Time2Graph: Revisiting Time Series Modeling with Dynamic Shapelets

Abstract

Time series modeling has attracted extensive research efforts; however, achieving both reliable efficiency and interpretability from a unified model still remains a challenging problem. Among the literature, shapelets offer interpretable and explanatory insights in the classification tasks, while most existing works ignore the differing representative power at different time slices, as well as (more importantly) the evolution pattern of shapelets. In this paper, we propose to extract time-aware shapelets by designing a two-level timing factor. Moreover, we define and construct the shapelet evolution graph, which captures how shapelets evolve over time and can be incorporated into the time series embeddings by graph embedding algorithms. To validate whether the representations obtained in this way can be applied effectively in various scenarios, we conduct experiments based on three public time series datasets, and two real-world datasets from different domains. Experimental results clearly show the improvements achieved by our approach compared with 17 state-of-the-art baselines.

An extended version with 11 pages including appendix; Accepted by AAAI'2020

References in corpus (6)

• XGBoost: A Scalable Tree Boosting System
• DeepWalk: Online Learning of Social Representations
• Graph Embedding Techniques, Applications, and Performance: A Survey
• LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection
• Scalable Link Prediction in Dynamic Networks via Non-Negative Matrix Factorization
• Composing graphical models with neural networks for structured representations and fast inference

Cited by in corpus (6)

• A Review of Deep Learning Methods for Irregularly Sampled Medical Time Series Data
• Self-Supervised Time Series Representation Learning by Inter-Intra Relational Reasoning
• Event2Graph: Event-driven Bipartite Graph for Multivariate Time-series Anomaly Detection
• Time-Series Event Prediction with Evolutionary State Graph
• Diffeomorphic Transformations for Time Series Analysis: An Efficient Approach to Nonlinear Warping
• From Cities to Series: Complex Networks and Deep Learning for Improved Spatial and Temporal Analytics*