{"id":"W4409672360","doi":"10.1002/cjs.70008","title":"Predicting rare events using training data from stratified sampling designs, with application to human‐caused wildfire prediction","year":2025,"lang":"en","type":"article","venue":"Canadian Journal of Statistics","topic":"Fire effects on ecosystems","field":"Environmental Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Actua; Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Statistical Sciences Institute; Ministry of Natural Resources of the People's Republic of China; Institute for Catastrophic Loss Reduction","keywords":"Stratified sampling; Rare events; Training (meteorology); Sampling (signal processing); Training set; Data sampling; Statistics; Computer science; Environmental science; Artificial intelligence; Meteorology; Mathematics; Geography","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":true,"about_ca":true,"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":[0.010699,0.0005402726,0.0006423729,0.0006151344,0.000512985,0.0006088472,0.001142117,0.0007636925,0.0008443483],"category_scores_gemma":[0.0273547,0.0004813461,0.0008085946,0.0006314495,0.000689789,0.0004100847,0.00057303,0.0007718167,0.0001125113],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00160825,"about_ca_system_score_gemma":0.00195794,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.1152074,"about_ca_topic_score_gemma":0.1231952,"domain_scores_codex":[0.998161,0.001340539,0.00005963304,0.0002152594,0.0001256401,0.00009786763],"domain_scores_gemma":[0.9792651,0.01553647,0.001740263,0.0018064,0.001302476,0.0003494087],"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":[0.0007082685,0.0001450403,0.09654748,0.0000729758,0.0002135868,0.0001382544,0.0002795248,0.8566322,0.001167503,0.002619178,0.000519232,0.04095676],"study_design_scores_gemma":[0.00004210138,0.000121866,0.01335311,0.00001444961,0.0000302369,0.00001620407,0.0000491593,0.9833001,0.000587347,0.002242252,0.0002293828,0.00001366679],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.7674364,0.0001830049,0.2301661,0.0001349088,0.00002994589,0.0003173389,0.0006860667,0.0004293036,0.0006169818],"genre_scores_gemma":[0.9451996,0.00005002711,0.05348571,0.0000392311,0.000008327966,0.0001994087,0.0005914323,0.00001347489,0.0004128837],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.1152074,"threshold_uncertainty_score":0.2290736,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06891623857096456,"score_gpt":0.2808156290936847,"score_spread":0.2118993905227201,"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."}}