Anomaly Detection via Mean Shift Density Enhancement
Pritam Kar, Rahul Bordoloi, Olaf Wolkenhauer, Saptarshi Bej · 2026
Read Anomaly Detection via Mean Shift Density Enhancement by Pritam Kar, Rahul Bordoloi, Olaf Wolkenhauer, Saptarshi Bej (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-03293.txt.
Description
Anomalies reveal themselves by how far they drift toward density, offers Pritam Kar, with 3 co-authors, in a fully unsupervised finding system. Earlier detectors excel only under given structural assumptions and degrade under noise, so results varies sharply across anomaly types. Their Mean Shift Density Enhancement instead watches each sample's geometric response as a density-driven process iteratively pulls points toward nearby modes on the data manifold over many steps.
Normal samples, well supported by local density, stay comparatively stable. Anomalous samples undergo large cumulative displacements. A weighted mean-shift procedure with changing, sample-given weights from a fuzzy neighborhood graph operationalizes the idea. Review spans dozens of real tabular datasets, several anomaly output mechanisms, and multiple noise levels. Against many established earlier methods, displacement-based scoring stayed consistently strong and balanced.
Catalog record
| CREATOR | Pritam Kar, Rahul Bordoloi, Olaf Wolkenhauer, Saptarshi Bej |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-03293