Instance-Based Counterfactual Explanations for Time Series Classification

In recent years, there has been a rapidly expanding focus on explaining the predictions made by black-box AI systems that handle image and tabular data. However, considerably less attention paid to of opaque handling time series this paper, we advance novel model-agnostic, case-based technique – Native Guide generates counterfactual explanations for classifiers. Given query series, \(T_{q}\), which classification system predicts class, c, explanation shows how \(T_{q}\) could change, such an alternative \(c'\). The proposed instance-based adapts existing instances in case-base highlighting modifying discriminative areas underlie classification. Quantitative qualitative results from two comparative experiments indicate plausible, proximal, sparse diverse are better than those produced key benchmark methods.