{"id":"W4318908907","doi":"10.1145/3582491","title":"Data Science---A Systematic Treatment","year":2023,"lang":"en","type":"article","venue":"Communications of the ACM","topic":"Artificial Intelligence in Healthcare","field":"Health Professions","cited_by":14,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Data science","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","open_science","insufficient_payload"],"consensus_categories":["open_science"],"category_scores_codex":[0.002300452,0.00007092118,0.0001982809,0.0001251734,0.002410849,0.000007248235,0.05183045,0.00005176948,0.00002744547],"category_scores_gemma":[0.04616629,0.00004355298,0.00003857437,0.001421142,0.0006673433,0.0001757538,0.04923808,0.0002216203,0.001115855],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002312761,"about_ca_system_score_gemma":0.0007731073,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00115329,"about_ca_topic_score_gemma":0.002259933,"domain_scores_codex":[0.9979081,0.0007635275,0.0006182373,0.0001782764,0.0002535295,0.0002783884],"domain_scores_gemma":[0.9101163,0.002887129,0.0003263151,0.08630762,0.0003039223,0.00005872746],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00004271252,0.0008319309,0.3727494,0.02523129,0.0002783161,0.000001494947,0.09097213,0.0002149078,0.007327677,0.2753568,0.2215388,0.005454504],"study_design_scores_gemma":[0.0007663909,0.0003119319,0.08015874,0.0344687,0.000598561,0.000007235988,0.2220758,0.1293077,0.004382919,0.4439894,0.08299136,0.0009413016],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5406702,0.002977104,0.00001381266,0.4153527,0.002791577,0.008868481,0.0006922898,0.0006607644,0.027973],"genre_scores_gemma":[0.993938,0.0008054203,0.003647365,0.0001802108,0.00003799539,0.0003126875,0.000043071,0.00001339811,0.001021832],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4532678,"threshold_uncertainty_score":0.9996619,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.7465776776167445,"score_gpt":0.6254157346368587,"score_spread":0.1211619429798858,"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."}}