Kernel-Based Anomaly Detection Using Generalized Hyperbolic Processes
Pauline Bourigault 1,2 Danilo P. Mandic 2 · 2025
Read Kernel-Based Anomaly Detection Using Generalized Hyperbolic Processes by Pauline Bourigault 1,2 Danilo P. Mandic 2 (2025) 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-2501-15265.txt.
Description
Pauline Bourigault and Danilo Mandic present a new way to spot odd data points, in a research paper from 2025. Their method builds on kernels, tools that judge how alike two data points are. It uses a flexible distribution that can bend to fit skewed data, heavy tails, and sharp peaks.
The team folds that distribution into two standard detectors: one based on density estimates and one based on a one-class support vector machine. They prove the new kernel is sound and stable for learning use. Tests on made-up and real data showed better detection when the data were heavy-tailed, lopsided, or uneven. Code for the method is shared in a public repository.
Catalog record
| CREATOR | Pauline Bourigault 1,2 Danilo P. Mandic 2 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-15265