{"id":"W4412125311","doi":"10.1080/10920277.2025.2517149","title":"Recurrent Neural Networks for Multivariate Loss Reserving and Risk Capital Analysis","year":2025,"lang":"en","type":"article","venue":"North American Actuarial Journal","topic":"Probability and Risk Models","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada; Shared Hierarchical Academic Research Computing Network; McMaster University","keywords":"Multivariate statistics; Multivariate analysis; Artificial neural network; Computer science; Econometrics; Artificial intelligence; Economics; Machine learning","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.003276753,0.0002299195,0.0007224068,0.000653383,0.0008789377,0.000756449,0.0008315752,0.00006653431,0.00004180535],"category_scores_gemma":[0.004150167,0.0001579879,0.0004897646,0.002417669,0.0004211199,0.0004313801,0.0002641962,0.0006968739,0.000002484845],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008491894,"about_ca_system_score_gemma":0.0001611286,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00095722,"about_ca_topic_score_gemma":0.005098792,"domain_scores_codex":[0.9964693,0.0006762546,0.001004557,0.0006049203,0.0007175256,0.0005274754],"domain_scores_gemma":[0.9952919,0.002438464,0.0009187516,0.0005021953,0.0005315305,0.0003171155],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006112656,0.00005947704,0.4944669,0.000001797167,0.0004561957,0.000006825083,0.0005572203,0.07985763,0.000001817788,0.00009298729,0.0004616618,0.4234262],"study_design_scores_gemma":[0.0008906577,0.0002448656,0.4434949,0.00000682069,0.0005873744,0.00001147334,0.0003796445,0.5463023,0.00000198476,0.006877246,0.001004405,0.0001983041],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6861857,0.0001575433,0.3119406,0.0005974462,0.0008300669,0.0001828522,0.00005101573,0.00001621874,0.00003847978],"genre_scores_gemma":[0.9948025,0.0002831376,0.004125772,0.0001778143,0.0005160766,0.00001036113,0.000006970507,0.000008679232,0.00006865994],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4664447,"threshold_uncertainty_score":0.7294455,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04420844994975798,"score_gpt":0.3664997982290382,"score_spread":0.3222913482792802,"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."}}