{"id":"W4385681410","doi":"10.48550/arxiv.2308.03369","title":"Variable importance for causal forests: breaking down the heterogeneity of treatment effects","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Safran Electronics (Canada)","funders":"","keywords":"Consistency (knowledge bases); Confounding; Econometrics; Variable (mathematics); Random forest; Computer science; Measure (data warehouse); Regression; Causal inference; Mathematics; Statistics; Machine learning; Artificial intelligence; Data mining","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":[0.02162428,0.001469356,0.00219546,0.002789539,0.001266986,0.001896143,0.002825554,0.00229657,0.003259686],"category_scores_gemma":[0.07164633,0.0009467044,0.002201296,0.002090172,0.002171793,0.003337215,0.002858821,0.004743304,0.0005619973],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001348506,"about_ca_system_score_gemma":0.00233048,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002466765,"about_ca_topic_score_gemma":0.003543759,"domain_scores_codex":[0.9905322,0.006273719,0.0003250996,0.00135832,0.001118437,0.0003922144],"domain_scores_gemma":[0.9482774,0.04516462,0.00166066,0.002850018,0.001547117,0.0005002036],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004528854,0.0002218332,0.008986607,0.0009906597,0.000527254,0.0003298509,0.0005931769,0.3097697,0.003514601,0.2131116,0.006857854,0.4546439],"study_design_scores_gemma":[0.0001134488,0.00009713094,0.001023478,0.0001668364,0.0001235089,0.0001662573,0.00004389284,0.7066616,0.002092664,0.2855966,0.003883786,0.00003080388],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003059088,0.0003495699,0.995799,0.000160004,0.00003835926,0.00007168682,0.00005290282,0.000183843,0.0002855122],"genre_scores_gemma":[0.2560284,0.0007769333,0.7395964,0.0005958713,0.0004253477,0.0005467425,0.0005066085,0.0003384002,0.001185326],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02162428,"threshold_uncertainty_score":0.1143615,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2595606165189933,"score_gpt":0.3130373605594232,"score_spread":0.05347674404042985,"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."}}