Stochastic Voronoi Ensembles for Anomaly Detection

TL;DR

This paper introduces SVEAD, a novel anomaly detection method using stochastic Voronoi ensembles that effectively identifies local anomalies with linear time complexity, outperforming existing approaches.

Contribution

The paper proposes SVEAD, a new ensemble-based anomaly detection technique utilizing stochastic Voronoi diagrams for improved local anomaly detection.

Findings

SVEAD achieves linear time complexity and constant space complexity.

SVEAD outperforms 12 state-of-the-art methods on 45 datasets.

The method effectively detects local anomalies in datasets with varying densities.

Abstract

Anomaly detection aims to identify data instances that deviate significantly from majority of data, which has been widely used in fraud detection, network security, and industrial quality control. Existing methods struggle with datasets exhibiting varying local densities: distance-based methods miss local anomalies, while density-based approaches require careful parameter selection and incur quadratic time complexity. We observe that local anomalies, though indistinguishable under global analysis, become conspicuous when the data space is decomposed into restricted regions and each region is examined independently. Leveraging this geometric insight, we propose SVEAD (Stochastic Voronoi Ensembles Anomaly Detector), which constructs ensemble random Voronoi diagrams and scores points by normalized cell-relative distances weighted by local scale. The proposed method achieves linear time complexity and constant space complexity. Experiments on 45 datasets demonstrate that SVEAD outperforms 12 state-of-the-art approaches.