{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":814,"total_is_capped":false,"direct_labels_cover":3,"predictions_cover":814,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"189c207923fc","filters":{"topic":"Data Analysis with R"}},"results":[{"id":"W1951724000","doi":"10.18637/jss.v067.i01","title":"Fitting Linear Mixed-Effects Models Using <b>lme4</b>","year":2015,"lang":"en","type":"article","venue":"Journal of Statistical Software","topic":"Data Analysis with R","field":"Computer Science","cited_by":85618,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University","funders":"Natural Sciences and Engineering Research Council of Canada; Banff International Research Station for Mathematical Innovation and Discovery","keywords":"Restricted maximum likelihood; Deviance (statistics); Mixed model; Smoothing; Applied mathematics; Likelihood function; Mathematics; Generalized linear model; Linear model; Maximum likelihood; Generalized linear mixed model; Covariate; Algorithm; Statistics; Mathematical optimization; Computer science","authors":[{"name":"Douglas M. Bates","is_ca":false},{"name":"Martin Mächler","is_ca":false},{"name":"Benjamin M. Bolker","is_ca":true},{"name":"Steve Walker","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05709843782498766,"gpt":0.3091346325908944,"spread":0.2520361947659067,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03759384,0.006159679,0.00493646,0.004387685,0.001649259,0.004892902,0.008161397,0.003644082,0.1035673],"category_scores_gemma":[0.09536354,0.004190404,0.00969612,0.005529113,0.001860122,0.003672303,0.004358776,0.008650399,0.04838625],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00240295,"about_ca_system_score_gemma":0.00601193,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007569758,"about_ca_topic_score_gemma":0.0115414,"domain_scores_codex":[0.9750354,0.01628146,0.001833357,0.003602795,0.00258165,0.0006653257],"domain_scores_gemma":[0.9609304,0.02831186,0.002079688,0.005703904,0.002660715,0.0003134811],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001184637,0.0007973037,0.007121037,0.008443345,0.007340832,0.00120365,0.001519633,0.09763058,0.01259656,0.1333104,0.3226248,0.4062271],"study_design_scores_gemma":[0.0009175665,0.0006699003,0.006530485,0.001760354,0.002069909,0.001090137,0.0003456797,0.3505747,0.01840807,0.2227892,0.393927,0.0009169288],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0008799061,0.0002410765,0.9752074,0.0003288715,0.0001936452,0.0006676998,0.006731031,0.0140062,0.001744183],"genre_scores_gemma":[0.005517451,0.0002839386,0.9712832,0.0003229876,0.00005947013,0.005430903,0.005262344,0.009718803,0.002120859],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1035673,"threshold_uncertainty_score":0.3464671,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2050345155","doi":"10.1186/1471-2105-12-35","title":"VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R","year":2011,"lang":"en","type":"article","venue":"BMC Bioinformatics","topic":"Data Analysis with R","field":"Computer Science","cited_by":3262,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Ontario Institute for Cancer Research","funders":"Canadian Institutes of Health Research; Government of Ontario; Ontario Institute for Cancer Research","keywords":"Venn diagram; Computer science; Visualization; Process (computing); Data mining; Engineering drawing; Theoretical computer science; Programming language; Mathematics","authors":[{"name":"Hanbo Chen","is_ca":true},{"name":"Paul C. Boutros","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.07426646512308042,"gpt":0.2469729188522922,"spread":0.1727064537292118,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01122103,0.003464084,0.002904505,0.006970113,0.001434353,0.003416202,0.005243167,0.001344343,0.08476263],"category_scores_gemma":[0.04580577,0.00236365,0.003340425,0.004122194,0.00121245,0.003840954,0.004033357,0.005063618,0.02695123],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008609319,"about_ca_system_score_gemma":0.003094797,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001802806,"about_ca_topic_score_gemma":0.002376098,"domain_scores_codex":[0.9939023,0.002676071,0.0007728085,0.001257557,0.001142402,0.0002488069],"domain_scores_gemma":[0.9752855,0.01775907,0.001523814,0.002866133,0.002068957,0.0004965786],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001138176,0.0001534464,0.004947765,0.008324218,0.001575737,0.001054428,0.001670644,0.0196947,0.01746877,0.03621138,0.6473024,0.2604584],"study_design_scores_gemma":[0.0007537932,0.000233704,0.005904801,0.001366485,0.0006237754,0.001842532,0.0002794941,0.1076371,0.03900562,0.1186189,0.7228094,0.0009244471],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.001558522,0.000521216,0.7922303,0.0002676489,0.0004279497,0.0003712908,0.02351511,0.1787196,0.002388377],"genre_scores_gemma":[0.01425159,0.0004649534,0.8749583,0.0002959798,0.0001350203,0.002549579,0.02316324,0.08185899,0.002322382],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.08476263,"threshold_uncertainty_score":0.2835593,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2753599755","doi":"10.1111/rssa.12378","title":"Visualization in Bayesian Workflow","year":2019,"lang":"en","type":"article","venue":"Journal of the Royal Statistical Society Series A (Statistics in Society)","topic":"Data Analysis with R","field":"Computer Science","cited_by":1128,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Office of Naval Research; Institute of Education Sciences; Defense Advanced Research Projects Agency; Columbia University; Alfred P. Sloan Foundation; National Science Foundation","keywords":"Workflow; Visualization; Computer science; Bayesian probability; Data science; Data mining; Artificial intelligence; Database","authors":[{"name":"Jonah Gabry","is_ca":false},{"name":"Daniel Simpson","is_ca":true},{"name":"Aki Vehtari","is_ca":false},{"name":"Michael Betancourt","is_ca":false},{"name":"Andrew Gelman","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.006838517424116364,"gpt":0.2549536180972047,"spread":0.2481151006730883,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01465472,0.002342244,0.001687293,0.006115708,0.001992852,0.01201606,0.003043093,0.00236057,0.06002238],"category_scores_gemma":[0.05234171,0.001439562,0.002551426,0.005135992,0.001556077,0.007012154,0.007331074,0.00392274,0.02300003],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002060831,"about_ca_system_score_gemma":0.005130902,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01020697,"about_ca_topic_score_gemma":0.007137607,"domain_scores_codex":[0.9916077,0.003925362,0.0009454367,0.001144424,0.00196623,0.0004108594],"domain_scores_gemma":[0.9712624,0.01536634,0.00116231,0.005855986,0.004672931,0.001679978],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001488598,0.0002188651,0.002583315,0.001418867,0.0002643648,0.0009953964,0.003510327,0.02673417,0.00457187,0.2821173,0.3543637,0.3217332],"study_design_scores_gemma":[0.0002874541,0.00005563578,0.0008275891,0.0007457272,0.00006823903,0.0003946735,0.0003988755,0.1274019,0.005774612,0.4227192,0.4411002,0.0002257673],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001844198,0.0005382202,0.8621624,0.002870761,0.0004166167,0.0003349576,0.01000266,0.1083499,0.01348034],"genre_scores_gemma":[0.05494836,0.001495468,0.8795797,0.001371703,0.0003719446,0.001222674,0.02124356,0.0287029,0.01106366],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.06002238,"threshold_uncertainty_score":0.2007949,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2891445448","doi":"10.1177/1536867x19830877","title":"Fast and wild: Bootstrap inference in Stata using boottest","year":2019,"lang":"en","type":"article","venue":"The Stata Journal Promoting communications on statistics and Stata","topic":"Data Analysis with R","field":"Computer Science","cited_by":884,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Carleton University; Queen's University","funders":"","keywords":"Inference; Econometrics; Statistics; Computer science; Mathematics; Artificial intelligence","authors":[{"name":"David Roodman","is_ca":false},{"name":"Morten Ørregaard Nielsen","is_ca":true},{"name":"James G. MacKinnon","is_ca":true},{"name":"Matthew D. Webb","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06334283729434241,"gpt":0.3511636108219209,"spread":0.2878207735275785,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01516879,0.002046529,0.002825719,0.004827036,0.001119821,0.003814855,0.003679827,0.001455486,0.1395724],"category_scores_gemma":[0.1256152,0.002087332,0.002369455,0.006198019,0.001601157,0.004998185,0.005096938,0.005982176,0.06907238],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007450634,"about_ca_system_score_gemma":0.003482867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001848549,"about_ca_topic_score_gemma":0.002927322,"domain_scores_codex":[0.9851096,0.009784421,0.001431866,0.001347089,0.001733672,0.0005933431],"domain_scores_gemma":[0.9153987,0.06836348,0.00328065,0.008174915,0.003642913,0.001139448],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001063676,0.0002300277,0.007092268,0.002557705,0.0008058422,0.001152022,0.001546912,0.009541032,0.001701323,0.04990916,0.6387368,0.2856631],"study_design_scores_gemma":[0.001808335,0.00031828,0.008489384,0.001945634,0.0004145946,0.001306284,0.0009396913,0.153482,0.008812312,0.2991993,0.5227573,0.000527011],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003377392,0.0005403768,0.817338,0.001017423,0.0007103305,0.0007853135,0.0190836,0.1472867,0.009860763],"genre_scores_gemma":[0.03659078,0.0005524684,0.8358226,0.001332295,0.0005255504,0.004533869,0.01429301,0.09868921,0.007660161],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1395724,"threshold_uncertainty_score":0.4669162,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2073175998","doi":"10.1038/nmeth.2811","title":"BoxPlotR: a web tool for generation of box plots","year":2014,"lang":"en","type":"letter","venue":"Nature Methods","topic":"Data Analysis with R","field":"Computer Science","cited_by":875,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer","funders":"National Institutes of Health; Wellcome Trust","keywords":"Box plot; Plot (graphics); Data set; Bar chart; Scatter plot; Outlier; Range (aeronautics); Computer science; Quartile; Visualization; Statistics; Skewness; Mathematics; Data mining; Confidence interval","authors":[{"name":"Michaela Spitzer","is_ca":false},{"name":"Jan Wildenhain","is_ca":false},{"name":"Juri Rappsilber","is_ca":false},{"name":"Mike Tyers","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05445951663217588,"gpt":0.3831012404065347,"spread":0.3286417237743589,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01473814,0.001481157,0.001748784,0.003106412,0.0008600849,0.002840845,0.003845189,0.003317453,0.1230547],"category_scores_gemma":[0.07903335,0.001243887,0.001588077,0.002307457,0.001485848,0.002504504,0.002243929,0.007468408,0.08757894],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009352362,"about_ca_system_score_gemma":0.001586223,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001061483,"about_ca_topic_score_gemma":0.001741249,"domain_scores_codex":[0.99106,0.003886213,0.0008127135,0.0006883912,0.003235948,0.0003167247],"domain_scores_gemma":[0.9335502,0.04402788,0.002440814,0.009413307,0.008999484,0.001568248],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001996569,0.00002964976,0.0003875108,0.0003483723,0.00006219135,0.0002462903,0.000135474,0.0005075617,0.001684807,0.01074019,0.8835003,0.102158],"study_design_scores_gemma":[0.000256916,0.00008549215,0.0008974876,0.0004056354,0.00005434527,0.0009755362,0.00006941109,0.01237986,0.005281787,0.06896066,0.9104838,0.0001489442],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.00119462,0.001167644,0.8304226,0.02118235,0.008808195,0.0004663095,0.01142706,0.1073449,0.01798632],"genre_scores_gemma":[0.01882246,0.002082082,0.8302239,0.02091607,0.005500034,0.004157753,0.01293845,0.05780639,0.04755286],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.1230547,"threshold_uncertainty_score":0.4116589,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2808135930","doi":"10.1097/ede.0000000000000864","title":"Web Site and R Package for Computing E-values","year":2018,"lang":"en","type":"article","venue":"Epidemiology","topic":"Data Analysis with R","field":"Computer Science","cited_by":755,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Institute for Clinical Evaluative Sciences","funders":"National Institute of Environmental Health Sciences; National Cancer Institute; Medical Research Council","keywords":"Web site; Computer science; R package; Web application; World Wide Web; The Internet; Programming language","authors":[{"name":"Maya B. Mathur","is_ca":false},{"name":"Peng Ding","is_ca":true},{"name":"Corinne A. Riddell","is_ca":true},{"name":"Tyler J. VanderWeele","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05159941123569833,"gpt":0.3519743605208481,"spread":0.3003749492851498,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.01098348,0.002876232,0.00356428,0.004586567,0.0007028148,0.005762172,0.00476103,0.002542799,0.4035178],"category_scores_gemma":[0.1285763,0.001974882,0.002854591,0.005767884,0.001280549,0.003244381,0.003733625,0.004540125,0.2565515],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00122448,"about_ca_system_score_gemma":0.004179465,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002335967,"about_ca_topic_score_gemma":0.003285371,"domain_scores_codex":[0.9901557,0.004278892,0.001319024,0.001628659,0.002252734,0.0003650384],"domain_scores_gemma":[0.9111208,0.0688179,0.004708089,0.006995898,0.006664632,0.001692635],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001609432,0.00002291505,0.001194465,0.002092541,0.0002886127,0.000134801,0.00006762229,0.002090303,0.0001414543,0.004537531,0.9550526,0.03421628],"study_design_scores_gemma":[0.0004808358,0.00007101599,0.002708133,0.001872209,0.0002401881,0.0005505427,0.00008146319,0.01142354,0.000826046,0.0819729,0.8995717,0.0002014217],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"software","genre_scores_codex":[0.001551521,0.00265774,0.1667141,0.007344758,0.003821446,0.001183001,0.6323362,0.1623266,0.02206464],"genre_scores_gemma":[0.03586116,0.003713582,0.3951705,0.0125712,0.002351176,0.01077762,0.2872816,0.2139898,0.03828336],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.4035178,"threshold_uncertainty_score":0.8508094,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2924517740","doi":"10.1186/s12874-019-0666-3","title":"A