Unsupervised Feature Construction for Anomaly Detection in Time Series -- An Evaluation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913658492157952 |
|---|---|
| author | Hamon, Marine Lemaire, Vincent Nair-Benrekia, Nour Eddine Yassine Berlemont, Samuel Cumin, Julien |
| author_facet | Hamon, Marine Lemaire, Vincent Nair-Benrekia, Nour Eddine Yassine Berlemont, Samuel Cumin, Julien |
| contents | To detect anomalies with precision and without prior knowledge in time series, is it better to build a detector from the initial temporal representation, or to compute a new (tabular) representation using an existing automatic variable construction library? In this article, we address this question by conducting an in-depth experimental study for two popular detectors (Isolation Forest and Local Outlier Factor). The obtained results, for 5 different datasets, show that the new representation, computed using the tsfresh library, allows Isolation Forest to significantly improve its performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_07999 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Unsupervised Feature Construction for Anomaly Detection in Time Series -- An Evaluation Hamon, Marine Lemaire, Vincent Nair-Benrekia, Nour Eddine Yassine Berlemont, Samuel Cumin, Julien Machine Learning To detect anomalies with precision and without prior knowledge in time series, is it better to build a detector from the initial temporal representation, or to compute a new (tabular) representation using an existing automatic variable construction library? In this article, we address this question by conducting an in-depth experimental study for two popular detectors (Isolation Forest and Local Outlier Factor). The obtained results, for 5 different datasets, show that the new representation, computed using the tsfresh library, allows Isolation Forest to significantly improve its performance. |
| title | Unsupervised Feature Construction for Anomaly Detection in Time Series -- An Evaluation |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2501.07999 |