About this document
Two-Step Anomaly Detection for PU Learning by uil.count1985 is a document available to read on EtoBox.
This document presents a novel two-step method for positive and unlabeled (PU) classification in imbalanced data sets, utilizing anomaly detection to identify hidden positives among unlabeled data. The method introduces a new semi-supervised anomaly detector called Nearest-Neighbor Isolation Forest (NNIF), which outperforms existing PU learning methods in various experimental settings. The paper addresses the challenges of class imbalance and label noise, providing empirical comparisons and practical guidan
- Author
- uil.count1985
- Language
- EN