A Robust Unsupervised Method For Outlier Set Detection

doi:10.1016/j.knosys.2025.114274

A Robust Unsupervised Method For Outlier Set Detection

2025-08-15

0

PRE

AI

DOI：10.1016/j.knosys.2025.114274

Original

Original request for help

Save

Abstract

En 中文

• New approach to label clustered sets as outliers sets instead of individual points. • Our two-step method combines clustering and distance-based outlier set detection. • Traditional outlier detection methods are not designed to tackle this problem. • Our method (OSTI) reliably identifies outlier sets across 8,000 synthetic datasets.

Keywords:

outlier detection

clustering

outlier sets

two-step method

synthetic datasets

A Robust Unsupervised Method For Outlier Set Detection

Abstract

Journal

Organization