Skip to Main content Skip to Navigation
Preprints, Working Papers, ...

How to Evaluate the Quality of Unsupervised Anomaly Detection Algorithms?

Abstract : When sufficient labeled data are available, classical criteria based on Receiver Operating Characteristic (ROC) or Precision-Recall (PR) curves can be used to compare the performance of un-supervised anomaly detection algorithms. However , in many situations, few or no data are labeled. This calls for alternative criteria one can compute on non-labeled data. In this paper, two criteria that do not require labels are empirically shown to discriminate accurately (w.r.t. ROC or PR based criteria) between algorithms. These criteria are based on existing Excess-Mass (EM) and Mass-Volume (MV) curves, which generally cannot be well estimated in large dimension. A methodology based on feature sub-sampling and aggregating is also described and tested, extending the use of these criteria to high-dimensional datasets and solving major drawbacks inherent to standard EM and MV curves.
Complete list of metadatas

Cited literature [23 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01341809
Contributor : Nicolas Goix <>
Submitted on : Monday, July 4, 2016 - 7:58:08 PM
Last modification on : Thursday, March 5, 2020 - 3:57:53 PM
Document(s) archivé(s) le : Wednesday, October 5, 2016 - 2:44:58 PM

Files

icml_workshop_with_suppl.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01341809, version 1
  • ARXIV : 1607.01152

Citation

Nicolas Goix. How to Evaluate the Quality of Unsupervised Anomaly Detection Algorithms?. 2016. ⟨hal-01341809⟩

Share

Metrics

Record views

202

Files downloads

576