[0.0034033032,0.00027577035,0.0006006884,0.0018492567,0.0004698569,0.0028010895,0.0011641514,0.001462332,0.00300788],"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":[0.0007517315,0.00024955347,0.0022710026,0.0004965406,0.00005789151,0.00016269991,0.00011498322,0.020896802,0.04245036,0.01508897,0.038039066,0.87942046],"study_design_scores_gemma":[0.00010008634,0.0004260703,0.0035842005,0.00018427365,0.0000509015,0.00064973295,0.00016305233,0.78113204,0.1413098,0.02530491,0.046990722,0.00010418882],"about_ca_topic_score_codex":0.0020017305,"about_ca_topic_score_gemma":0.0023017696,"teacher_disagreement_score":0.0062041637,"about_ca_system_score_codex":0.00059756055,"about_ca_system_score_gemma":0.0006903269,"threshold_uncertainty_score":0.020754993},"labels":[],"label_agreement":null},{"id":"W4398770290","doi":"10.6339/24-jds1131","title":"Evaluating Perceptual Judgements on 3D Printed Bar Charts","year":2024,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"3D Shape Modeling and Analysis","field":"Engineering","cited_by":1,"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":"Bar (unit); Bar chart; Perception; Computer science; Engineering drawing; Psychology; Statistics; Mathematics; Engineering; Geology","score_opus":0.15466865673993677,"score_gpt":0.3983999130363163,"score_spread":0.2437312562963795,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4398770290","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":[0.9532722,0.0004752616,0.0322924,0.00014934246,0.00020667579,0.00024725476,0.00030060628,0.00043166432,0.012624517],"genre_scores_gemma":[0.9756638,0.00019105079,0.022221211,0.00009139657,0.0000626938,0.00011791555,0.00031273026,0.00014259171,0.0011967295],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9851127,0.005893348,0.001152412,0.0014943223,0.005870691,0.00047652583],"domain_scores_gemma":[0.85933805,0.10628612,0.01161016,0.009252711,0.01119118,0.0023217828],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010987233,0.00091885217,0.00074321154,0.003278885,0.00061437266,0.0046976362,0.0016552804,0.0020214452,0.0065369667],"category_scores_gemma":[0.18980613,0.00047125094,0.0012046966,0.0013069933,0.002205028,0.0040487004,0.002793044,0.0011733733,0.0012730605],"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":[0.023314385,0.0011548789,0.15698281,0.00340738,0.001190974,0.0019379245,0.05143939,0.041296322,0.29552504,0.016740689,0.008392027,0.3986182],"study_design_scores_gemma":[0.0011927509,0.014487322,0.6721518,0.0016450546,0.000825984,0.0018183169,0.03909434,0.1432033,0.071462564,0.033714276,0.01875511,0.0016492166],"about_ca_topic_score_codex":0.0026260766,"about_ca_topic_score_gemma":0.0023470426,"teacher_disagreement_score":0.010987233,"about_ca_system_score_codex":0.0009879044,"about_ca_system_score_gemma":0.00043049493,"threshold_uncertainty_score":0.05810678},"labels":[],"label_agreement":null},{"id":"W4400245371","doi":"10.6339/24-jds1143","title":"Physician Effects in Critical Care: A Causal Inference Approach Through Propensity Weighting with Parametric and Super Learning Methods","year":2024,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Advanced Causal Inference Techniques","field":"Mathematics","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":"Western University","funders":"","keywords":"Propensity score matching; Weighting; Causal inference; Covariate; Confounding; Inverse probability weighting; Estimator; Parametric statistics; Intensive care unit; Computer science; Medicine; Statistics; Intensive care medicine; Machine learning; Mathematics","score_opus":0.32995507111031813,"score_gpt":0.5379351258148638,"score_spread":0.20798005470454567,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400245371","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":[0.006010718,0.0001466232,0.9930454,0.000251707,0.000017943583,0.00006864599,0.000035427573,0.00006876241,0.00035471353],"genre_scores_gemma":[0.44474405,0.00078438275,0.5505573,0.00043146586,0.00026088418,0.000651732,0.0002907125,0.00008631095,0.002193138],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98368686,0.013132438,0.00041365513,0.001195563,0.0012455835,0.00032583848],"domain_scores_gemma":[0.89263093,0.09252554,0.005293754,0.0064433613,0.0024838967,0.00062248105],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03688011,0.0008403507,0.0016356418,0.0024760386,0.0007713399,0.001666361,0.002895145,0.0016378281,0.0040227477],"category_scores_gemma":[0.09378984,0.0008275164,0.0018218941,0.0028030733,0.0023951752,0.0034945118,0.0034258266,0.0034745634,0.00027933635],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00026231294,0.00027985653,0.014273856,0.000318161,0.000543161,0.00031682602,0.00063476316,0.48114848,0.00061463274,0.327477,0.0023203464,0.17181058],"study_design_scores_gemma":[0.000038524733,0.00005850409,0.0009145621,0.000038426508,0.000050637198,0.000058794485,0.000041637333,0.8317536,0.00023921044,0.165842,0.00094640057,0.00001771402],"about_ca_topic_score_codex":0.004684018,"about_ca_topic_score_gemma":0.0037570214,"teacher_disagreement_score":0.03688011,"about_ca_system_score_codex":0.0014485064,"about_ca_system_score_gemma":0.0026380597,"threshold_uncertainty_score":0.19504297},"labels":[],"label_agreement":null},{"id":"W4404422782","doi":"10.6339/24-jds1134","title":"Rethinking