{"id":"W2125610074","doi":"10.2196/medinform.3251","title":"Clinical Data Miner: An Electronic Case Report Form System With Integrated Data Preprocessing and Machine-Learning Libraries Supporting Clinical Diagnostic Model Research","year":2014,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"AI in cancer detection","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Computer science; Workflow; Preprocessor; Data pre-processing; Machine learning; Software; Data mining; Interface (matter); Artificial intelligence; Software engineering; Database; Programming language; Operating system","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.02464139,0.0002555378,0.000552386,0.0001656287,0.0005253271,0.000772382,0.003259376,0.0003803934,0.000005630611],"category_scores_gemma":[0.01359384,0.0001869927,0.000032117,0.0006432692,0.0006207379,0.005901824,0.004310307,0.002800495,0.000008272355],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009420656,"about_ca_system_score_gemma":0.002126203,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001185957,"about_ca_topic_score_gemma":0.000319748,"domain_scores_codex":[0.9935644,0.0006816858,0.002515926,0.0008533746,0.001558608,0.0008259969],"domain_scores_gemma":[0.9903243,0.003869643,0.0009607892,0.003765556,0.0003313095,0.0007484416],"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":[0.0001754676,0.0003541177,0.03243195,0.00171497,0.0001632128,0.00445957,0.005529919,0.0005530676,6.116014e-7,0.005402398,0.00455814,0.9446566],"study_design_scores_gemma":[0.0006218665,0.000637096,0.00008889884,0.0004128354,0.00002589282,0.01450219,0.001341404,0.9780935,0.000003031509,0.0002631342,0.003774841,0.0002352818],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.140309,0.00008484556,0.8576835,0.0003458582,0.0001896029,0.0004601227,0.0000152694,0.0004584604,0.0004533376],"genre_scores_gemma":[0.9130184,0.0000828136,0.08577586,0.0002240979,0.000346583,0.00004129104,0.0004415076,0.000031006,0.0000384097],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9775404,"threshold_uncertainty_score":0.9995001,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1539092547912468,"score_gpt":0.4471745632613304,"score_spread":0.2932653084700836,"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."}}