{"id":"W4394890400","doi":"10.1101/2024.04.16.24305900","title":"Early prediction of <i>Mycobacterium tuberculosis</i> transmission clusters using supervised learning models","year":2024,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Tuberculosis Research and Epidemiology","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute of Infection and Immunity; BC Centre for Disease Control; University of British Columbia; Simon Fraser University","funders":"","keywords":"Demographics; Transmission (telecommunications); Mycobacterium tuberculosis; Tuberculosis; Disease; Machine learning; Random forest; Artificial intelligence; Predictive modelling; Medicine; Disease transmission; Computer science; Demography; Internal medicine; Pathology; Virology","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004243434,0.0007114676,0.0007527156,0.0009759036,0.0002917548,0.001122007,0.0008137565,0.0008022953,0.001711046],"category_scores_gemma":[0.01020946,0.0002997335,0.0009742949,0.0004221385,0.0003951653,0.000859572,0.0006747915,0.001375915,0.0008946335],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006674292,"about_ca_system_score_gemma":0.0008580302,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005375318,"about_ca_topic_score_gemma":0.005099003,"domain_scores_codex":[0.9988599,0.0006553884,0.00005008682,0.0002094981,0.00009440739,0.0001306703],"domain_scores_gemma":[0.9923646,0.005215135,0.0009636832,0.0004875198,0.0006397652,0.0003292905],"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":[0.001083514,0.0008720299,0.3600706,0.0003236871,0.000584066,0.0003768948,0.0003396481,0.5044719,0.005019791,0.003408779,0.01320122,0.1102479],"study_design_scores_gemma":[0.00002115141,0.00008436257,0.01102105,0.00004194587,0.00003119038,0.00006404881,0.0000545101,0.9828583,0.001210935,0.004051194,0.0005476121,0.00001372405],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7740577,0.001214431,0.2145481,0.002234191,0.000166261,0.0001623628,0.00319321,0.001535146,0.002888437],"genre_scores_gemma":[0.9770499,0.0001401541,0.01961755,0.0001906138,0.00007647976,0.00003628347,0.002144668,0.00004309192,0.0007011784],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005375318,"threshold_uncertainty_score":0.02244169,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06185526649826256,"score_gpt":0.3105612365306369,"score_spread":0.2487059700323743,"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."}}