{"id":"W3142990567","doi":"10.1016/j.watres.2021.117089","title":"Exploration of E. coli contamination drivers in private drinking water wells: An application of machine learning to a large, multivariable, geo-spatio-temporal dataset","year":2021,"lang":"en","type":"article","venue":"Water Research","topic":"Groundwater flow and contamination studies","field":"Environmental Science","cited_by":35,"is_retracted":false,"has_abstract":false,"ca_institutions":"McMaster University; Queen's University; Global Institute for Water Security; University of Saskatchewan; Kingston Health Sciences Centre; Public Health Ontario","funders":"","keywords":"Contamination; Aquifer; Hydrogeology; Bivariate analysis; Environmental science; Water well; Groundwater; Bedrock; Univariate; Hydrology (agriculture); Environmental health; Multivariate statistics; Engineering; Statistics; Geology; Ecology; Mathematics; Medicine","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":[0.0008902636,0.0004006728,0.0004964527,0.001749471,0.000345737,0.0005948457,0.0008530469,0.000906888,0.0009121482],"category_scores_gemma":[0.002601829,0.00022794,0.00118837,0.001882647,0.0002717822,0.0006144011,0.0009616393,0.0006643178,0.0003703469],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005854328,"about_ca_system_score_gemma":0.0009560821,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.04610392,"about_ca_topic_score_gemma":0.06215151,"domain_scores_codex":[0.9995784,0.0001175206,0.00003892471,0.0001337956,0.00006151068,0.0000697878],"domain_scores_gemma":[0.9983911,0.0009416657,0.0002064384,0.0001753132,0.0001758371,0.0001096618],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0009181259,0.00156429,0.8168353,0.0005133375,0.0008442695,0.0008987626,0.0004968819,0.1047463,0.00487718,0.001216272,0.01936226,0.04772694],"study_design_scores_gemma":[0.0001436476,0.0003438361,0.536541,0.00005825674,0.0001832452,0.0004523626,0.001752211,0.444422,0.001971286,0.002859745,0.01118893,0.00008347535],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9544665,0.0002903531,0.004153308,0.0006736044,0.0000312402,0.00003533773,0.0395691,0.0003272733,0.0004533377],"genre_scores_gemma":[0.9197643,0.0001807085,0.009449684,0.00009525674,0.00003212177,0.00006767642,0.06976165,0.00003105363,0.0006176535],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04610392,"threshold_uncertainty_score":0.09167117,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04510934528432027,"score_gpt":0.320124583976891,"score_spread":0.2750152386925707,"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."}}