Repository for "Scaling Time Series Classification via XAI-Driven Data Reduction" accepted for AALTD workshop at ECML-PKDD 2026.
Explainable AI (XAI) for time series has seen significant algorithmic growth, but its utility in providing measurable performance gains for downstream tasks remains under-explored. This paper bridges this gap by introducing drXAI, a novel methodology that repurposes XAI attribution methods for effective data reduction in Time Series Classification (TSC). The core challenge in modern TSC is scalability; state-of-the-art models, such as Transformers, exhibit quadratic complexity relative to sequence length and linear complexity relative to the number of channels. This renders them c omputationally prohibitive for massive datasets. drXAI addresses this by using a fast, GPU-accelerated classifier (Hydra) to generate local attributions. We aggregate these into global feature importance scores and employ an automated elbow-cut heuristic to select the most salient features without requiring manual thresholds. We evaluate our approach on both synthetic and real-world univariate and multivariate datasets. On synthetic benchmarks, drXAI successfully recovers ground-truth features where traditional baselines fail. On real-world data, drXAI achieves 80–90% data reduction while maintaining classification accuracy comparable to models trained on the full dataset. Most importantly, we show that drXAI allows resource-intensive models like ConvTran to scale to datasets that were previously inaccessible due to memory constraints. Our results show the benefits of using XAI not just for interpretability, but as a robust tool for feature selection and scalability in time series analysis. All our code and data are openly available.
Mean accuracy of each selection (and All Features) for the 4 MTSC real-world datasets and the 3 SOTA classifiers,
yielding 12 results in total.
Mean percentage of data saved by each selection for the 4 MTSC datasets.
Looking at the two above tables, only our methodology (drXAI) is at the top both for accuracy and data reduction.
Mean accuracy of each selection (and All Features) for the 5 UTSC datasets used and the 3 classifiers,
yielding 15 results in total.
Mean percentage of data saved of each selection for the 5 UTSC datasets.
Looking at the two above tables, only our methodology (drXAI) is at the top both for accuracy and data reduction.
Code developed using python 3.11.14. Install libraries in requirements.txt
Executables are:
Trains a classifier on either original dataset or a reduced dataset.
Generates a selection of features based on the global feature importance scores. Only tested using hydra as "explainer classifier"
Used to achieve baseline selections in our experiments. It used a different virtual evitomen
Models, datasets, selections and results are available here https://zenodo.org/records/19045954