review of spline function procedures in R","year":2019,"lang":"en","type":"review","venue":"BMC Medical Research Methodology","topic":"Data Analysis with R","field":"Computer Science","cited_by":577,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University Health Centre","funders":"","keywords":"Univariate; Computer science; Spline (mechanical); Observational study; Software; Data science; Regression analysis; Machine learning; Thin plate spline; Data mining; Management science; Statistics; Mathematics; Multivariate statistics; Spline interpolation; Engineering","authors":[{"name":"Aris Perperoglou","is_ca":false},{"name":"Willi Sauerbrei","is_ca":false},{"name":"Michał Abrahamowicz","is_ca":true},{"name":"Matthias Schmid","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7468500022474139,"gpt":0.6371901037000626,"spread":0.1096598985473514,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0149363,0.002391054,0.003142081,0.007711663,0.0006480801,0.002198479,0.003663397,0.003147384,0.01336427],"category_scores_gemma":[0.04285845,0.001399526,0.004109886,0.0114223,0.002702896,0.002302079,0.001539684,0.005448965,0.01519846],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001368254,"about_ca_system_score_gemma":0.004586164,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003130061,"about_ca_topic_score_gemma":0.001916147,"domain_scores_codex":[0.9877084,0.006514472,0.001668681,0.001042439,0.002853433,0.0002125712],"domain_scores_gemma":[0.9580916,0.03415959,0.001770761,0.001783761,0.003840757,0.0003535466],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007652606,0.00005935636,0.0004379016,0.02629793,0.0003223444,0.0001685566,0.0002393837,0.003722701,0.0005941755,0.03500368,0.08982168,0.8432557],"study_design_scores_gemma":[0.00004108881,0.0001167931,0.001076669,0.01100196,0.0001813451,0.0009115448,0.00006635038,0.001998233,0.00063143,0.04126847,0.9425708,0.0001353456],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0002674607,0.8710923,0.1144439,0.004800854,0.001618241,0.0001339801,0.001002244,0.001336544,0.005304567],"genre_scores_gemma":[0.003348768,0.9182253,0.06929962,0.002037794,0.002481518,0.0005684694,0.00115732,0.000939187,0.001942059],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.0149363,"threshold_uncertainty_score":0.07899165,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4297515751","doi":"","title":"glmulti: An R Package for Easy Automated Model Selection with (Generalized) Linear Models","year":2010,"lang":"en","type":"preprint","venue":"HAL (Le Centre pour la Communication Scientifique Directe)","topic":"Data Analysis with R","field":"Computer Science","cited_by":147,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"","keywords":"Selection (genetic algorithm); R package; Computer science; Mathematics; Artificial intelligence; Statistics","authors":[{"name":"Vincent Calcagno","is_ca":true},{"name":"Claire de Mazancourt","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02357698296147909,"gpt":0.2616525405533713,"spread":0.2380755575918922,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006821741,0.004162005,0.002860919,0.00362924,0.0007873905,0.003030913,0.004098609,0.001207124,0.07793045],"category_scores_gemma":[0.04032709,0.002161269,0.003539974,0.003278647,0.0008907252,0.002634119,0.003115604,0.00486896,0.05542272],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008092738,"about_ca_system_score_gemma":0.003919254,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004719668,"about_ca_topic_score_gemma":0.007943294,"domain_scores_codex":[0.9953098,0.002683631,0.000358429,0.0006466306,0.0007977856,0.0002037861],"domain_scores_gemma":[0.9806777,0.01424853,0.001299351,0.002059969,0.001467893,0.0002464838],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003073343,0.0001497856,0.004878785,0.002727224,0.001907534,0.0006894596,0.0004300951,0.04885023,0.003697607,0.04165212,0.6140383,0.2806715],"study_design_scores_gemma":[0.0004558336,0.0001836235,0.005348157,0.0005655942,0.0006622001,0.0008520846,0.000158177,0.2712963,0.00785201,0.1593507,0.5528926,0.0003829044],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.001223543,0.0005522819,0.8788904,0.0004931717,0.0001889828,0.0002239283,0.02904336,0.08703955,0.002344788],"genre_scores_gemma":[0.01323792,0.0006903997,0.8940255,0.0005169837,0.0001835579,0.002116586,0.03269894,0.05243907,0.004091057],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.07793045,"threshold_uncertainty_score":0.2607033,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2149851338","doi":"10.1111/2041-210x.12044","title":"Strategies for fitting nonlinear ecological models in <scp>R</scp>,<scp> AD M</scp>odel <scp>B</scp>uilder, and <scp>BUGS</scp>","year":2013,"lang":"en","type":"article","venue":"Methods in Ecology and Evolution","topic":"Data Analysis with R","field":"Computer Science","cited_by":123,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa; University of British Columbia Hospital; Dalhousie University; McMaster University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science","authors":[{"name":"Benjamin M. Bolker","is_ca":true},{"name":"Beth Gardner","is_ca":false},{"name":"Mark N. Maunder","is_ca":false},{"name":"Casper Willestofte Berg","is_ca":false},{"name":"M. Brooks","is_ca":false},{"name":"Liza S. Comita","is_ca":false},{"name":"Elizabeth E. Crone","is_ca":false},{"name":"Sarah Cubaynes","is_ca":false},{"name":"T. D. Davies","is_ca":true},{"name":"Perry de Valpine","is_ca":false},{"name":"Jessica H. Ford","is_ca":false},{"name":"Olivier Giménez","is_ca":false},{"name":"Marc Kéry","is_ca":false},{"name":"Eun Jung Kim","is_ca":false},{"name":"Cleridy E. Lennert‐Cody","is_ca":false},{"name":"Árni Magnússon","is_ca":false},{"name":"Steve Martell","is_ca":true},{"name":"John C. Nash","is_ca":true},{"name":"Anders Nielsen","is_ca":false},{"name":"Jim Regetz","is_ca":false},{"name":"Hans J. Skaug","is_ca":false},{"name":"Elise F. Zipkin","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03835212703600921,"gpt":0.3283769825518429,"spread":0.2900248555158337,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01812579,0.002997418,0.001540172,0.003874349,0.001305578,0.002881696,0.004778958,0.001821855,0.01875996],"category_scores_gemma":[0.087825,0.002668986,0.003365995,0.002607122,0.001413313,0.003486736,0.005515331,0.002960273,0.00883672],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001324071,"about_ca_system_score_gemma":0.002608468,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01032128,"about_ca_topic_score_gemma":0.01562326,"domain_scores_codex":[0.9942179,0.003346851,0.0005448157,0.0006445816,0.001048969,0.0001968525],"domain_scores_gemma":[0.9653109,0.02720154,0.001588349,0.002776576,0.002727941,0.0003946101],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002785179,0.0002204063,0.009380216,0.001281633,0.0004643836,0.0006165687,0.001615292,0.4384744,0.005873853,0.1138247,0.06726157,0.3607085],"study_design_scores_gemma":[0.00007300105,0.00005852584,0.001027513,0.0002060573,0.00007812928,0.0002881691,0.0002561917,0.8782219,0.005577929,0.09056868,0.02348744,0.0001564776],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001674803,0.00005773209,0.9918429,0.0002400426,0.00002265567,0.00005155082,0.0002552069,0.005036796,0.0008183206],"genre_scores_gemma":[0.02587122,0.0001530996,0.9653388,0.0002031888,0.0000224551,0.0005520295,0.000749153,0.005793347,0.001316716],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01875996,"threshold_uncertainty_score":0.09585947,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2412645763","doi":"10.1371/journal.pcbi.1004961","title":"Ten Simple Rules for Effective Statistical Practice","year":2016,"lang":"en","type":"editorial","venue":"PLoS Computational Biology","topic":"Data Analysis with R","field":"Computer Science","cited_by":116,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"National Cancer Institute; National Institute of Biomedical Imaging and Bioengineering; National Institute of Mental Health; National Institutes of Health; National Science Foundation","keywords":"Simple (philosophy); Praise; Computer science; Data science; Psychology; Epistemology; Philosophy; Social psychology","authors":[{"name":"Robert E. Kass","is_ca":false},{"name":"Brian Caffo","is_ca":false},{"name":"Marie Davidian","is_ca":false},{"name":"Xiao‐Li Meng","is_ca":false},{"name":"Bin Yu","is_ca":false},{"name":"Nancy Reid","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01164210549544279,"gpt":0.3312051418532067,"spread":0.3195630363577638,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2176358,0.003590869,0.003897604,0.007856688,0.006065942,0.0175355,0.006303997,0.01272525,0.01130122],"category_scores_gemma":[0.4089833,0.003247584,0.003567235,0.004124186,0.04713051,0.02113015,0.01237812,0.02792914,0.01211402],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007739708,"about_ca_system_score_gemma":0.01311345,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003055441,"about_ca_topic_score_gemma":0.001946557,"domain_scores_codex":[0.7206084,0.185535,0.02943367,0.01613517,0.04572589,0.002561911],"domain_scores_gemma":[0.5511985,0.3577601,0.01012619,0.04106195,0.03498268,0.004870585],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007220368,0.00008966615,0.0004610207,0.0005310449,0.0000849654,0.0001642877,0.001725873,0.002111637,0.0002195756,0.9111505,0.02867494,0.05471427],"study_design_scores_gemma":[0.00005567222,0.00004295911,0.00009765506,0.0006116112,0.00001416289,0.00007883693,0.0001440956,0.002780932,0.0001638834,0.9552527,0.04071651,0.00004103901],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"editorial","genre_scores_codex":[0.0009843514,0.004260262,0.9064968,0.05132526,0.003612265,0.0009708197,0.0002327344,0.001047769,0.03106974],"genre_scores_gemma":[0.0442964,0.003420837,0.9193746,0.01624819,0.00472971,0.005341978,0.0003627262,0.0007383475,0.005487145],"genre_candidate":"editorial","genre_consensus":null,"teacher_disagreement_score":0.7823642,"threshold_uncertainty_score":0.9647944,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4390461978","doi":"10.3929/ethz-b-000073064","title":"Nonparametric Econometrics: The np Package","year":2008,"lang":"en","type":"article","venue":"Repository for Publications and Research Data (ETH Zurich)","topic":"Data Analysis with R","field":"Computer Science","cited_by":100,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada","keywords":"Nonparametric statistics; Econometrics; R package; Statistics; Computer science; Mathematics; Economics","authors":[{"name":"Tristen Hayfield","is_ca":false},{"name":"Jeffrey S. Racine","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2633315052726449,"gpt":0.38542908937566,"spread":0.1220975841030151,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007041323,0.001568152,0.001508784,0.002871666,0.0005380759,0.002512068,0.00278492,0.001204079,0.1410555],"category_scores_gemma":[0.0567342,0.001188063,0.001792425,0.003412412,0.0007911405,0.002389696,0.003029875,0.002989628,0.07682318],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005993259,"about_ca_system_score_gemma":0.002759182,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003159806,"about_ca_topic_score_gemma":0.003465094,"domain_scores_codex":[0.9947577,0.002809203,0.000413003,0.000586085,0.001221101,0.0002128418],"domain_scores_gemma":[0.9708403,0.0207179,0.001202825,0.004390705,0.002495561,0.0003526986],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001416878,0.00007875192,0.00274998,0.001143586,0.0003595284,0.0002194968,0.0002075698,0.01388719,0.0009269397,0.08357364,0.5931501,0.3035615],"study_design_scores_gemma":[0.0001352016,0.00006341668,0.003291982,0.0003041262,0.0001162001,0.0005818394,0.00007065971,0.05613734,0.001404903,0.179815,0.7579677,0.0001116899],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.000739853,0.0005761789,0.9239125,0.0008506809,0.0002820167,0.0002527823,0.02928039,0.03277367,0.01133182],"genre_scores_gemma":[0.02034033,0.001436232,0.8719736,0.001244834,0.0005155392,0.003835489,0.03625137,0.04740847,0.01699422],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.1410555,"threshold_uncertainty_score":0.4718777,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2128422078","doi":"10.4310/sii.2016.v9.n4.a1","title":"Statistical methods and computing for big data","year":2016,"lang":"en","type":"article","venue":"Statistics and Its Interface","topic":"Data Analysis with R","field":"Computer Science","cited_by":100,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Institute of General Medical Sciences; National Cancer Institute; Banff International Research Station for Mathematical Innovation and Discovery; National Institutes of Health; National Science Foundation","keywords":"Big data; Computational statistics; Software; Scale (ratio); Selection (genetic algorithm); Data stream mining; Feature selection","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.1010523441108938,"gpt":0.4147756650059232,"spread":0.3137233208950294,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03213668,0.003615539,0.004501236,0.008764946,0.001403502,0.00868287,0.004737963,0.004283709,0.02024074],"category_scores_gemma":[0.1192265,0.00178292,0.002934332,0.01095294,0.007374236,0.007412996,0.00650194,0.01139254,0.0178189],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002176724,"about_ca_system_score_gemma":0.007333084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001822258,"about_ca_topic_score_gemma":0.001223151,"domain_scores_codex":[0.9621179,0.02456734,0.003063285,0.003130852,0.006640645,0.000480041],"domain_scores_gemma":[0.8642259,0.1061578,0.004045074,0.01751081,0.00658792,0.001472359],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001234788,0.00008229526,0.001188677,0.003747583,0.000696582,0.0004643217,0.0008640618,0.01111404,0.0009439522,0.630055,0.1250202,0.2256998],"study_design_scores_gemma":[0.00006240916,0.00004727615,0.0005221869,0.0008505712,0.00006776596,0.0002442632,0.0001158743,0.0140261,0.0003484969,0.7972913,0.1863542,0.00006955688],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0004380588,0.01513737,0.9534352,0.009525912,0.002617168,0.0005103017,0.002046578,0.005093526,0.01119595],"genre_scores_gemma":[0.01515337,0.0205534,0.9402451,0.00390729,0.005675843,0.004494267,0.002752846,0.002644511,0.004573401],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03213668,"threshold_uncertainty_score":0.169957,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2792985690","doi":"10.1002/ece3.3807","title":"Count data in biology—Data transformation or