{"id":"W2792845961","doi":"10.1016/j.spl.2018.02.018","title":"Statisticians can do better in the big data era","year":2018,"lang":"en","type":"article","venue":"Statistics & Probability Letters","topic":"Data Analysis with R","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Big data; Compromise; Data science; Computer science; Mathematics; Data mining; Political science; Law","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.08541648,0.003062734,0.005222937,0.006480381,0.003313371,0.02014246,0.00440855,0.01289584,0.01945612],"category_scores_gemma":[0.2885739,0.001993169,0.002716947,0.005368182,0.020528,0.05758153,0.01042341,0.03251738,0.01853017],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00233725,"about_ca_system_score_gemma":0.006718492,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002453007,"about_ca_topic_score_gemma":0.001567067,"domain_scores_codex":[0.9567703,0.02607049,0.002206197,0.004229441,0.009380314,0.001343271],"domain_scores_gemma":[0.6283251,0.2679217,0.01080116,0.05613182,0.02264209,0.01417811],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004971267,0.0003089608,0.004282143,0.001275233,0.0007553539,0.0003684836,0.001798411,0.004340976,0.001096458,0.3966766,0.3920444,0.196556],"study_design_scores_gemma":[0.00007813548,0.00005433391,0.0005690223,0.000243078,0.00004812277,0.00008977048,0.0003812921,0.002427127,0.0003626481,0.875809,0.1198236,0.0001138351],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"commentary","genre_gemma":"commentary","genre_scores_codex":[0.002128059,0.03027107,0.1529247,0.7842549,0.01294173,0.00006056738,0.0007420446,0.002640425,0.01403645],"genre_scores_gemma":[0.1254079,0.05545611,0.249663,0.4587679,0.08098572,0.0004192991,0.001743001,0.006898734,0.02065832],"genre_candidate":"commentary","genre_consensus":"commentary","teacher_disagreement_score":0.9145835,"threshold_uncertainty_score":0.4517309,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05480528962070819,"score_gpt":0.2823701365293838,"score_spread":0.2275648469086756,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}