Robust and Efficient Noisy-Label Time-Series Classification via Dynamic Time Warping Based Granular Ball Computing

2026-08-12

AI Summary

This work addresses the susceptibility of dynamic time warping (DTW)-based nearest neighbor classifiers to label noise and their high computational cost during inference in time series classification. To mitigate these issues, the authors propose a granular-ball-based classification framework that aggregates temporally similar training samples into granular balls and performs classification at the granularity level. Two DTW-driven strategies for constructing granular balls are introduced. Experimental results on four benchmark datasets demonstrate that the proposed method significantly enhances robustness against symmetric label noise while substantially reducing the number of DTW computations required during inference. The approach thus achieves a favorable trade-off between classification accuracy, computational efficiency, and noise resilience.

Abstract

Dynamic Time Warping (DTW)-based Nearest-Neighbor (NN) classifiers are effective for time-series classification but are vulnerable to mislabeled training samples and require numerous DTW computations during inference. We propose DTW-based Granular Ball Computing (DTW-GBC), which organizes temporally similar training samples into granular balls and performs classification at the granule level. We further develop two granular-ball construction strategies for DTW-GBC. Experiments on four benchmark datasets with symmetric label noise show that the two DTW-GBC variants generally mitigate the performance degradation caused by label noise while requiring substantially fewer comparisons than DTW-based 1-NN during inference. These findings suggest that DTW-GBC provides a favorable balance between classification robustness and inference efficiency.

Problem

• time-series classification
• label noise
• Dynamic Time Warping
• nearest-neighbor classifier
• inference efficiency

Innovation

• Dynamic Time Warping
• Granular Ball Computing
• Noisy-Label Robustness
• Time-Series Classification
• Inference Efficiency