Detecting outliers from pairwise proximities: Proximity isolation forests
Pattern Recognition(2023)
摘要
•We propose a Random Forest-based outlier detector that works with pairwise proximities.•We propose nine training strategies, both random and optimized ones.•Experiments confirm the suitability of the approach.•We compare favorably to other proximity-based techniques, including recent ones.
更多查看译文
关键词
Random forest,Outlier detection,Isolation,Pairwise distances
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要