Efficient Time Series Clustering from Multiscale Reservoir Dynamics with Granular-Ball Anchoring Graph Optimization

2026-06-10

AI Summary

This work addresses the longstanding challenge in time series clustering of balancing accuracy and computational efficiency—where conventional similarity-based methods suffer from high computational complexity and deep learning approaches incur substantial training overhead. To overcome these limitations, we propose MSRGC-Net, a novel framework that integrates training-free multi-scale reservoir representations, granular-ball-computation-driven adaptive anchor graph construction, and a cross-scale consensus optimization strategy. By design, MSRGC-Net eliminates the need for backpropagation and avoids reliance on large numbers of trainable parameters. Extensive experiments on multiple univariate and multivariate benchmark datasets demonstrate that MSRGC-Net achieves superior clustering performance compared to state-of-the-art methods while significantly improving computational efficiency.

Abstract

Time-series clustering remains challenging due to the inherent trade-off between clustering effectiveness and computational efficiency. Similarity-based methods often suffer from quadratic complexity caused by pairwise distance computations, while deep learning-based approaches typically rely on costly iterative training and a large number of trainable parameters. In this paper, we propose MSRGC-Net, an efficient time-series clustering framework that integrates multiscale reservoir computing, granular-ball-based anchoring graph construction, and consensus learning. MSRGC-Net adopts a training-free reservoir computing paradigm to extract multiscale temporal representations from raw time series without backpropagation, significantly reducing computational overhead. To capture the intrinsic structure of the resulting representations, granular-ball computing is employed to adaptively model data distributions via density-consistent regions, yielding compact and robust anchor graph representations. Furthermore, a consensus-based anchoring graph optimization strategy is introduced to effectively align multiscale reservoir representations and integrate complementary information across temporal scales. Extensive experiments on widely used univariate and multivariate benchmark datasets demonstrate that MSRGC-Net consistently outperforms state-of-the-art methods in clustering performance while maintaining superior computational efficiency.

Problem

• time-series clustering
• computational efficiency
• similarity-based methods
• deep learning
• quadratic complexity

Innovation

• multiscale reservoir computing
• granular-ball computing
• anchoring graph optimization
• training-free clustering
• time series representation