Out-of-Distribution Detection in Time-Series Domain: A Novel Seasonal Ratio Scoring Approach

TL;DR

This paper introduces a novel Seasonal Ratio Scoring (SRS) method for detecting out-of-distribution data in time-series classification, addressing unique challenges and outperforming baseline methods on real-world benchmarks.

Contribution

The paper proposes the SRS approach, a new method specifically designed for time-series OOD detection that leverages seasonal decomposition and deep generative models.

Findings

SRS outperforms baseline methods on real-world benchmarks.

The seasonal ratio score effectively detects OOD time-series data.

Decomposition improves the accuracy of likelihood estimation.

Abstract

Safe deployment of time-series classifiers for real-world applications relies on the ability to detect the data which is not generated from the same distribution as training data. This task is referred to as out-of-distribution (OOD) detection. We consider the novel problem of OOD detection for the time-series domain. We discuss the unique challenges posed by time-series data and explain why prior methods from the image domain will perform poorly. Motivated by these challenges, this paper proposes a novel {\em Seasonal Ratio Scoring (SRS)} approach. SRS consists of three key algorithmic steps. First, each input is decomposed into class-wise semantic component and remainder. Second, this decomposition is employed to estimate the class-wise conditional likelihoods of the input and remainder using deep generative models. The seasonal ratio score is computed from these estimates. Third, a threshold interval is identified from the in-distribution data to detect OOD examples. Experiments on diverse real-world benchmarks demonstrate that the SRS method is well-suited for time-series OOD detection when compared to baseline methods. Open-source code for SRS method is provided at https://github.com/tahabelkhouja/SRS