model reformation?","year":2018,"lang":"en","type":"article","venue":"Ecology and Evolution","topic":"Data Analysis with R","field":"Computer Science","cited_by":97,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Memorial University of Newfoundland","funders":"Natural Sciences and Engineering Research Council of Canada; Memorial University of Newfoundland","keywords":"Residual; Transformation (genetics); Statistics; Generalized linear model; Econometrics; Type I and type II errors; Data transformation; Statistical model; Plot (graphics); Mathematics; Computer science; Data mining; Algorithm; Biology","authors":[{"name":"Anne P. St‐Pierre","is_ca":true},{"name":"Violaine Shikon","is_ca":true},{"name":"David C. Schneider","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04467150556521438,"gpt":0.3145071598013819,"spread":0.2698356542361676,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.1548434,0.001398794,0.003547902,0.006166765,0.001201069,0.008210621,0.005592268,0.003308187,0.005523337],"category_scores_gemma":[0.5115272,0.001084529,0.002457068,0.01371547,0.01019885,0.01105425,0.004433412,0.00947213,0.002250229],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004276687,"about_ca_system_score_gemma":0.00707806,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002497869,"about_ca_topic_score_gemma":0.001833986,"domain_scores_codex":[0.8094844,0.1443013,0.01928569,0.008776684,0.01732969,0.0008222694],"domain_scores_gemma":[0.5019865,0.3926686,0.03660547,0.03253052,0.03535254,0.0008562313],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002612649,0.00009577344,0.005980463,0.02124841,0.001101158,0.0003906593,0.006172739,0.003753648,0.001188187,0.3520138,0.08102155,0.5267724],"study_design_scores_gemma":[0.0001037348,0.0002891406,0.00534957,0.03308535,0.0006803187,0.0006497438,0.003217889,0.008681926,0.003815134,0.4939073,0.4499199,0.0003001187],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00755273,0.1603243,0.6393887,0.1626005,0.01675467,0.0007497575,0.001887898,0.001212808,0.009528476],"genre_scores_gemma":[0.1678152,0.1544826,0.5777149,0.07138582,0.01591754,0.004082415,0.002559663,0.001981799,0.004060132],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8451567,"threshold_uncertainty_score":0.8188998,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W597599483","doi":"10.1017/cbo9780511803642","title":"A First Course in Statistical Programming with R","year":2007,"lang":"en","type":"book","venue":"Cambridge University Press eBooks","topic":"Data Analysis with R","field":"Computer Science","cited_by":92,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia; Western University","funders":"","keywords":"Computer science; Download; Open source; Code (set theory); Course (navigation); Source code; Programming language; Software engineering; World Wide Web; Core (optical fiber); C programming language; Software; Engineering; Set (abstract data type)","authors":[{"name":"W. John Braun","is_ca":true},{"name":"Duncan J. Murdoch","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01447406246458246,"gpt":0.2168288740390328,"spread":0.2023548115744504,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003318429,0.002056181,0.002254209,0.002127594,0.00087396,0.004097669,0.002334161,0.001854806,0.1245518],"category_scores_gemma":[0.01425403,0.001494832,0.001997601,0.004305717,0.001298821,0.004135857,0.002251644,0.008451879,0.1866283],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009299434,"about_ca_system_score_gemma":0.00218787,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008518273,"about_ca_topic_score_gemma":0.001578405,"domain_scores_codex":[0.9971994,0.0006437257,0.0001938828,0.0004567019,0.001423116,0.00008320175],"domain_scores_gemma":[0.9908159,0.005961441,0.0002493294,0.001084403,0.001525607,0.0003633743],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001320172,0.00003043181,0.000102071,0.0003639727,0.00002932947,0.0000833468,0.0001539354,0.0005301653,0.0003081725,0.02347452,0.7885889,0.1863219],"study_design_scores_gemma":[0.00001085158,0.00001613442,0.0002197457,0.0002063454,0.000007373933,0.0002661948,0.00003549914,0.0007564862,0.0001794214,0.02790936,0.9703755,0.00001698174],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0006124079,0.02953617,0.5699868,0.02799328,0.008241773,0.0004743486,0.008158593,0.0402372,0.3147593],"genre_scores_gemma":[0.005016742,0.03217293,0.5384711,0.01729304,0.006040317,0.001484565,0.009712599,0.02466271,0.365146],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1245518,"threshold_uncertainty_score":0.4166671,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2156242793","doi":"10.1093/aje/kwr241","title":"The Quality of Modern Cross-Sectional Ecologic Studies: A Bibliometric Review","year":2011,"lang":"en","type":"review","venue":"American Journal of Epidemiology","topic":"Data Analysis with R","field":"Computer Science","cited_by":75,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba","funders":"","keywords":"Cross-sectional study; Medicine; Fallacy; MEDLINE; Epidemiology; Clinical study design; Environmental health; Pathology; Biology","authors":[{"name":"Brenden Dufault","is_ca":true},{"name":"Neil Klar","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.4254039160845181,"gpt":0.5413397750599583,"spread":0.1159358589754403,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.087947,0.001504447,0.009383977,0.1475083,0.001706295,0.008743306,0.00256135,0.001627513,0.002374936],"category_scores_gemma":[0.3331026,0.001627481,0.004680176,0.1400145,0.00280964,0.007933167,0.00334563,0.001046575,0.0004241054],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.008500217,"about_ca_system_score_gemma":0.01919953,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006209141,"about_ca_topic_score_gemma":0.01193002,"domain_scores_codex":[0.8950565,0.03389469,0.04034218,0.003868366,0.02604083,0.0007974066],"domain_scores_gemma":[0.5862548,0.2975948,0.05098032,0.00946711,0.05404759,0.001655382],"domain_codex":null,"domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.0002702883,0.00005401787,0.02268403,0.5295187,0.01247584,0.0002340875,0.002207766,0.0005288065,0.0004625002,0.002386316,0.01239924,0.4167785],"study_design_scores_gemma":[0.00033937,0.0005233633,0.1426646,0.6344352,0.06546241,0.002056969,0.006100829,0.001639293,0.001449416,0.007851873,0.137052,0.0004247691],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00625521,0.9856806,0.001558312,0.002793199,0.0004541727,0.000508922,0.001199769,0.00005718657,0.001492708],"genre_scores_gemma":[0.06133185,0.928612,0.00612866,0.0006365311,0.0005009191,0.001097783,0.001456416,0.00003431584,0.0002015181],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.912053,"threshold_uncertainty_score":0.4651138,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2941763735","doi":"10.1002/ece3.6747","title":"A contrast of meta and metafor packages for meta‐analyses in R","year":2020,"lang":"en","type":"article","venue":"Ecology and Evolution","topic":"Data Analysis with R","field":"Computer Science","cited_by":69,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta; York University","funders":"","keywords":"Contrast (vision); Benchmark (surveying); Consistency (knowledge bases); Computer science; Meta-analysis; Data science; Checklist; Statistics; Artificial intelligence; Psychology; Cartography; Geography; Cognitive psychology; Mathematics","authors":[{"name":"Christopher J. Lortie","is_ca":true},{"name":"Alessandro Filazzola","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.09534385885876914,"gpt":0.320078127650376,"spread":0.2247342687916068,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.05849527,0.002467182,0.002908581,0.005930658,0.000939102,0.003991074,0.005113258,0.002978691,0.1131976],"category_scores_gemma":[0.221632,0.002248906,0.008903545,0.006846345,0.001511711,0.004010929,0.005375836,0.007118405,0.02848328],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001067479,"about_ca_system_score_gemma":0.004370709,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001392745,"about_ca_topic_score_gemma":0.002068253,"domain_scores_codex":[0.947385,0.04010976,0.004672979,0.003560698,0.00356312,0.0007084427],"domain_scores_gemma":[0.7631064,0.1838756,0.009491289,0.03420041,0.008003229,0.001323065],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002249333,0.0001714446,0.004524572,0.01999651,0.01552217,0.0006893473,0.001221018,0.005795734,0.003407125,0.07641234,0.704021,0.1659894],"study_design_scores_gemma":[0.001483287,0.0004286175,0.007639417,0.005114645,0.007253021,0.001328022,0.0002230134,0.01850023,0.00656238,0.1778147,0.7730343,0.000618391],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001803701,0.002667068,0.8779417,0.004265167,0.002972754,0.001480528,0.03206214,0.07059684,0.006210183],"genre_scores_gemma":[0.02211155,0.001236503,0.9031317,0.002839372,0.0008081176,0.01361648,0.007268121,0.04389561,0.005092518],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9415047,"threshold_uncertainty_score":0.3786836,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W123767535","doi":"","title":"GETTING STARTED WITH THE R COMMANDER: A BASIC-STATISTICS GRAPHICAL USER INTERFACE TO R","year":2004,"lang":"en","type":"article","venue":"","topic":"Data Analysis with R","field":"Computer Science","cited_by":66,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University","funders":"","keywords":"Dialog box; Computer science; Graphical user interface; Interface (matter); Window (computing); Programming language; User interface; Human–computer interaction; World Wide Web; Operating system","authors":[{"name":"John Fox","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01010641793212767,"gpt":0.2466048792513933,"spread":0.2364984613192656,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.01043141,0.003016535,0.002885503,0.003211063,0.0006561633,0.0030487,0.004838484,0.001911078,0.3374586],"category_scores_gemma":[0.05347888,0.001841045,0.001763694,0.002785523,0.001118755,0.003514712,0.003031518,0.005155357,0.2478825],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007816159,"about_ca_system_score_gemma":0.002401806,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001348168,"about_ca_topic_score_gemma":0.001330038,"domain_scores_codex":[0.9944523,0.002617961,0.0007190821,0.0007973379,0.001165388,0.0002477892],"domain_scores_gemma":[0.9659015,0.02442911,0.001256282,0.002984026,0.004325467,0.001103639],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004473406,0.00005803347,0.0006375028,0.001340839,0.0001606259,0.0003478996,0.0002669235,0.00118763,0.003551587,0.005470511,0.902329,0.08420208],"study_design_scores_gemma":[0.000920396,0.0001964648,0.003056466,0.0007168982,0.000127934,0.0008677091,0.0001297533,0.01783564,0.009711803,0.02673604,0.9393836,0.0003172535],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.000989537,0.0005541501,0.5007517,0.001397315,0.001076018,0.001152605,0.05474004,0.4243371,0.0150016],"genre_scores_gemma":[0.01628249,0.0009042829,0.7126446,0.003132783,0.0008106744,0.007614354,0.04714324,0.1857663,0.02570128],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3374586,"threshold_uncertainty_score":0.9450349,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1968805547","doi":"10.1080/02664760500168648","title":"Interpretable dimension reduction","year":2005,"lang":"en","type":"article","venue":"Journal of Applied Statistics","topic":"Data Analysis with R","field":"Computer Science","cited_by":62,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University; Acadia University","funders":"University of Waterloo; Natural Sciences and Engineering Research Council of Canada; Mitacs","keywords":"Principal component analysis; Linear subspace; Dimensionality reduction; Constraint (computer-aided design); Dimension (graph theory); Mathematics; Homogeneity (statistics); Reduction (mathematics); Mathematical optimization; Computer science; Algorithm; Statistics; Artificial intelligence; Combinatorics","authors":[{"name":"Hugh Chipman","is_ca":true},{"name":"Hong Gu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.007379734791390265,"gpt":0.2424087083072268,"spread":0.2350289735158365,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004539817,0.002256584,0.001329491,0.003719404,0.001000033,0.005507966,0.001741806,0.001108425,0.01455543],"category_scores_gemma":[0.02320533,0.0006157265,0.001709255,0.00270179,0.001626097,0.003257933,0.003373345,0.002714315,0.005957552],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008130361,"about_ca_system_score_gemma":0.001291273,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007465695,"about_ca_topic_score_gemma":0.0008771886,"domain_scores_codex":[0.9955505,0.001577929,0.0003673785,0.0009438913,0.001392362,0.0001678602],"domain_scores_gemma":[0.9916182,0.00289448,0.000493999,0.00272031,0.00215013,0.000122902],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002982156,0.0001576274,0.002631839,0.001200439,0.0002732303,0.000501393,0.001155769,0.03124949,0.0147182,0.3275157,0.0608918,0.5594063],"study_design_scores_gemma":[0.00007758653,0.0000988991,0.002345499,0.0003877078,0.0001362837,0.0004575988,0.0007361236,0.175672,0.01053574,0.6885085,0.1209311,0.0001129972],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006198063,0.001573205,0.971693,0.001945822,0.0005865162,0.0002142092,0.002882488,0.002058537,0.01284824],"genre_scores_gemma":[0.1113616,0.002700404,0.8652505,0.0007680852,0.0006558154,0.0008057467,0.008748297,0.0007441205,0.008965378],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01455543,"threshold_uncertainty_score":0.04869276,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4294557430","doi":"10.18637/jss.v103.i07","title":"Hierarchical Clustering with Contiguity Constraint in <i>R</i>","year":2022,"lang":"en","type":"article","venue":"Journal of Statistical Software","topic":"Data Analysis with R","field":"Computer Science","cited_by":51,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Contiguity; Cluster analysis; Computer science; Hierarchical clustering; Theoretical computer science; Function (biology); Algorithm; Artificial intelligence","authors":[{"name":"Guillaume Guénard","is_ca":true},{"name":"Pierre