Aarhus University Seal / Aarhus Universitets segl

Active Learning of SVDD Hyperparameter Values

Research output: Contribution to book/anthology/report/proceedingArticle in proceedingsResearchpeer-review

  • Holger Trittenbach, KIT
  • ,
  • Klemens Böhm, KIT
  • ,
  • Ira Assent

Support Vector Data Description (SVDD) is a popular one-class classifier, and well-suited for outlier detection. However, the effectiveness of SVDD depends on selecting good hyperparameter values - a difficult problem that has received significant attention in the literature. Since SVDD is an unsupervised classifier, tuning of hyperparameter values is difficult. This has motivated several methods to estimate hyperparameter values based on data characteristics. But existing methods are purely heuristic, and the conditions under which they work well are largely unclear. This has created a situation where instead of selecting hyperparameter values, one has to choose among several, equally plausible heuristics.In this article, we make some strides towards a principled approach to estimate SVDD hyperparameter values. We propose LAMA (Local Active Min-Max Alignment), the first method to select SVDD hyperparameter values by active learning. The core idea is based on kernel alignment, which we adapt to active learning with small sample sizes. LAMA provides estimates for both of the SVDD hyperparameters. These estimates are evidence-based, i.e., rely on actual class labels, and come with a quality score. This eliminates the need for manual validation, an issue with current heuristics. LAMA outperforms state-of-theart competitors in extensive experiments on real-world data. In several cases, LAMA even yields results close to the empirical upper bound.

Original languageEnglish
Title of host publicationProceedings - 2020 IEEE 7th International Conference on Data Science and Advanced Analytics, DSAA 2020
EditorsGeoff Webb, Zhongfei Zhang, Vincent S. Tseng, Graham Williams, Michalis Vlachos, Longbing Cao
Number of pages9
PublisherIEEE
Publication yearOct 2020
Pages109-117
Article number9260103
ISBN (Electronic)9781728182063
DOIs
Publication statusPublished - Oct 2020
Event7th International Conference on Data Science and Advanced Analytics - Sydney, Australia
Duration: 6 Oct 20209 Oct 2020
Conference number: 7

Conference

Conference7th International Conference on Data Science and Advanced Analytics
Nummer7
LandAustralia
BySydney
Periode06/10/202009/10/2020

    Research areas

  • Active Learning, Outlier Detection, Support Vector Data Description

See relations at Aarhus University Citationformats

ID: 208007109