On Robust Variance Filtering and Change of Variance Detection

Abstract

This paper studies the variance filtering and change of variance (CoV) detection under multiple change points in time series signal. In real world scenarios, CoV detection can be challenging since the time series signal may contain not only outliers but also abrupt trend changes. To deal with these challenges, we propose a robust CoV detection algorithm based on robust statistics and sparse regularizations. Specifically, we adopt Huber loss to suppress outliers both in trend removal and variance filtering, utilize sparse regu-larizations to capture trend and variance changes, and obtain accurate change points locations by using breakpoint detection for centered cumulative sum of the estimated variance. We compare our proposed robust CoV algorithm with other state-of-the-art CoV detection algorithms on both synthetic and public datasets. The experiments demonstrate that our algorithm outperforms existing methods.