Legendre","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01225028693388872,"gpt":0.2517438408299075,"spread":0.2394935538960188,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01130151,0.003643523,0.002984426,0.003669961,0.001745717,0.004548099,0.005311263,0.002352357,0.0435128],"category_scores_gemma":[0.05382933,0.002597226,0.004660007,0.005237719,0.001600622,0.003077168,0.00450451,0.004697212,0.03803607],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001638057,"about_ca_system_score_gemma":0.00480534,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009737865,"about_ca_topic_score_gemma":0.009413737,"domain_scores_codex":[0.9891291,0.004547297,0.0009556154,0.002593085,0.00225427,0.0005206959],"domain_scores_gemma":[0.9852635,0.008174986,0.001163093,0.003032386,0.002086699,0.0002793679],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007080929,0.0001288009,0.005617324,0.003013858,0.001805337,0.0009775717,0.001835566,0.08054757,0.01352903,0.1349473,0.5103124,0.2465771],"study_design_scores_gemma":[0.0003224497,0.0001280038,0.004404438,0.0005860876,0.000429847,0.001146497,0.0001943463,0.2863388,0.024509,0.1772109,0.5041659,0.0005637879],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0009822333,0.0003303881,0.94249,0.0003120165,0.000241352,0.00009806934,0.005866027,0.0473778,0.002302185],"genre_scores_gemma":[0.01700724,0.0002795323,0.9340979,0.0006399167,0.0001424386,0.001499773,0.01196198,0.03120352,0.003167594],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0435128,"threshold_uncertainty_score":0.1455648,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3044013993","doi":"10.1098/rsos.200566","title":"Are replication rates the same across academic fields? Community forecasts from the DARPA SCORE programme","year":2020,"lang":"en","type":"article","venue":"Royal Society Open Science","topic":"Data Analysis with R","field":"Computer Science","cited_by":46,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Bishop's University","funders":"Defense Advanced Research Projects Agency; Royal Marsden NHS Foundation Trust","keywords":"Replication (statistics); Confidence interval; Agency (philosophy); Psychology; Empirical research; Computer science; Applied psychology; Statistics; Social science; Sociology; Mathematics","authors":[{"name":"Michael Gordon","is_ca":false},{"name":"Domenico Viganola","is_ca":false},{"name":"Michaël Bishop","is_ca":true},{"name":"Yiling Chen","is_ca":false},{"name":"Anna Dreber","is_ca":false},{"name":"Brandon Goldfedder","is_ca":false},{"name":"Felix Holzmeister","is_ca":false},{"name":"Magnus Johannesson","is_ca":false},{"name":"Yang Liu","is_ca":false},{"name":"Charles Twardy","is_ca":false},{"name":"Juntao Wang","is_ca":false},{"name":"Thomas Pfeiffer","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1294043474949684,"gpt":0.3691238483329833,"spread":0.239719500838015,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.3812931,0.0007245408,0.001341707,0.006926184,0.001524136,0.006150196,0.002953087,0.003608983,0.006114711],"category_scores_gemma":[0.7640811,0.000768487,0.001642491,0.005540929,0.003501787,0.005561632,0.004288882,0.004507913,0.001480381],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003403825,"about_ca_system_score_gemma":0.004359134,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01007945,"about_ca_topic_score_gemma":0.006565472,"domain_scores_codex":[0.7031236,0.1846156,0.01724731,0.01754883,0.07285583,0.004608871],"domain_scores_gemma":[0.1365976,0.6507133,0.05942829,0.04705375,0.1007606,0.005446319],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.004626473,0.0004314112,0.3064187,0.002997234,0.001302477,0.0005534777,0.02264603,0.02153889,0.002542656,0.09142739,0.1463748,0.3991405],"study_design_scores_gemma":[0.001333861,0.002136235,0.4116193,0.005741935,0.0009150083,0.0009999067,0.01497415,0.1700729,0.005441796,0.2529131,0.1325169,0.001334879],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.3858862,0.0113325,0.366305,0.1341956,0.003953179,0.003670399,0.01568227,0.001884097,0.07709063],"genre_scores_gemma":[0.9317612,0.0009880677,0.05519257,0.004077249,0.000764754,0.001470385,0.003374071,0.0003081656,0.002063544],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6187069,"threshold_uncertainty_score":0.7629758,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2924029102","doi":"10.1080/00031305.2018.1556735","title":"The <i>p</i> -value Function and Statistical Inference","year":2019,"lang":"en","type":"article","venue":"The American Statistician","topic":"Data Analysis with R","field":"Computer Science","cited_by":45,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; York University","keywords":"Inference; Function (biology); Statistical inference; Scalar (mathematics); Value (mathematics); Mathematics; Power function; Statistics; Computer science; Artificial intelligence; Mathematical analysis","authors":[{"name":"D. A. S. Fraser","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.006202599462732283,"gpt":0.2597219331783565,"spread":0.2535193337156242,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.07861543,0.002905675,0.00345857,0.007757063,0.001840005,0.0096771,0.004730791,0.007528342,0.005974737],"category_scores_gemma":[0.2565901,0.001677231,0.00283715,0.01066488,0.0293203,0.01417275,0.005860439,0.01730083,0.004201103],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004082754,"about_ca_system_score_gemma":0.004476569,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002511118,"about_ca_topic_score_gemma":0.000838501,"domain_scores_codex":[0.9112544,0.07149482,0.004101017,0.005009397,0.007438313,0.0007020053],"domain_scores_gemma":[0.6925891,0.2754994,0.007007039,0.0180263,0.006028975,0.0008491282],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00004703457,0.00003157583,0.001069679,0.0005005417,0.0001247857,0.0003581523,0.0004074286,0.007111423,0.0002581846,0.9217675,0.01292252,0.05540127],"study_design_scores_gemma":[0.00001994555,0.00004360947,0.0004041245,0.0003301808,0.00003259617,0.0005109682,0.00009542533,0.01825319,0.0005013201,0.9543754,0.02537214,0.00006113497],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0007973434,0.004654926,0.9780496,0.007322507,0.0007413798,0.00008503569,0.0002211768,0.0004548545,0.00767331],"genre_scores_gemma":[0.05533237,0.007545006,0.9216154,0.005638658,0.003588002,0.001423742,0.0004180337,0.0009243779,0.003514462],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.07861543,"threshold_uncertainty_score":0.4157631,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2277736791","doi":"10.20982/tqmp.09.2.p043","title":"Conducting Simulation Studies in the R Programming Environment","year":2013,"lang":"en","type":"article","venue":"Tutorials in Quantitative Methods for Psychology","topic":"Data Analysis with R","field":"Computer Science","cited_by":40,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"National Institute on Alcohol Abuse and Alcoholism","keywords":"Computer science; Bootstrapping (finance); Variety (cybernetics); Syntax; Simulation modeling; Data science; Management science; Artificial intelligence; Econometrics; Mathematics; Engineering","authors":[{"name":"Kevin A. Hallgren","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.5490431352157252,"gpt":0.6000822268446929,"spread":0.05103909162896769,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02609685,0.002697384,0.002314995,0.003245889,0.0007384359,0.003138513,0.003013438,0.001662458,0.05711316],"category_scores_gemma":[0.1254196,0.001903347,0.002561037,0.003380789,0.001719741,0.002856069,0.002826113,0.00579378,0.02750191],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009138623,"about_ca_system_score_gemma":0.004626162,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001631651,"about_ca_topic_score_gemma":0.001972754,"domain_scores_codex":[0.9789563,0.01456964,0.002122189,0.001432761,0.002481701,0.000437387],"domain_scores_gemma":[0.8357876,0.1415639,0.004849197,0.01053478,0.006271438,0.0009930457],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001190372,0.0005416052,0.006144005,0.009510075,0.001576666,0.00108608,0.002481051,0.0560637,0.004993957,0.1340386,0.4788314,0.3035425],"study_design_scores_gemma":[0.001034461,0.0004836603,0.003493124,0.00237161,0.0004839621,0.001256164,0.0006137811,0.1639994,0.01187077,0.2228745,0.5911044,0.0004140802],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002078692,0.0003781763,0.9394004,0.001055616,0.0004583347,0.001572267,0.006443461,0.03915475,0.009458426],"genre_scores_gemma":[0.01284065,0.0005345613,0.9606935,0.0005671452,0.0001397463,0.00751359,0.00353099,0.01117514,0.003004626],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05711316,"threshold_uncertainty_score":0.1910626,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W7074663863","doi":"","title":"Our Own Master Race: Eugenics in Canada, 1885-1945","year":2023,"lang":"en","type":"article","venue":"Project Muse (Johns Hopkins University)","topic":"Data Analysis with R","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Eugenics; George (robot); Subject (documents); Agency (philosophy); Perspective (graphical); Argument (complex analysis)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.02700623818115424,"gpt":0.2251063727721264,"spread":0.1981001345909722,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009395289,0.0002244476,0.000472092,0.002602267,0.01164318,0.002935823,0.001100017,0.0009665729,0.01249099],"category_scores_gemma":[0.004964544,0.0001957292,0.0002620391,0.007512277,0.004197043,0.0009898522,0.001832344,0.002168112,0.0004536038],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.08006196,"about_ca_system_score_gemma":0.07510572,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.9961313,"about_ca_topic_score_gemma":0.9986309,"domain_scores_codex":[0.9982713,0.0001232953,0.00003117723,0.0001816317,0.0004962583,0.0008962858],"domain_scores_gemma":[0.9975349,0.0002175448,0.0001622278,0.00006680024,0.001337866,0.0006807089],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0005146449,0.0001401529,0.2076248,0.0003397402,0.0001468428,0.001045181,0.2056952,0.001076396,0.000453854,0.2095983,0.194752,0.1786129],"study_design_scores_gemma":[0.00001976121,0.00003288731,0.4586695,0.0003788522,0.00003853572,0.0001247232,0.06350925,0.000237836,0.0002281091,0.003679476,0.4730203,0.00006065492],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7064933,0.01050999,0.0004537548,0.01726099,0.0004451965,0.00008815875,0.01397501,0.00006297124,0.2507107],"genre_scores_gemma":[0.9503027,0.001963897,0.0002897152,0.0008254097,0.00006521979,0.0000251057,0.001269703,0.00002882376,0.04522941],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.08006196,"threshold_uncertainty_score":0.5808929,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2161482435","doi":"10.1890/04-1655","title":"SAMPLING VARIABILITY AND ESTIMATES OF DENSITY DEPENDENCE: A COMPOSITE-LIKELIHOOD APPROACH","year":2006,"lang":"en","type":"article","venue":"Ecology","topic":"Data Analysis with R","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Sampling (signal processing); Statistics; Population; Inference; Statistical inference; Series (stratigraphy); Computer science; Econometrics; Mathematics; Ecology; Artificial intelligence","authors":[{"name":"Subhash R. Lele","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01214202770356461,"gpt":0.2457870415087425,"spread":0.2336450138051779,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01489853,0.0009391977,0.001607066,0.003363833,0.0007392516,0.002510929,0.003705812,0.001494028,0.001867992],"category_scores_gemma":[0.05827821,0.00110782,0.001923243,0.003054269,0.001975928,0.003472075,0.002999237,0.003458499,0.0004584398],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001167876,"about_ca_system_score_gemma":0.001415827,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00343388,"about_ca_topic_score_gemma":0.002466842,"domain_scores_codex":[0.9937926,0.003819933,0.0002237439,0.0008662958,0.001104659,0.0001928397],"domain_scores_gemma":[0.9635214,0.0300041,0.001858319,0.002470997,0.001807761,0.0003373745],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002490153,0.0001504168,0.02742288,0.0005663565,0.0007282356,0.0006492262,0.0009177206,0.3513455,0.002746454,0.3443956,0.003376212,0.2674523],"study_design_scores_gemma":[0.00001830841,0.00003954077,0.003949911,0.00005155112,0.00004885479,0.0002468624,0.00006569004,0.8597531,0.0005668639,0.1322493,0.002954309,0.0000556644],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002477681,0.0002141102,0.996713,0.00007913364,0.00001194499,0.00002021391,0.00004484036,0.00007450816,0.0003645916],"genre_scores_gemma":[0.2176563,0.001116478,0.7773281,0.0001809599,0.0002513887,0.0004066176,0.000801381,0.0002282995,0.002030488],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01489853,"threshold_uncertainty_score":0.07879186,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4401042314","doi":"10.18653/v1/2024.naacl-long.335","title":"TableLlama: Towards Open Large Generalist Models for Tables","year":2024,"lang":"en","type":"article","venue":"","topic":"Data Analysis with R","field":"Computer Science","cited_by":35,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Atomic Energy of Canada Limited; National Science Foundation","keywords":"Generalist and specialist species; Computer science; Programming language","authors":[{"name":"Tianshu Zhang","is_ca":false},{"name":"Xiang Yue","is_ca":false},{"name":"Yifei Li","is_ca":false},{"name":"Huan Sun","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05470949858411823,"gpt":0.320992085302541,"spread":0.2662825867184228,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004811309,0.002204619,0.002842696,0.002666905,0.00186459,0.005978984,0.006319854,0.003336729,0.01886741],"category_scores_gemma":[0.02368053,0.002635554,0.004965674,0.004394806,0.001266057,0.01537734,0.005967093,0.005675209,0.01480866],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001597237,"about_ca_system_score_gemma":0.0029138,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00876628,"about_ca_topic_score_gemma":0.02180117,"domain_scores_codex":[0.9977361,0.0009458332,0.0001628946,0.0006092994,0.00040506,0.0001406975],"domain_scores_gemma":[0.9919491,0.004751078,0.0002120022,0.002203107,0.0006764278,0.0002082696],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001443733,0.0002783434,0.003763047,0.001252802,0.00109383,0.0004129604,0.0007007611,0.1145093,0.002491188,0.1631981,0.3478509,0.363005],"study_design_scores_gemma":[0.0001663487,0.00004251576,0.000232328,0.00009371543,0.00009988713,0.00008440251,0.00007305774,0.6370248,0.0009384486,0.3103759,0.05082304,0.00004563581],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0033843,0.00177129,0.9221933,0.001088642,0.0003741171,0.0001421281,0.01146565,0.05697042,0.002610062],"genre_scores_gemma":[0.08148012,0.001501081,0.8652695,0.001230681,0.0004197564,0.0008692423,0.03478044,0.008564374,0.005884758],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01886741,"threshold_uncertainty_score":0.06311774,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2683414882","doi":"10.1101/156067","title":"BPG: