{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:7CC2PYEBVY6MCSXINX4QKWZGEF","short_pith_number":"pith:7CC2PYEB","schema_version":"1.0","canonical_sha256":"f885a7e081ae3cc14ae86df9055b26215a67c20c259e26286130c53f2e803ec4","source":{"kind":"arxiv","id":"1906.10845","version":2},"attestation_state":"computed","paper":{"title":"A Debiased MDI Feature Importance Measure for Random Forests","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Bin Yu, Karl Kumbier, Sumanta Basu, Xiao Li, Yu Wang","submitted_at":"2019-06-26T04:51:52Z","abstract_excerpt":"Tree ensembles such as Random Forests have achieved impressive empirical success across a wide variety of applications. To understand how these models make predictions, people routinely turn to feature importance measures calculated from tree ensembles. It has long been known that Mean Decrease Impurity (MDI), one of the most widely used measures of feature importance, incorrectly assigns high importance to noisy features, leading to systematic bias in feature selection. In this paper, we address the feature selection bias of MDI from both theoretical and methodological perspectives. Based on "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"1906.10845","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2019-06-26T04:51:52Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"c02922e83a55e126ae73e677cc719a1db63129013e80da9f8be5e57fffb44ed7","abstract_canon_sha256":"ca8e7b5062234502b05a3ee9adef400a33e4e4ad3989e298ad3678313dd42c9f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:15:06.783056Z","signature_b64":"9JGNRLGAbVbrBCmb2VEa70x+ZC2fqZItJiVIsgMQdxeZfFGhlIry8QYskurSwgEAffRIYaOg3A6gGgemO9E2Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f885a7e081ae3cc14ae86df9055b26215a67c20c259e26286130c53f2e803ec4","last_reissued_at":"2026-07-05T00:15:06.782600Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:15:06.782600Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Debiased MDI Feature Importance Measure for Random Forests","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Bin Yu, Karl Kumbier, Sumanta Basu, Xiao Li, Yu Wang","submitted_at":"2019-06-26T04:51:52Z","abstract_excerpt":"Tree ensembles such as Random Forests have achieved impressive empirical success across a wide variety of applications. To understand how these models make predictions, people routinely turn to feature importance measures calculated from tree ensembles. It has long been known that Mean Decrease Impurity (MDI), one of the most widely used measures of feature importance, incorrectly assigns high importance to noisy features, leading to systematic bias in feature selection. In this paper, we address the feature selection bias of MDI from both theoretical and methodological perspectives. Based on "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1906.10845","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1906.10845/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"1906.10845","created_at":"2026-07-05T00:15:06.782651+00:00"},{"alias_kind":"arxiv_version","alias_value":"1906.10845v2","created_at":"2026-07-05T00:15:06.782651+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1906.10845","created_at":"2026-07-05T00:15:06.782651+00:00"},{"alias_kind":"pith_short_12","alias_value":"7CC2PYEBVY6M","created_at":"2026-07-05T00:15:06.782651+00:00"},{"alias_kind":"pith_short_16","alias_value":"7CC2PYEBVY6MCSXI","created_at":"2026-07-05T00:15:06.782651+00:00"},{"alias_kind":"pith_short_8","alias_value":"7CC2PYEB","created_at":"2026-07-05T00:15:06.782651+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.07153","citing_title":"Feature Importance Depends on Properties of the Data: Towards Choosing the Correct Explanations for Your Data and Decision Trees based Models","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF","json":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF.json","graph_json":"https://pith.science/api/pith-number/7CC2PYEBVY6MCSXINX4QKWZGEF/graph.json","events_json":"https://pith.science/api/pith-number/7CC2PYEBVY6MCSXINX4QKWZGEF/events.json","paper":"https://pith.science/paper/7CC2PYEB"},"agent_actions":{"view_html":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF","download_json":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF.json","view_paper":"https://pith.science/paper/7CC2PYEB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1906.10845&json=true","fetch_graph":"https://pith.science/api/pith-number/7CC2PYEBVY6MCSXINX4QKWZGEF/graph.json","fetch_events":"https://pith.science/api/pith-number/7CC2PYEBVY6MCSXINX4QKWZGEF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF/action/storage_attestation","attest_author":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF/action/author_attestation","sign_citation":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF/action/citation_signature","submit_replication":"https://pith.science/pith/7CC2PYEBVY6MCSXINX4QKWZGEF/action/replication_record"}},"created_at":"2026-07-05T00:15:06.782651+00:00","updated_at":"2026-07-05T00:15:06.782651+00:00"}