Evaluating Simplification Algorithms for Interpretability of Time Series Classification

TL;DR

This paper introduces metrics to evaluate simplified time series data for interpretability of classifiers, assesses four algorithms across datasets, and confirms the metrics' practical utility through human evaluation.

Contribution

It proposes new metrics for assessing simplifications in time series interpretability and provides an evaluation framework for their effectiveness.

Findings

Simplifications with fewer segments tend to be more loyal to original classifications.

Four algorithms were systematically evaluated across diverse datasets.

Human evaluation supports the effectiveness of the proposed metrics.

Abstract

In this work, we introduce metrics to evaluate the use of simplified time series in the context of interpretability of a TSC -- a Time Series Classifier. Such simplifications are important because time series data, in contrast to text and image data, are not intuitively under- standable to humans. These metrics are related to the complexity of the simplifications -- how many segments they contain -- and to their loyalty -- how likely they are to maintain the classification of the original time series. We focus on simplifications that select a subset of the original data points, and show that these typically have high Shapley value, thereby aiding interpretability. We employ these metrics to experimentally evaluate four distinct simplification algorithms, across several TSC algorithms and across datasets of varying characteristics, from seasonal or stationary to short or long. We subsequently perform a human-grounded evaluation with forward simulation, that confirms also the practical utility of the introduced metrics to evaluate the use of simplifications in the context of interpretability of TSC. Our findings are summarized in a framework for deciding, for a given TSC, if the various simplifications are likely to aid in its interpretability.