Seamless, Automated and Interactive Visualization of Scientific Data","year":2017,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Data Analysis with R","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Ontario Institute for Cancer Research","funders":"Natural Sciences and Engineering Research Council of Canada; University of Toronto; Government of Ontario; Canadian Institutes of Health Research; National Science Foundation; Ontario Institute for Cancer Research; University of Pennsylvania; Center for Translational Molecular Medicine; Ontario Genomics Institute; Movember Foundation; Ontario Genomics; Genome Canada; Prostate Cancer Canada","keywords":"Computer science; Personalization; Visualization; Open source; Software; Interactive visualization; Software engineering; Source code; Web application; Data mining; Data science; World Wide Web; Programming language","authors":[{"name":"Christine P’ng","is_ca":true},{"name":"Jeffrey Green","is_ca":true},{"name":"Lauren C. Chong","is_ca":true},{"name":"Daryl Waggott","is_ca":true},{"name":"Stephenie D. Prokopec","is_ca":true},{"name":"Mehrdad Shamsi","is_ca":true},{"name":"Francis Nguyen","is_ca":true},{"name":"Denise Mak","is_ca":true},{"name":"Felix Lam","is_ca":true},{"name":"Marco Albuquerque","is_ca":true},{"name":"Ying Wu","is_ca":true},{"name":"Esther H. Jung","is_ca":true},{"name":"Maud H. W. Starmans","is_ca":true},{"name":"Michelle Chan‐Seng‐Yue","is_ca":true},{"name":"Cindy Q. Yao","is_ca":true},{"name":"Bianca Liang","is_ca":true},{"name":"Emilie Lalonde","is_ca":true},{"name":"Syed Haider","is_ca":true},{"name":"Nicole A. Simone","is_ca":true},{"name":"Dorota H.S. Sendorek","is_ca":true},{"name":"Kenneth C. Chu","is_ca":true},{"name":"Nathalie C. Moon","is_ca":true},{"name":"Natalie S. Fox","is_ca":true},{"name":"Michal R. Grzadkowski","is_ca":true},{"name":"Nicholas J. Harding","is_ca":true},{"name":"Clement Fung","is_ca":true},{"name":"Amanda R. Murdoch","is_ca":true},{"name":"Kathleen E. Houlahan","is_ca":true},{"name":"Jianxin Wang","is_ca":true},{"name":"David R. Garcia","is_ca":true},{"name":"Richard de Borja","is_ca":true},{"name":"Ren Sun","is_ca":true},{"name":"Xihui Lin","is_ca":true},{"name":"Gregory M. Chen","is_ca":true},{"name":"Aileen Lu","is_ca":true},{"name":"Yu-Jia Shiah","is_ca":true},{"name":"Amin Zia","is_ca":true},{"name":"Ryan Kearns","is_ca":true},{"name":"Paul C. Boutros","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03137588375988596,"gpt":0.2915388983916274,"spread":0.2601630146317415,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007895345,0.003104395,0.002005001,0.00576171,0.0008494764,0.00531368,0.003759776,0.00152354,0.04581551],"category_scores_gemma":[0.02070076,0.001692742,0.001808538,0.003863494,0.001002425,0.003702286,0.006577908,0.005170556,0.02875731],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006782832,"about_ca_system_score_gemma":0.002346326,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002414,"about_ca_topic_score_gemma":0.002988305,"domain_scores_codex":[0.9958044,0.001260548,0.0002978689,0.0006313692,0.001741401,0.0002645877],"domain_scores_gemma":[0.9928646,0.00327107,0.0004919288,0.001663898,0.001152414,0.0005560254],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005206978,0.0001134009,0.001762649,0.001459045,0.0003851975,0.0006303428,0.0006635184,0.006721762,0.02045306,0.01374802,0.783173,0.1703693],"study_design_scores_gemma":[0.0008119012,0.0001220414,0.006940981,0.0008250266,0.0001766159,0.00123484,0.000211702,0.1524285,0.04554805,0.1078506,0.6833078,0.0005419177],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.001775214,0.0004337019,0.5686455,0.001177155,0.0004390848,0.0002085968,0.02053055,0.4042197,0.002570507],"genre_scores_gemma":[0.02604062,0.001084385,0.8394297,0.001150797,0.0003168506,0.00142902,0.03037593,0.09647971,0.003692972],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.04581551,"threshold_uncertainty_score":0.1532681,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4245274090","doi":"10.1201/9781315380537","title":"Using the R Commander","year":2016,"lang":"en","type":"book","venue":"","topic":"Data Analysis with R","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University","funders":"","keywords":"History; Computer science","authors":[{"name":"John P. Fox","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06007906383447632,"gpt":0.2889410885068781,"spread":0.2288620246724018,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007873406,0.002723718,0.002522531,0.005287232,0.001176401,0.005557615,0.004183636,0.00212475,0.2061614],"category_scores_gemma":[0.03590275,0.002345242,0.002038002,0.005115251,0.001432352,0.005518949,0.004048994,0.006159857,0.3381047],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000994683,"about_ca_system_score_gemma":0.003324623,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002474935,"about_ca_topic_score_gemma":0.00256479,"domain_scores_codex":[0.9881182,0.003575543,0.00138192,0.002316697,0.00430009,0.0003074046],"domain_scores_gemma":[0.983876,0.00842645,0.001044274,0.003127841,0.003068285,0.00045714],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005444112,0.00001862629,0.0002419893,0.0008784672,0.00008788813,0.0001134892,0.0001167721,0.0007310832,0.0009045321,0.01010626,0.8720402,0.1147063],"study_design_scores_gemma":[0.000026408,0.0000206943,0.0003468494,0.0001968619,0.00001936884,0.0002465665,0.00002636833,0.0009984649,0.001011885,0.01462731,0.9824197,0.00005948183],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0009282466,0.008041623,0.5108285,0.004806944,0.005049334,0.001039733,0.1022954,0.2616128,0.1053975],"genre_scores_gemma":[0.00803859,0.006855277,0.573868,0.006254955,0.002215367,0.003726911,0.09064973,0.1867751,0.1216161],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2061614,"threshold_uncertainty_score":0.6896784,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2790635366","doi":"10.1177/0962280218759693","title":"MethodCompare: An R package to assess bias and precision in method comparison studies","year":2018,"lang":"en","type":"article","venue":"Statistical Methods in Medical Research","topic":"Data Analysis with R","field":"Computer Science","cited_by":31,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Alberta Children's Hospital; University of Calgary","funders":"","keywords":"R package; Statistics; Computer science; Econometrics; Mathematics","authors":[{"name":"Patrick Taffé","is_ca":false},{"name":"Mingkai Peng","is_ca":true},{"name":"Victoria Stagg","is_ca":false},{"name":"Tyler Williamson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6339406249091168,"gpt":0.7103905885533488,"spread":0.07644996364423196,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.07168199,0.004099906,0.004525821,0.008901943,0.001104482,0.00497057,0.005004589,0.002276741,0.06088536],"category_scores_gemma":[0.3469515,0.003015488,0.006059503,0.007755423,0.001884841,0.003452221,0.005167983,0.004832078,0.01522622],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001648371,"about_ca_system_score_gemma":0.008014223,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002640218,"about_ca_topic_score_gemma":0.003404071,"domain_scores_codex":[0.9300627,0.04603973,0.008148369,0.005783973,0.009282628,0.0006826764],"domain_scores_gemma":[0.5821936,0.3569021,0.02200575,0.020072,0.01785212,0.0009744965],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002257395,0.0003228027,0.01926455,0.02584628,0.01743014,0.0006360883,0.00167143,0.01169921,0.002759031,0.02943072,0.5771627,0.3115197],"study_design_scores_gemma":[0.002804174,0.000724624,0.02787487,0.007984879,0.008854012,0.001839435,0.0004717051,0.06756065,0.008312063,0.1359482,0.7366154,0.001009971],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.004613862,0.004458363,0.8523039,0.001621552,0.001731159,0.003483702,0.05572179,0.07139593,0.004669758],"genre_scores_gemma":[0.03808002,0.002360523,0.8530623,0.001683574,0.0006496678,0.0237412,0.01974406,0.05689598,0.003782606],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.928318,"threshold_uncertainty_score":0.3790951,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2075203254","doi":"10.1890/09-0460.1","title":"Multiscale codependence analysis: an integrated approach to analyze relationships across scales","year":2010,"lang":"en","type":"article","venue":"Ecology","topic":"Data Analysis with R","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; Université de Montréal","funders":"","keywords":"Ecology; Scale (ratio); Geography; Computer science; Biology; Cartography","authors":[{"name":"Guillaume Guénard","is_ca":true},{"name":"Pierre Legendre","is_ca":true},{"name":"Daniel Boisclair","is_ca":true},{"name":"Martin Bilodeau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02587042978143286,"gpt":0.299021606548138,"spread":0.2731511767667051,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00349623,0.0009530571,0.001416188,0.004021875,0.0005243991,0.001606364,0.001432389,0.0007301898,0.001230226],"category_scores_gemma":[0.01495893,0.0005288253,0.001480462,0.002791561,0.001374877,0.002124626,0.002286028,0.001368193,0.0002024259],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005377944,"about_ca_system_score_gemma":0.001038678,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002947574,"about_ca_topic_score_gemma":0.001991621,"domain_scores_codex":[0.9987508,0.0004733896,0.00008843294,0.0003257117,0.0002910368,0.00007060204],"domain_scores_gemma":[0.991924,0.004989711,0.001130423,0.001281232,0.0004646897,0.000209968],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00019948,0.0001989709,0.04512157,0.0006186927,0.001573981,0.0007648816,0.001300876,0.2400092,0.04652065,0.2223204,0.004643414,0.4367279],"study_design_scores_gemma":[0.0000193739,0.00007737354,0.01110678,0.00003218265,0.00009439065,0.0002351971,0.00008206098,0.8430349,0.00225745,0.1393685,0.00362218,0.00006958251],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008523203,0.0001254916,0.9906159,0.00007969376,0.00001279939,0.00002621231,0.00007904689,0.00026681,0.0002708073],"genre_scores_gemma":[0.2689351,0.0003351606,0.7292969,0.000132768,0.0001036492,0.0003541354,0.0002870068,0.0002096919,0.0003454851],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004021875,"threshold_uncertainty_score":0.01849008,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4386637689","doi":"10.1007/978-3-031-34583-8","title":"Quantitative Methods for the Social Sciences","year":2023,"lang":"en","type":"book","venue":"Springer texts in political science and international relations","topic":"Data Analysis with R","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa","funders":"","keywords":"Data science; Management science; Computer science; Social science; Sociology; Engineering","authors":[{"name":"Daniel Stockemer","is_ca":true},{"name":"Jean‐Nicolas Bordeleau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1088483196193201,"gpt":0.4700007971054457,"spread":0.3611524774861256,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03224983,0.003177613,0.003887764,0.006890728,0.001404808,0.007185357,0.003729856,0.00290651,0.04059373],"category_scores_gemma":[0.07229576,0.002129138,0.002002269,0.008621716,0.005379751,0.004806985,0.003263876,0.006768661,0.0150482],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002554898,"about_ca_system_score_gemma":0.005562941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002659195,"about_ca_topic_score_gemma":0.003064058,"domain_scores_codex":[0.9657506,0.02510343,0.001805648,0.002015829,0.005000454,0.0003241192],"domain_scores_gemma":[0.8910767,0.09180023,0.002082942,0.009728663,0.004699165,0.0006122554],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00006970325,0.0001321749,0.0007224174,0.003265417,0.0003640793,0.0001020121,0.0006819252,0.002666565,0.0009153294,0.5846654,0.1156985,0.2907165],"study_design_scores_gemma":[0.00007524747,0.00005826392,0.0009954881,0.001040308,0.0001149729,0.000232475,0.0003060294,0.006530611,0.0005810846,0.7174745,0.2725253,0.00006579245],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00034904,0.02001036,0.9509766,0.002995504,0.001476726,0.0005861632,0.002424104,0.002260221,0.01892135],"genre_scores_gemma":[0.01416247,0.01599274,0.9362079,0.001828652,0.001714782,0.005637495,0.002414906,0.002261293,0.01977981],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.04059373,"threshold_uncertainty_score":0.1705554,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3027464055","doi":"10.1186/s12859-020-3494-x","title":"MEPHAS: an interactive graphical user interface for medical and pharmaceutical statistical analysis with R and Shiny","year":2020,"lang":"en","type":"article","venue":"BMC Bioinformatics","topic":"Data Analysis with R","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Institute of Genetics; Chinese Government Scholarship; China Scholarship Council","keywords":"Computer science; Graphical user interface; Visualization; Interface (matter); Analytics; Workstation; User interface; Scatter plot; OS X; Data visualization; Statistical analysis; Data mining; Software; Programming language; Machine learning; Operating system; Statistics","authors":[{"name":"Yi Zhou","is_ca":false},{"name":"Siu-wai Leung","is_ca":false},{"name":"Shosuke Mizutani","is_ca":false},{"name":"Tatsuya Takagi","is_ca":false},{"name":"Yu‐Shi Tian","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.029111670027292,"gpt":0.333294396465094,"spread":0.304182726437802,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009095216,0.002890767,0.001830232,0.003096673,0.0005221852,0.002843689,0.003804646,0.001782245,0.2127432],"category_scores_gemma":[0.04745369,0.001700501,0.002367768,0.002100606,0.00100174,0.003504413,0.004002578,0.002877114,0.07501019],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006171471,"about_ca_system_score_gemma":0.002642378,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001396974,"about_ca_topic_score_gemma":0.001999985,"domain_scores_codex":[0.9956995,0.001837262,0.0005158637,0.0006831614,0.001021389,0.0002427821],"domain_scores_gemma":[0.9661539,0.02654508,0.001461902,0.002092916,0.003024623,0.0007216279],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00119709,0.00009099457,0.001901832,0.004634239,0.0005628115,0.0007190362,0.0007174488,0.004025514,0.007052992,0.01554098,0.7897412,0.1738159],"study_design_scores_gemma":[0.001434437,0.0003066569,0.006650079,0.001977061,0.0003913903,0.00164186,0.000214687,0.05340423,0.01598241,0.06020407,0.8572251,0.000568167],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.002847713,0.001257765,0.4966382,0.001457193,0.0007239074,0.0008252214,0.0493349,0.4378218,0.009093369],"genre_scores_gemma":[0.03552444,0.001552736,0.7444035,0.002816737,0.0005769723,0.0062394,0.03551928,0.1592328,0.01413423],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.2127432,"threshold_uncertainty_score":0.711697,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2900082925","doi":"10.1007/978-3-319-96978-7_2","title":"Use