Drop-in density-aware partitioning for tree-based anomaly detection

Unsupervised isolation tree-based anomaly detection methods are widely used for their simplicity, scalability, and strong empirical performance. However, the standard partition rule, whether drawing thresholds uniformly from feature ranges or sampling random normal vectors for oblique hyperplanes, i...

詳細記述

書誌詳細

主要な著者:	Sijin Yeom, Jae-Hun Jung
フォーマット:	Artigo
言語:	Inglês
出版事項:	IOP Publishing 2026-01-01
シリーズ:	Machine Learning: Science and Technology
主題:	unsupervised anomaly detection isolation forest density-aware partitioning density-aware split density-aware direction
オンライン･アクセス:	https://doi.org/10.1088/2632-2153/ae5390
タグ:	タグ追加 タグなし, このレコードへの初めてのタグを付けませんか!