{"id":"W2076514995","doi":"10.1214/12-aoas584","title":"Robust VIF regression with application to variable selection in large data sets","year":2013,"lang":"en","type":"article","venue":"The Annals of Applied Statistics","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":37,"is_retracted":false,"has_abstract":true,"ca_institutions":"HEC Montréal","funders":"","keywords":"Outlier; Robustness (evolution); Feature selection; Computer science; Estimator; Regression; Robust regression; Lasso (programming language); Covariate; Regression analysis; Variance (accounting); Variance inflation factor; Econometrics; Artificial intelligence; Machine learning; Statistics; Mathematics; Multicollinearity; Economics","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":[0.02301944,0.002101356,0.002973266,0.003638728,0.001070633,0.001790844,0.002973313,0.002440035,0.002830846],"category_scores_gemma":[0.08605167,0.001040685,0.002712406,0.004892746,0.001671371,0.001698873,0.002935951,0.004473626,0.001225168],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001076976,"about_ca_system_score_gemma":0.002759867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00813426,"about_ca_topic_score_gemma":0.006631867,"domain_scores_codex":[0.9869788,0.009645884,0.0004703641,0.001086714,0.001517836,0.0003003939],"domain_scores_gemma":[0.95035,0.0417905,0.001866852,0.003241047,0.002469741,0.0002819875],"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":[0.0001871393,0.0001229979,0.005550471,0.0005405953,0.0008088162,0.000552516,0.0003157363,0.5704059,0.001935376,0.1141637,0.008422008,0.2969947],"study_design_scores_gemma":[0.00002971961,0.00004770424,0.0007322481,0.00005094403,0.00003099548,0.00009466768,0.00002372389,0.93725,0.0006570885,0.05667783,0.004369946,0.00003503399],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001570804,0.0005687797,0.9968672,0.0002159997,0.00004556744,0.00004410466,0.00007655106,0.0003852376,0.0002256541],"genre_scores_gemma":[0.06811411,0.001331807,0.9270338,0.0002040263,0.0002884686,0.0005853163,0.0006604818,0.0004564355,0.001325622],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02301944,"threshold_uncertainty_score":0.1217399,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2047160822476765,"score_gpt":0.4366883251623959,"score_spread":0.2319722429147194,"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."}}