of Machine Learning (ML) for Predicting and Analyzing Ecological and ‘Presence Only’ Data: An Overview of Applications and a Good Outlook","year":2018,"lang":"en","type":"book-chapter","venue":"","topic":"Data Analysis with R","field":"Computer Science","cited_by":26,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université Laval; Mount Allison University","funders":"","keywords":"Machine learning; Computer science; Artificial intelligence; Boosting (machine learning); Artificial neural network; Field (mathematics); Data mining; Mathematics","authors":[{"name":"Falk Huettmann","is_ca":false},{"name":"Erica H. Craig","is_ca":false},{"name":"Keiko A. Herrick","is_ca":false},{"name":"Andrew P. Baltensperger","is_ca":false},{"name":"Grant Humphries","is_ca":false},{"name":"David J. Lieske","is_ca":true},{"name":"Katharine B. Miller","is_ca":false},{"name":"Timothy C. Mullet","is_ca":false},{"name":"Steffen Oppel","is_ca":false},{"name":"Cynthia Resendiz","is_ca":false},{"name":"Imme Rutzen","is_ca":false},{"name":"Moritz S. Schmid","is_ca":true},{"name":"Madan Krishna Suwal","is_ca":false},{"name":"Brian Young","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1424685242687778,"gpt":0.3336138633195477,"spread":0.1911453390507699,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005390222,0.001954994,0.001790721,0.003972064,0.0004804335,0.005197984,0.003177192,0.002688459,0.007449155],"category_scores_gemma":[0.01006044,0.001456297,0.002058223,0.005909864,0.002271029,0.007662186,0.002774561,0.005551444,0.008501088],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001271704,"about_ca_system_score_gemma":0.001044378,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001500611,"about_ca_topic_score_gemma":0.001636143,"domain_scores_codex":[0.997088,0.0009156099,0.0001902205,0.0006846551,0.001040877,0.00008070013],"domain_scores_gemma":[0.9896354,0.008286672,0.0003033078,0.0008034999,0.0008314563,0.0001396731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003502623,0.00006903856,0.001261001,0.001568585,0.0001239777,0.0001103616,0.0001550295,0.007253127,0.002146875,0.09179429,0.03498509,0.8604978],"study_design_scores_gemma":[0.00001449731,0.0001289118,0.002246319,0.001374405,0.00009427856,0.001154577,0.0001406507,0.06383847,0.00827009,0.4173866,0.5051271,0.0002241964],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.001446903,0.1778109,0.7896187,0.006790485,0.001254494,0.0000934772,0.0009636758,0.002320311,0.01970101],"genre_scores_gemma":[0.02077117,0.180164,0.770258,0.00309458,0.004425259,0.0002675527,0.001809405,0.001149324,0.01806067],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.007449155,"threshold_uncertainty_score":0.02850658,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W103961220","doi":"","title":"The R software: Fundamentals of programming and statistical analysis","year":2013,"lang":"en","type":"preprint","venue":"QUT ePrints (Queensland University of Technology)","topic":"Data Analysis with R","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal","funders":"","keywords":"Computer science; Fortran; Debugging; Documentation; Section (typography); Software; Programming language; Biostatistics; Software engineering","authors":[{"name":"Pierre Lafaye de Micheaux","is_ca":true},{"name":"Rémy Drouilhet","is_ca":false},{"name":"Benoît Liquet","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.008401666252845294,"gpt":0.2194462348443054,"spread":0.2110445685914601,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02201675,0.005875874,0.007116552,0.01056682,0.00146405,0.007229535,0.006394759,0.002752691,0.04594877],"category_scores_gemma":[0.1027656,0.004238338,0.004293694,0.01097973,0.005105768,0.004311346,0.003208018,0.01100117,0.05055501],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001383759,"about_ca_system_score_gemma":0.005549435,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002383061,"about_ca_topic_score_gemma":0.002251771,"domain_scores_codex":[0.9742883,0.01416799,0.004327317,0.003288693,0.003375174,0.0005524732],"domain_scores_gemma":[0.8806266,0.08253768,0.008543676,0.0212242,0.006051678,0.001016078],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0006992132,0.0003684426,0.003627441,0.0106392,0.001875549,0.0008765917,0.000893086,0.0219483,0.01022429,0.1205937,0.5590536,0.2692006],"study_design_scores_gemma":[0.000860644,0.0003710781,0.006487069,0.002808044,0.001350244,0.002635924,0.0002906059,0.05102576,0.01892089,0.340886,0.573713,0.0006507268],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001105562,0.002382919,0.8986587,0.001359564,0.0007778091,0.0008082988,0.01819113,0.07189272,0.004823179],"genre_scores_gemma":[0.009612506,0.001551999,0.94217,0.0006609414,0.0005069305,0.00379402,0.008043726,0.03030272,0.003357235],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.04594877,"threshold_uncertainty_score":0.1537139,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3168850057","doi":"10.32614/rj-2022-012","title":"RFpredInterval: An R Package for Prediction Intervals with Random Forests and Boosted Forests","year":2022,"lang":"en","type":"article","venue":"The R Journal","topic":"Data Analysis with R","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; HEC Montréal; Fondation HEC","keywords":"Random forest; R package; Computer science; Reliability (semiconductor); Predictive modelling; Prediction interval; Data mining; Point (geometry); Mean squared prediction error; Set (abstract data type); Statistics; Machine learning; Mathematics","authors":[{"name":"Cansu Alakuş","is_ca":false},{"name":"Denis Larocque","is_ca":false},{"name":"Aurélie Labbe","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01908452781822515,"gpt":0.2593394716969587,"spread":0.2402549438787336,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01128616,0.0033175,0.002653799,0.003753375,0.0009615284,0.002957477,0.004850954,0.001593925,0.03105386],"category_scores_gemma":[0.05454515,0.002337729,0.003960316,0.003521638,0.0008578869,0.002925364,0.002358241,0.005241902,0.02658451],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007310205,"about_ca_system_score_gemma":0.002700565,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004260196,"about_ca_topic_score_gemma":0.006113143,"domain_scores_codex":[0.9942292,0.002596909,0.0004667601,0.001309193,0.001171732,0.0002262364],"domain_scores_gemma":[0.9792401,0.01446416,0.001693366,0.002569739,0.001727168,0.0003054563],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0009683571,0.0001900076,0.01606967,0.004523851,0.003244539,0.0005550587,0.0005025879,0.0888794,0.004633408,0.02724501,0.5713404,0.2818478],"study_design_scores_gemma":[0.0007490458,0.0002022735,0.009452549,0.0009587547,0.001049992,0.0008887998,0.0001077346,0.5228721,0.01129195,0.1300711,0.3219071,0.000448567],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.002500282,0.001442602,0.843117,0.0004469645,0.0004538129,0.0002576522,0.03182436,0.1174531,0.002504172],"genre_scores_gemma":[0.03540552,0.001077768,0.8588853,0.0006719147,0.0004045438,0.002187586,0.04586811,0.05250311,0.002996123],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.03105386,"threshold_uncertainty_score":0.1038855,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4300185383","doi":"10.1017/cbo9781316451090","title":"A First Course in Statistical Programming with R","year":2016,"lang":"en","type":"book","venue":"Cambridge University Press eBooks","topic":"Data Analysis with R","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University; University of British Columbia, Okanagan Campus; University of British Columbia","funders":"","keywords":"Computer science; Advice (programming); Graphics; Markov chain; Code (set theory); Programming language; Software engineering; Computer graphics (images); Machine learning","authors":[{"name":"W. John Braun","is_ca":true},{"name":"Duncan J. Murdoch","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01250430285110974,"gpt":0.2057397241362296,"spread":0.1932354212851199,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00295844,0.001995207,0.002145165,0.002389835,0.0007733212,0.004045343,0.002429322,0.00185001,0.1300371],"category_scores_gemma":[0.01453483,0.001566191,0.001824858,0.00439602,0.001287984,0.003991474,0.002287138,0.006936762,0.191419],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000959649,"about_ca_system_score_gemma":0.002128257,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00104087,"about_ca_topic_score_gemma":0.001822662,"domain_scores_codex":[0.9968636,0.0007534636,0.0002305097,0.00049802,0.001573352,0.00008099891],"domain_scores_gemma":[0.9916304,0.005265951,0.0002827398,0.001151928,0.001403298,0.0002656809],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001116068,0.00002117227,0.00009114807,0.0003726397,0.00002544057,0.00005917304,0.0001153369,0.0005881191,0.0002465165,0.02276659,0.7850236,0.1906792],"study_design_scores_gemma":[0.000010027,0.00001249943,0.0001815716,0.0002284631,0.000007179488,0.0001858762,0.00002894399,0.0008402089,0.0001638608,0.03396504,0.9643592,0.00001716206],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0005176874,0.03608045,0.6070986,0.02099782,0.007763179,0.0003999329,0.01141222,0.0470001,0.26873],"genre_scores_gemma":[0.005475452,0.0352499,0.5524874,0.01402029,0.005540951,0.001826032,0.01436377,0.03180644,0.3392298],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1300371,"threshold_uncertainty_score":0.4350174,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W982522955","doi":"10.1007/978-1-4614-9020-3_8","title":"Programming in R","year":2013,"lang":"en","type":"book-chapter","venue":"Statisctics and computing/Statistics and computing","topic":"Data Analysis with R","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Programming language; Computer science; Programming paradigm; Order (exchange); Programming domain; C programming language; Object-oriented programming; Inductive programming; Artificial intelligence; Software","authors":[{"name":"Pierre Lafaye de Micheaux","is_ca":true},{"name":"Rémy Drouilhet","is_ca":false},{"name":"Benoît Liquet","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01385810052166029,"gpt":0.2451940253181226,"spread":0.2313359247964624,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003231557,0.001923056,0.001930588,0.001947266,0.0006704515,0.004374498,0.00196688,0.001176709,0.1091987],"category_scores_gemma":[0.01367703,0.001311984,0.001635925,0.002145698,0.001298473,0.002610096,0.001739858,0.003534074,0.1488043],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005753277,"about_ca_system_score_gemma":0.001591137,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009087429,"about_ca_topic_score_gemma":0.001245796,"domain_scores_codex":[0.9960073,0.001781191,0.0004106285,0.0006965392,0.0009580019,0.000146368],"domain_scores_gemma":[0.99324,0.004059293,0.0002922079,0.0014753,0.0008001822,0.0001330172],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004442428,0.00003159516,0.0001928545,0.0009443962,0.00008038124,0.0001037943,0.0001573805,0.004669054,0.0008948308,0.2769138,0.4414925,0.2744751],"study_design_scores_gemma":[0.0000389592,0.00002275124,0.0001472717,0.0003250589,0.00004809791,0.0003042394,0.00004883127,0.01324471,0.001865884,0.3073719,0.6765426,0.00003975641],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0003419749,0.001578828,0.9013045,0.001251184,0.0005551502,0.000128443,0.004244797,0.02973781,0.06085732],"genre_scores_gemma":[0.01068409,0.002180889,0.8733022,0.001568856,0.000628639,0.00104635,0.006640485,0.02941247,0.07453606],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1091987,"threshold_uncertainty_score":0.3653059,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3000522262","doi":"10.1002/ece3.5970","title":"A checklist for choosing between R packages in ecology and evolution","year":2020,"lang":"en","type":"article","venue":"Ecology and Evolution","topic":"Data Analysis with R","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta; York University","funders":"","keywords":"Workflow; Ecology; Computer science; Evolutionary ecology; Data science; Checklist; Process (computing); Management science; Biology; Engineering","authors":[{"name":"Christopher J. Lortie","is_ca":true},{"name":"Jenna Braun","is_ca":true},{"name":"Alessandro Filazzola","is_ca":true},{"name":"M. Florencia Miguel","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01662091244123715,"gpt":0.25156243227777,"spread":0.2349415198365328,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.09223571,0.003907083,0.002566847,0.01476872,0.004573687,0.005885204,0.007202784,0.004889517,0.06664996],"category_scores_gemma":[0.2720358,0.0035293,0.003274685,0.01315969,0.003947159,0.007224353,0.007602847,0.01183282,0.05308489],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002516459,"about_ca_system_score_gemma":0.01165406,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002631621,"about_ca_topic_score_gemma":0.00550581,"domain_scores_codex":[0.91431,0.04736647,0.02380974,0.00286624,0.01012596,0.00152159],"domain_scores_gemma":[0.6856188,0.2117014,0.0150515,0.02851247,0.05520811,0.003907722],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003806726,0.0001225927,0.001659361,0.004960834,0.0001124888,0.0006784197,0.00177061,0.001138726,0.00279374,0.04869613,0.7993362,0.1383503],"study_design_scores_gemma":[0.000159844,0.0001003316,0.001545228,0.004046429,0.00007631783,0.0007409025,0.0004122664,0.001207259,0.001554058,0.04205196,0.947868,0.0002374883],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002861566,0.005676627,0.80114,0.02306084,0.004576116,0.009403273,0.04794238,0.06799749,0.03734183],"genre_scores_gemma":[0.004210892,0.002211569,0.9271202,0.008400466,0.0007779638,0.01334232,0.01395831,0.02323723,0.006740955],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.09223571,"threshold_uncertainty_score":0.4877949,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3088716273","doi":"10.1111/2041-210x.13494","title":"Going further