{"id":"W4402126847","doi":"10.1016/j.jempfin.2024.101538","title":"Pooling and winsorizing machine learning forecasts to predict stock returns with high-dimensional data","year":2024,"lang":"en","type":"article","venue":"Journal of Empirical Finance","topic":"Stock Market Forecasting Methods","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"University of Denver","keywords":"Predictability; Machine learning; Pooling; Artificial intelligence; Computer science; Gradient boosting; Elastic net regularization; Robustness (evolution); Lasso (programming language); Random forest; Econometrics; Artificial neural network; Boosting (machine learning); Stock market; Economics; Statistics; Feature selection; Mathematics","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.008289701,0.0002232216,0.0005402308,0.0004798509,0.0002012018,0.0003818106,0.0009969827,0.0000904041,0.00005108726],"category_scores_gemma":[0.01432652,0.0001362669,0.00008302891,0.00132432,0.00009479676,0.000801452,0.0006556365,0.0009571935,0.00001463751],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006955482,"about_ca_system_score_gemma":0.0002740576,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001455981,"about_ca_topic_score_gemma":0.00001633087,"domain_scores_codex":[0.9958168,0.0004268799,0.0009966152,0.0006562948,0.001744377,0.0003590366],"domain_scores_gemma":[0.9933619,0.005021438,0.0004258337,0.0005376456,0.0004025298,0.0002505943],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001689095,0.00009369005,0.1457295,0.00004744818,0.0001512373,0.001411303,0.001306551,0.01949228,0.0007660685,0.0002945084,0.04447782,0.7845405],"study_design_scores_gemma":[0.001187423,0.002989748,0.2302997,0.002013983,0.0001400827,0.004651358,0.0001001706,0.5035249,0.0002364783,0.009509237,0.2447355,0.0006113667],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9278641,0.002552869,0.0638879,0.004300083,0.0009352588,0.0001395126,0.00003693439,0.0000343644,0.0002490263],"genre_scores_gemma":[0.8371229,0.00003814152,0.1611736,0.0002685635,0.0005046817,0.000001520492,0.000002455172,0.00002920037,0.0008589919],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7839291,"threshold_uncertainty_score":0.9939762,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1893237070993366,"score_gpt":0.4380208558987659,"score_spread":0.2486971487994293,"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."}}