{"id":"W2942190862","doi":"10.1080/10920277.2020.1745656","title":"Boosting Insights in Insurance Tariff Plans with Tree-Based Machine Learning Methods","year":2020,"lang":"en","type":"preprint","venue":"North American Actuarial Journal","topic":"Financial Risk and Volatility Modeling","field":"Economics, Econometrics and Finance","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"Actua; Université Laval","funders":"KU Leuven; Vlaamse regering; Vlaams Supercomputer Centrum; Fonds Wetenschappelijk Onderzoek; McGill University","keywords":"Decision tree; Computer science; Boosting (machine learning); Machine learning; Gradient boosting; Random forest; Tariff; Artificial intelligence; Guard (computer science); Ensemble learning; Analytics; Focus (optics); Data mining; Economics","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":["metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.0009619655,0.0005664352,0.00166506,0.0006588955,0.0003185048,0.0003111247,0.0006455136,0.0002083192,0.00004557322],"category_scores_gemma":[0.0009599897,0.0005732981,0.0002997132,0.000761381,0.0001872538,0.0002163307,0.0002149415,0.004714029,0.0000237761],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003441091,"about_ca_system_score_gemma":0.0004866644,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006438635,"about_ca_topic_score_gemma":0.00727798,"domain_scores_codex":[0.9964124,0.0002717115,0.001604618,0.0009217249,0.0001593731,0.0006301626],"domain_scores_gemma":[0.996487,0.0003051973,0.002426214,0.0003717232,0.00009315527,0.0003167583],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0006842178,0.00008691874,0.8326674,0.00005263873,0.00008867571,0.0001028442,0.001676925,0.09648687,0.000008463333,0.0001672199,0.00001292908,0.06796491],"study_design_scores_gemma":[0.002293313,0.0007055273,0.5927153,0.0002356024,0.00003632786,0.00001979764,0.0001097567,0.3897412,0.00002625359,0.003619316,0.009327902,0.001169635],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6880342,0.0009869841,0.308209,0.0004901705,0.0007027306,0.0002793948,0.0001898452,0.00006700441,0.00104062],"genre_scores_gemma":[0.9336253,0.0004506568,0.06449279,0.0003727539,0.0008338807,0.00001982183,0.0001052469,0.00008627117,0.00001320896],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2932544,"threshold_uncertainty_score":0.9996719,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04665748895399771,"score_gpt":0.2703459494564732,"score_spread":0.2236884605024755,"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."}}