with model verification and deep learning","year":2020,"lang":"en","type":"article","venue":"Methods in Ecology and Evolution","topic":"Data Analysis with R","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Center for Northern Studies; Université de Moncton","funders":"New Brunswick Innovation Foundation; Polar Knowledge Canada","keywords":"Deep learning; Workflow; Computer science; Artificial intelligence; Machine learning; Data science; Database","authors":[{"name":"Sylvain Christin","is_ca":true},{"name":"Éric Hervet","is_ca":true},{"name":"Nicolas Lecomte","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02567183586669051,"gpt":0.3102287904532507,"spread":0.2845569545865602,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03929202,0.001763362,0.002063203,0.002698026,0.001027533,0.006697294,0.004200243,0.003468059,0.008243073],"category_scores_gemma":[0.1229243,0.001327826,0.00428646,0.001920387,0.007406312,0.01648413,0.006499421,0.01147092,0.002176646],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00357316,"about_ca_system_score_gemma":0.00671797,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004835792,"about_ca_topic_score_gemma":0.003739756,"domain_scores_codex":[0.9761878,0.013989,0.001812181,0.002731056,0.004691484,0.000588605],"domain_scores_gemma":[0.8496069,0.113358,0.004197216,0.02153735,0.01051245,0.0007880831],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001041036,0.0000801067,0.002250984,0.002818424,0.0006232731,0.0002624339,0.0006811332,0.04897209,0.001500055,0.7548057,0.02046072,0.167441],"study_design_scores_gemma":[0.0000291124,0.00004313767,0.0002848685,0.001501516,0.00009145842,0.000101357,0.0001311251,0.1130271,0.003751278,0.8277187,0.05324088,0.00007948334],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00233944,0.006324394,0.9686443,0.01648532,0.0006502102,0.00006685496,0.0003064928,0.002138266,0.003044748],"genre_scores_gemma":[0.1620868,0.01113376,0.8063617,0.0111308,0.001176074,0.0005155836,0.001401177,0.0023918,0.003802311],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.960708,"threshold_uncertainty_score":0.2077985,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W6943970825","doi":"10.17605/osf.io/p46mb","title":"ChrontouR","year":2020,"lang":"en","type":"article","venue":"OSF Preprints (OSF Preprints)","topic":"Data Analysis with R","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.01921679443882397,"gpt":0.2496680565044222,"spread":0.2304512620655983,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003716663,0.003510567,0.002742282,0.004504333,0.001484618,0.004012638,0.00326579,0.00127823,0.3970371],"category_scores_gemma":[0.01389646,0.003150701,0.002645257,0.003208363,0.0009000456,0.002831914,0.002901414,0.002978623,0.2223627],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001505785,"about_ca_system_score_gemma":0.003190447,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006494727,"about_ca_topic_score_gemma":0.00988755,"domain_scores_codex":[0.9984152,0.0002622665,0.0002536163,0.0005999291,0.0002736272,0.0001952804],"domain_scores_gemma":[0.9930071,0.003589946,0.0005326714,0.001343629,0.001092721,0.0004339727],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0008555445,0.0001008488,0.003287767,0.002231694,0.0002329323,0.0002413073,0.0006643118,0.001659141,0.004004642,0.006157792,0.9192272,0.06133695],"study_design_scores_gemma":[0.0009181474,0.00009296957,0.006583436,0.0004058494,0.000194767,0.000432172,0.0002048896,0.007285979,0.0123826,0.02240122,0.9488103,0.0002877558],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"software","genre_gemma":"software","genre_scores_codex":[0.001761941,0.0002653026,0.0747249,0.0001936977,0.0002470601,0.0005297678,0.329065,0.5840483,0.009164101],"genre_scores_gemma":[0.01173263,0.0004242915,0.1596118,0.0009275141,0.00020994,0.004480452,0.2934867,0.5065619,0.02256477],"genre_candidate":"software","genre_consensus":"software","teacher_disagreement_score":0.3970371,"threshold_uncertainty_score":0,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4406534133","doi":"10.1016/0967-0653(93)94868-y","title":"10.1016/0967-0653(93)94868-y","year":2000,"lang":"en","type":"article","venue":"Time to knit","topic":"Data Analysis with R","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"","keywords":"Multivariate statistics; Multivariate analysis; Econometrics; Environmental science; Statistics; Mathematics","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.006010319443260933,"gpt":0.1739484471530853,"spread":0.1679381277098244,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.001602333,0.00303636,0.002168705,0.002985083,0.002525595,0.003935916,0.003450887,0.004826494,0.9914002],"category_scores_gemma":[0.002383866,0.001281744,0.001609618,0.00279871,0.002749937,0.006417322,0.003852744,0.002940111,0.9934126],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001107859,"about_ca_system_score_gemma":0.001081629,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006318009,"about_ca_topic_score_gemma":0.004807316,"domain_scores_codex":[0.9991431,0.0000741771,0.0000787576,0.0003040585,0.0002008959,0.0001990143],"domain_scores_gemma":[0.9972993,0.0007098332,0.000189947,0.0003734666,0.000550042,0.0008773822],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004894657,0.0002505006,0.001459835,0.0005729517,0.00005016643,0.0003060058,0.0001195981,0.0005426833,0.002109051,0.004379788,0.4273132,0.5624068],"study_design_scores_gemma":[0.00009634872,0.0001707315,0.001561425,0.0004970685,0.0000250564,0.0005409402,0.0002222252,0.0003528196,0.0005355071,0.001047092,0.9949121,0.00003862966],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.0008343558,0.0005817598,0.001415512,0.0007432821,0.0004402138,0.0002085723,0.002110744,0.001972074,0.9916934],"genre_scores_gemma":[0.0009900384,0.0003331528,0.0007948785,0.0003829687,0.00008516597,0.00009657152,0.0008467645,0.0003090442,0.9961615],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.008599758,"threshold_uncertainty_score":0.01256245,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1486450196","doi":"10.18637/jss.v069.i04","title":"Parallel and Other Simulations in<i>R</i>Made Easy: An End-to-End Study","year":2016,"lang":"en","type":"article","venue":"Journal of Statistical Software","topic":"Data Analysis with R","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Eidgenössische Technische Hochschule Zürich","keywords":"Computer science; Graphics; Computation; Set (abstract data type); Table (database); Scale (ratio); Contrast (vision); Contingency table; Algorithm; Computational science; Data mining; Computer graphics (images); Artificial intelligence; Programming language; Machine learning","authors":[{"name":"Marius Hofert","is_ca":true},{"name":"Martin Mächler","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02613877045930482,"gpt":0.3101626526354363,"spread":0.2840238821761315,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01936333,0.002010673,0.00180999,0.001100664,0.001185595,0.004237211,0.003922253,0.002088184,0.04112123],"category_scores_gemma":[0.1033371,0.001406739,0.003138833,0.001938623,0.001454644,0.005026298,0.004670346,0.00511322,0.02257074],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001198721,"about_ca_system_score_gemma":0.002245857,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002667231,"about_ca_topic_score_gemma":0.003048741,"domain_scores_codex":[0.9847489,0.01021779,0.0008219755,0.001518807,0.002293111,0.0003994353],"domain_scores_gemma":[0.9234363,0.05172091,0.001780143,0.01603673,0.006014254,0.0010117],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001381766,0.0006133857,0.00599524,0.002195496,0.0007405416,0.00102295,0.001420893,0.1826917,0.006981722,0.249892,0.2587219,0.2883425],"study_design_scores_gemma":[0.0005006089,0.0004905591,0.001932579,0.0006955309,0.0001948745,0.000514462,0.0002993953,0.4262099,0.01277388,0.2492163,0.3069381,0.0002338124],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004686392,0.0005697152,0.9582276,0.002665828,0.0006685382,0.0004530544,0.002877964,0.01622268,0.01362823],"genre_scores_gemma":[0.03760629,0.001007114,0.9289778,0.001995945,0.0003229606,0.002083768,0.005223176,0.01602118,0.006761802],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.04112123,"threshold_uncertainty_score":0.1375642,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4200001778","doi":"10.3233/sji-210875","title":"A quality framework for statistical algorithms","year":2021,"lang":"en","type":"article","venue":"Statistical Journal of the IAOS","topic":"Data Analysis with R","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Acadia University; Statistics Canada","funders":"","keywords":"Toolbox; Computer science; Quality (philosophy); Machine learning; Statistical learning; Transparency (behavior); Artificial intelligence; Data science; Process management; Operations research; Algorithm; Mathematics; Engineering; Computer security","authors":[{"name":"Wesley Yung","is_ca":true},{"name":"Siu‐Ming Tam","is_ca":false},{"name":"Bart Buelens","is_ca":false},{"name":"Hugh Chipman","is_ca":true},{"name":"Florian Dumpert","is_ca":false},{"name":"Gabriele Ascari","is_ca":false},{"name":"Fabiana Rocci","is_ca":false},{"name":"Joep Burger","is_ca":false},{"name":"Inkyung Choi","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04160504749905197,"gpt":0.3772035243872404,"spread":0.3355984768881884,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1154484,0.002327498,0.002594488,0.007805967,0.002451161,0.01420815,0.005906576,0.004948202,0.009045642],"category_scores_gemma":[0.2252442,0.001665543,0.004690837,0.006320588,0.01135131,0.01495962,0.008882251,0.01121632,0.004650369],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007642553,"about_ca_system_score_gemma":0.01401751,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007165193,"about_ca_topic_score_gemma":0.003785591,"domain_scores_codex":[0.8887354,0.06160505,0.009834683,0.006717533,0.03113985,0.001967537],"domain_scores_gemma":[0.7818567,0.1260532,0.008221927,0.03082076,0.04983865,0.003208808],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00003183963,0.00003072641,0.0004486792,0.00026873,0.00004741725,0.00004690976,0.0002175587,0.008721805,0.000204018,0.9490777,0.006449106,0.0344555],"study_design_scores_gemma":[0.00004740442,0.00007151406,0.0002294022,0.0003473934,0.00003378662,0.00009756001,0.00008064048,0.0570135,0.000544028,0.8968909,0.04459867,0.00004523979],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0003880448,0.0009709455,0.9905465,0.003508994,0.0002429055,0.0001087837,0.0001292689,0.0004314164,0.003673146],"genre_scores_gemma":[0.03355635,0.001657338,0.9570851,0.001729096,0.001164633,0.0006634961,0.0006067645,0.0007807675,0.002756384],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1154484,"threshold_uncertainty_score":0.6105568,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3117276733","doi":"","title":"NbClust Package. An examination of indices for determining the number of clusters","year":2012,"lang":"fr","type":"preprint","venue":"HAL (Le Centre pour la Communication Scientifique Directe)","topic":"Data Analysis with R","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université Laval","funders":"","keywords":"R package; Statistics; Computer science; Mathematics","authors":[{"name":"Malika Charrad","is_ca":true},{"name":"Nadia Ghazzali","is_ca":true},{"name":"Véronique Boiteau","is_ca":true},{"name":"Azam Niknafs","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03115078739807493,"gpt":0.2749493531850409,"spread":0.243798565786966,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03453443,0.004949078,0.006904849,0.00990178,0.005933887,0.007664388,0.01247013,0.003159675,0.0543362],"category_scores_gemma":[0.1228321,0.00300533,0.005738879,0.01526837,0.001950253,0.00524837,0.005273207,0.008499778,0.0490648],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001896617,"about_ca_system_score_gemma":0.007204189,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01317566,"about_ca_topic_score_gemma":0.02145736,"domain_scores_codex":[0.970771,0.01542647,0.002771317,0.005159515,0.005175556,0.0006961431],"domain_scores_gemma":[0.8762991,0.08961755,0.003256988,0.01675398,0.01162373,0.002448738],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001257266,0.0002669678,0.02125401,0.004714441,0.003770943,0.0005376466,0.001752834,0.006107847,0.003681745,0.01224311,0.8445961,0.09981709],"study_design_scores_gemma":[0.001414014,0.0004233027,0.02661091,0.001647037,0.002943691,0.002477353,0.002552903,0.1283462,0.01341512,0.07086403,0.7483726,0.0009328882],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02936728,0.006812804,0.3900747,0.004273296,0.002336653,0.001701653,0.3435769,0.2092851,0.01257165],"genre_scores_gemma":[0.07275615,0.002243204,0.6098701,0.001196857,0.000368124,0.004231601,0.1996639,0.09982354,0.009846531],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0543362,"threshold_uncertainty_score":0.1826377,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4406887161","doi":"10.21105/joss.07492","title":"ssdtools v2: An R package to fit Species Sensitivity Distributions","year":2025,"lang":"en","type":"article","venue":"The Journal of Open Source Software","topic":"Data Analysis with R","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Ministry of Environment","keywords":"R package; Sensitivity (control systems); Mathematics; Statistics; Environmental science; Statistical physics; Physics; Engineering","authors":[{"name":"Joe Thorley","is_ca":false},{"name":"Rebecca Fisher","is_ca":false},{"name":"David R. Fox","is_ca":false},{"name":"Carl J. Schwarz","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03521525920627432,"gpt":0.3171296502199887,"spread":0.2819143910137144,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00430133,0.003625843,0.002700312,0.003131933,0.0008676226,0.002810523,0.003417971,0.00116926,0.1139796],"category_scores_gemma":[0.02703266,0.001674335,0.003089359,0.002217989,0.0008040851,0.002710326,0.003342746,0.003335916,0.06215924],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008116331,"about_ca_system_score_gemma":0.002841247,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005449861,"about_ca_topic_score_gemma":0.007966414,"domain_scores_codex":[0.9976224,0.0007756996,0.000244237,0.0007174336,0.0004492214,0.0001910136],"domain_scores_gemma":[0.9916441,0.006188434,0.0005458231,0.0007121353,0.0007573938,0.0001521542],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0005568307,0.0001200217,0.01291696,0.005451402,0.00203593,0.0004654341,0.001040196,0.016403,0.00424913,0.01611293,0.8496305,0.09101768],"study_design_scores_gemma":[0.0004089464,0.0002109807,0.01239182,0.0007770666,0.0008210861,0.0007538008,0.0002981164,0.07777058,0.007949729,0.06742193,0.8307309,0.0004650878],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"software","genre_scores_codex":[0.01138751,0.001083378,0.2932618,0.000729274,0.0006493009,0.0005534446,0.347995,0.3351066,0.009233575],"genre_scores_gemma":[0.05876636,0.0008372868,0.4693499,0.001289026,0.0002212583,0.004869685,0.1609042,0.2901066,0.01365565],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.1139796,"threshold_uncertainty_score":0.3812997,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4285465633","doi":"10.1017/9781108993456","title":"A