Anomaly Attribution of Multivariate Time Series using Counterfactual Reasoning

TL;DR

This paper introduces a counterfactual reasoning-based attribution method for multivariate time series anomalies, enhancing understanding by identifying which variables contribute to anomalies, demonstrated on climate event data.

Contribution

The novel attribution scheme uses counterfactual reasoning to explain anomalies in multivariate time series, improving interpretability over existing detection methods.

Findings

Effective in attributing anomalies in climate data

Accurately identifies key variables causing anomalies

Validated on heatwaves and hurricanes

Abstract

There are numerous methods for detecting anomalies in time series, but that is only the first step to understanding them. We strive to exceed this by explaining those anomalies. Thus we develop a novel attribution scheme for multivariate time series relying on counterfactual reasoning. We aim to answer the counterfactual question of would the anomalous event have occurred if the subset of the involved variables had been more similarly distributed to the data outside of the anomalous interval. Specifically, we detect anomalous intervals using the Maximally Divergent Interval (MDI) algorithm, replace a subset of variables with their in-distribution values within the detected interval and observe if the interval has become less anomalous, by re-scoring it with MDI. We evaluate our method on multivariate temporal and spatio-temporal data and confirm the accuracy of our anomaly attribution of multiple well-understood extreme climate events such as heatwaves and hurricanes.