ShapeCond: Fast Shapelet-Guided Dataset Condensation for Time Series Classification

2026-02-09

AI Summary

The rapid growth of time series data poses significant storage and computational challenges, yet existing compression methods—primarily designed for images—often fail to preserve critical local patterns such as shapelets. To address this gap, this work proposes ShapeCond, a novel framework that, for the first time, integrates shapelet prior knowledge into time series dataset compression. By leveraging shapelet-guided optimization, ShapeCond synthesizes compact training sets that explicitly retain discriminative local structures while decoupling synthesis cost from sequence length. Extensive experiments on multiple benchmark datasets demonstrate that ShapeCond substantially outperforms current state-of-the-art methods, achieving up to a 10,000-fold speedup in synthesis (e.g., on the Sleep dataset) and consistently improving downstream classification accuracy.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Data Mining & Knowledge Management: Data Compression
• Search and Optimization: Combinatorial Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Time series data supports many domains (e.g., finance and climate science), but its rapid growth strains storage and computation. Dataset condensation can alleviate this by synthesizing a compact training set that preserves key information. Yet most condensation methods are image-centric and often fail on time series because they miss time-series-specific temporal structure, especially local discriminative motifs such as shapelets. In this work, we propose ShapeCond, a novel and efficient condensation framework for time series classification that leverages shapelet-based dataset knowledge via a shapelet-guided optimization strategy. Our shapelet-assisted synthesis cost is independent of sequence length: longer series yield larger speedups in synthesis (e.g., 29$\times$ faster over prior state-of-the-art method CondTSC for time-series condensation, and up to 10,000$\times$ over naively using shapelets on the Sleep dataset with 3,000 timesteps). By explicitly preserving critical local patterns, ShapeCond improves downstream accuracy and consistently outperforms all prior state-of-the-art time series dataset condensation methods across extensive experiments. Code is available at https://github.com/lunaaa95/ShapeCond.

Problem

• dataset condensation
• time series classification
• shapelets
• temporal structure
• local discriminative motifs

Innovation

• dataset condensation
• time series classification
• shapelets
• efficient synthesis
• temporal structure