First Course in Statistical Programming with R","year":2021,"lang":"en","type":"book","venue":"Cambridge University Press eBooks","topic":"Data Analysis with R","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University; University of British Columbia","funders":"","keywords":"Computer science; Code (set theory); Course (navigation); Programming language; C programming language; Software engineering; Data science; Software; Engineering; Set (abstract data type)","authors":[{"name":"W. John Braun","is_ca":true},{"name":"Duncan J. Murdoch","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01345541887272999,"gpt":0.2065228774397496,"spread":0.1930674585670196,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003668451,0.002045343,0.00235793,0.002371276,0.0008582073,0.003717024,0.002640014,0.001839979,0.1730692],"category_scores_gemma":[0.01498226,0.001611346,0.001930759,0.004386023,0.001169527,0.004034633,0.002519806,0.006784673,0.248499],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009629109,"about_ca_system_score_gemma":0.002118301,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009335281,"about_ca_topic_score_gemma":0.001700263,"domain_scores_codex":[0.996577,0.0008628466,0.0002551818,0.0005694764,0.001631301,0.0001041958],"domain_scores_gemma":[0.9911174,0.005461016,0.0002620866,0.001202316,0.001624257,0.0003330677],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001136882,0.00002822454,0.0001086902,0.0003513457,0.00002262371,0.00006652429,0.000115958,0.0004120533,0.0002770159,0.01948095,0.809961,0.1691642],"study_design_scores_gemma":[0.00000925558,0.00001407426,0.0001959288,0.0001828073,0.000006509568,0.0001850851,0.00002760309,0.0005813732,0.0001657313,0.02455506,0.9740615,0.00001508669],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0007039102,0.03100869,0.5466582,0.02630029,0.007622768,0.0004870812,0.01484522,0.05161093,0.3207629],"genre_scores_gemma":[0.005562562,0.02943381,0.5343249,0.01456969,0.004874028,0.00170439,0.016529,0.03114278,0.3618588],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1730692,"threshold_uncertainty_score":0.578974,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2801007199","doi":"10.18637/jss.v084.c01","title":"<b>stampr</b>: Spatial-Temporal Analysis of Moving Polygons in <i>R</i>","year":2018,"lang":"en","type":"article","venue":"Journal of Statistical Software","topic":"Data Analysis with R","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Wilfrid Laurier University","funders":"","keywords":"Polygon (computer graphics); Computer science; R package; Core (optical fiber); Computer graphics (images); Algorithm; Computational science","authors":[{"name":"Jed Long","is_ca":false},{"name":"Colin Robertson","is_ca":true},{"name":"Trisalyn Nelson","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01174774280374646,"gpt":0.2764526692167041,"spread":0.2647049264129577,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00624609,0.002757444,0.002279941,0.003679453,0.001072478,0.003759369,0.00272913,0.001158858,0.1079382],"category_scores_gemma":[0.04874041,0.001745273,0.003326665,0.004949291,0.001195842,0.003005777,0.003300861,0.003147692,0.0632631],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006591267,"about_ca_system_score_gemma":0.002459786,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005119526,"about_ca_topic_score_gemma":0.005781554,"domain_scores_codex":[0.9962558,0.001429369,0.0003997133,0.0008324531,0.0008770858,0.0002055755],"domain_scores_gemma":[0.9750743,0.01566088,0.002292873,0.004544391,0.001937419,0.0004901089],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004408507,0.00007409437,0.01044578,0.003187735,0.0009861421,0.0004495289,0.001030351,0.01358191,0.007193303,0.02771741,0.8094471,0.1254458],"study_design_scores_gemma":[0.0004326596,0.0002400361,0.01992405,0.0007123846,0.0005852898,0.001457979,0.0003066006,0.07874462,0.0208522,0.06068456,0.815453,0.0006066855],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.003523309,0.0004153573,0.7135835,0.000632355,0.0006783992,0.0002680838,0.09225928,0.1832113,0.005428483],"genre_scores_gemma":[0.04936689,0.0007951885,0.7033441,0.000819931,0.0003649596,0.004054158,0.07674163,0.1565101,0.008002975],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.1079382,"threshold_uncertainty_score":0.3610892,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3010960300","doi":"10.1002/eap.2123","title":"Ecological prediction at macroscales using big data: Does sampling design matter?","year":2020,"lang":"en","type":"article","venue":"Ecological Applications","topic":"Data Analysis with R","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Cégep Marie-Victorin","funders":"National Institute of Food and Agriculture; Directorate for Biological Sciences","keywords":"Sampling (signal processing); Ecology; Sampling design; Big data; Environmental science; Geography; Computer science; Biology; Data mining; Population; Sociology","authors":[{"name":"Patricia A. Soranno","is_ca":false},{"name":"Kendra Spence Cheruvelil","is_ca":false},{"name":"Boyang Liu","is_ca":false},{"name":"Qi Wang","is_ca":false},{"name":"Pang‐Ning Tan","is_ca":false},{"name":"Jiayu Zhou","is_ca":false},{"name":"Katelyn King","is_ca":false},{"name":"Ian M. McCullough","is_ca":false},{"name":"Jemma Stachelek","is_ca":false},{"name":"Meridith L. Bartley","is_ca":false},{"name":"Christopher T. Filstrup","is_ca":false},{"name":"Ephraim M. Hanks","is_ca":false},{"name":"Jean‐François Lapierre","is_ca":true},{"name":"Noah R. Lottig","is_ca":false},{"name":"Erin M. Schliep","is_ca":false},{"name":"Tyler Wagner","is_ca":false},{"name":"Katherine E. Webster","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.204191952903336,"gpt":0.319121722104141,"spread":0.1149297692008051,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.09489474,0.0007631374,0.0008154033,0.0006570707,0.0006682522,0.001674635,0.001874517,0.001126218,0.0007019928],"category_scores_gemma":[0.1910985,0.0006355591,0.001100072,0.001350181,0.001746065,0.002680168,0.001257962,0.001003696,0.0002136153],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008762543,"about_ca_system_score_gemma":0.001359989,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007337532,"about_ca_topic_score_gemma":0.008465504,"domain_scores_codex":[0.9507272,0.04339894,0.001501021,0.002344665,0.001665877,0.0003623272],"domain_scores_gemma":[0.7516398,0.2070763,0.01137752,0.02323024,0.005678443,0.0009977465],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001278502,0.0003581239,0.7477549,0.0003941324,0.002400031,0.0001612228,0.0009947364,0.173299,0.002998666,0.004865757,0.00235525,0.06313976],"study_design_scores_gemma":[0.0003872569,0.0009725199,0.1811019,0.0002789887,0.0007310526,0.0001836863,0.000863361,0.7798408,0.008043684,0.02373348,0.0037183,0.000144899],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6668044,0.0006472202,0.326699,0.001765761,0.000153106,0.0007605358,0.001160455,0.0003837252,0.001625749],"genre_scores_gemma":[0.9576917,0.0001110083,0.04074991,0.0002851609,0.00003683694,0.0004119846,0.0005689298,0.00003933662,0.0001051539],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9051052,"threshold_uncertainty_score":0.5018574,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1907586038","doi":"10.1201/9780429031069","title":"Graphics for Statistics and Data Analysis with R","year":2010,"lang":"en","type":"book","venue":"","topic":"Data Analysis with R","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Northern British Columbia","funders":"","keywords":"Graphics; Statistical graphics; Bar chart; Computer science; Scatter plot; Plot (graphics); Chart; Pie chart; Nonparametric statistics; Visualization; Data mining; Computer graphics; Parametric statistics; Statistics; Computer graphics (images); Mathematics; Machine learning","authors":[{"name":"Kevin J. Keen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02855733066667223,"gpt":0.2883750817602925,"spread":0.2598177510936203,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.008148741,0.003532897,0.003407166,0.007755537,0.001202341,0.006339816,0.004413269,0.002657061,0.3190989],"category_scores_gemma":[0.06180669,0.001818105,0.002833659,0.01226013,0.00161061,0.005822835,0.003405633,0.008134675,0.2898211],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001723326,"about_ca_system_score_gemma":0.004425915,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002063938,"about_ca_topic_score_gemma":0.002365041,"domain_scores_codex":[0.9818459,0.006633374,0.002013882,0.002287396,0.006860119,0.0003593542],"domain_scores_gemma":[0.9589406,0.02403196,0.002235001,0.006620069,0.007545474,0.0006269339],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004429398,0.00002951026,0.0002019624,0.001576481,0.00009310666,0.00007824122,0.0001946671,0.0008863152,0.0004085956,0.01662106,0.8779942,0.1018714],"study_design_scores_gemma":[0.00003106922,0.00002425597,0.0004484486,0.0003742211,0.0000289492,0.0001245319,0.0000491738,0.001256446,0.000398858,0.02675096,0.9704674,0.00004578754],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.000915934,0.01655108,0.5250596,0.0122366,0.009792013,0.001881016,0.09693249,0.1468193,0.189812],"genre_scores_gemma":[0.009186022,0.01837725,0.6346633,0.009151727,0.004823233,0.007159091,0.08237474,0.07425271,0.160012],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3190989,"threshold_uncertainty_score":0.9712228,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4280635309","doi":"10.1371/journal.pone.0268426","title":"Addressing the need for interactive, efficient, and reproducible data processing in ecology with the datacleanr R package","year":2022,"lang":"en","type":"article","venue":"PLoS ONE","topic":"Data Analysis with R","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université TÉLUQ; Université du Québec à Montréal","funders":"Helmholtz-Gemeinschaft; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Computer science; Workflow; Data science; Data quality; Data processing; Outlier; Data type; Data mining; Quality (philosophy); Database; Service (business); Artificial intelligence; Programming language","authors":[{"name":"Alexander Hurley","is_ca":false},{"name":"Richard L. Peters","is_ca":false},{"name":"Christoforos Pappas","is_ca":true},{"name":"David N. Steger","is_ca":false},{"name":"Ingo Heinrich","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1763984246597676,"gpt":0.3123639623910735,"spread":0.1359655377313059,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05281568,0.003436408,0.003681189,0.004425373,0.002311932,0.009618538,0.007069865,0.002576547,0.01648036],"category_scores_gemma":[0.1367923,0.00343038,0.005407551,0.005763336,0.004262418,0.006368819,0.01220844,0.009922777,0.01816745],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001218494,"about_ca_system_score_gemma":0.01035304,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004285141,"about_ca_topic_score_gemma":0.006895307,"domain_scores_codex":[0.9642435,0.02169995,0.002715108,0.004246215,0.006233248,0.0008619201],"domain_scores_gemma":[0.8659194,0.09437052,0.007321288,0.02137463,0.008988289,0.002025926],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001005844,0.0003311649,0.0230681,0.009005305,0.003537775,0.001413228,0.006632223,0.02735789,0.02795359,0.07738071,0.5792705,0.2430436],"study_design_scores_gemma":[0.0007342305,0.0003241725,0.01290686,0.001836469,0.0009238549,0.001460121,0.0007212068,0.08533382,0.03470796,0.2017375,0.6582787,0.001035067],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.003569574,0.00051097,0.8787217,0.002906265,0.0003565035,0.000345553,0.01150782,0.09996682,0.002114898],"genre_scores_gemma":[0.01822723,0.0004937322,0.9153088,0.002003551,0.000202022,0.002646611,0.01108496,0.04882465,0.001208398],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.05281568,"threshold_uncertainty_score":0.2793193,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2118093838","doi":"10.25336/p6531z","title":"Multistate Analysis of Life Histories with R","year":2015,"lang":"en","type":"article","venue":"Canadian Studies in Population","topic":"Data Analysis with R","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"","keywords":"Sociology; Demography","authors":[{"name":"David A. Swanson","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.08285411426822298,"gpt":0.3129009743640734,"spread":0.2300468600958504,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04175155,0.00226917,0.003453433,0.005248321,0.002347375,0.004264637,0.004067919,0.001276363,0.04245984],"category_scores_gemma":[0.2003601,0.001573837,0.005426483,0.008633018,0.003194235,0.002265615,0.002979602,0.00370958,0.01072109],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003413916,"about_ca_system_score_gemma":0.01239282,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.09581869,"about_ca_topic_score_gemma":0.1051189,"domain_scores_codex":[0.9668373,0.01863823,0.002414793,0.007570031,0.003116059,0.001423549],"domain_scores_gemma":[0.8511037,0.1082054,0.004681083,0.02811039,0.006827168,0.001072269],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.002625648,0.0004123456,0.1097507,0.003590218,0.01127341,0.001009689,0.005381292,0.07534843,0.005050598,0.1414367,0.2000336,0.4440873],"study_design_scores_gemma":[0.0007527253,0.0007027415,0.09661525,0.001151445,0.003562917,0.001265647,0.00139878,0.4443294,0.008386278,0.204559,0.2366215,0.0006543871],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0285174,0.0005565694,0.9063125,0.0008543627,0.0003932301,0.001073934,0.03477439,0.02375187,0.003765805],"genre_scores_gemma":[0.3005782,0.000343284,0.6594012,0.0003908832,0.0001127041,0.005521601,0.01655844,0.006922663,0.01017101],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.09581869,"threshold_uncertainty_score":0.2208059,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}