Attention Weights as Bidirectional Coefficients","year":2024,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Explainable Artificial Intelligence (XAI)","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":false,"ca_institutions":"","funders":"Instituto de Ciencias del Mar y Limnología, Universidad Nacional Autónoma de México; Ministry of Education, India; Institute for Catastrophic Loss Reduction","keywords":"Computer science; Statistics; Mathematics; Psychology","score_opus":0.08340618000620546,"score_gpt":0.3567068886842551,"score_spread":0.2733007086780496,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4404422782","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":[0.059953533,0.00034347692,0.9359758,0.00047690296,0.00008224317,0.00005291294,0.000075750926,0.0005508189,0.0024886653],"genre_scores_gemma":[0.8856376,0.0004110719,0.10942039,0.00018079723,0.00007087444,0.00009606098,0.00011213542,0.00017121194,0.003899907],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99960965,0.00011038119,0.000023718861,0.000108419066,0.00009748671,0.00005041849],"domain_scores_gemma":[0.9990214,0.00049906067,0.000096872696,0.00013653022,0.00019320063,0.000053017862],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009641593,0.0007948478,0.0004283487,0.000709576,0.00027560108,0.00086333754,0.001039729,0.0007890937,0.0018886854],"category_scores_gemma":[0.00532884,0.00033206472,0.00048199025,0.00061914284,0.00074579974,0.0024518159,0.001643199,0.001790899,0.000292235],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00020441419,0.00014646642,0.0048005306,0.00021170966,0.00015062814,0.00021870436,0.000695546,0.32388568,0.051491313,0.13809979,0.00345822,0.476637],"study_design_scores_gemma":[0.000012089016,0.000036988687,0.000601252,0.00001909701,0.000028667651,0.00004009765,0.000029109051,0.955318,0.004535737,0.038038775,0.0013273284,0.0000128976635],"about_ca_topic_score_codex":0.0040152827,"about_ca_topic_score_gemma":0.0051678484,"teacher_disagreement_score":0.0040152827,"about_ca_system_score_codex":0.0005510871,"about_ca_system_score_gemma":0.0008050086,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W4415273830","doi":"10.6339/25-jds1200","title":"Q-learning with Compound Outcome and Mixed Misclassification and Measurement Error in Covariates","year":2025,"lang":"en","type":"article","venue":"Journal of Data Science","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":0,"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":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Covariate; Outcome (game theory); Observational error; Univariate; Estimation; Scope (computer science); Psychological intervention; Binary number","score_opus":0.3429165163248257,"score_gpt":0.45237121675090675,"score_spread":0.10945470042608102,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4415273830","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":[0.012546185,0.0005703739,0.9852754,0.00079446327,0.00010197573,0.00009545344,0.00014308053,0.00012410013,0.00034907128],"genre_scores_gemma":[0.56257457,0.0011343857,0.42877984,0.0013435957,0.0005962815,0.0008199188,0.00074809685,0.0001090005,0.0038943114],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9717281,0.01813161,0.0013776973,0.00592288,0.002015912,0.0008238324],"domain_scores_gemma":[0.828021,0.14138332,0.010365675,0.01637367,0.0030213557,0.0008349541],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.061556295,0.0017669845,0.004324988,0.0016794284,0.0013673921,0.0031119967,0.004428136,0.003202707,0.0026924426],"category_scores_gemma":[0.15535258,0.0011485082,0.0026955532,0.0028482,0.0052662925,0.0048172926,0.0038787832,0.0053533437,0.00044434099],"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":[0.0010375024,0.00029829546,0.029436382,0.0007769941,0.0009878284,0.0009098903,0.00080805086,0.31015742,0.00087695866,0.4598245,0.0038425117,0.1910438],"study_design_scores_gemma":[0.00018982403,0.00022152912,0.002832867,0.00016090609,0.00019047067,0.00023175727,0.000054397216,0.5491047,0.0010381415,0.44361553,0.002290949,0.00006890336],"about_ca_topic_score_codex":0.005397001,"about_ca_topic_score_gemma":0.0027750838,"teacher_disagreement_score":0.061556295,"about_ca_system_score_codex":0.002335507,"about_ca_system_score_gemma":0.003786031,"threshold_uncertainty_score":0.32554471},"labels":[],"label_agreement":null}]}