ExplanationIntermediate
What is the purpose of TS-GAC in multivariate time series classification?
TS-GAC is a framework that integrates graph-based augmentations and graph-level contrasting to maintain spatial consistency during contrastive learning for multivariate time series classification. It promotes spatial consistency across sensors, allowing an encoder to be trained without supervision so that meaningful feature representations can be obtained for downstream classifiers.
Key points
- TS-GAC maintains spatial consistency during contrastive learning for MTS classification.
- It uses node and edge perturbations to create weak and strong views of each sample.
- Node-level contrasting enforces similarity between corresponding sensors across views.
- Graph-level contrasting aligns global features of the same sample across views while differentiating them from other samples.
- The framework enhances representation quality, allowing an unsupervised encoder to produce useful features for a downstream classifier.
Source:AI for Time Series_ Volume 1_ Unlocking Patterns with Deep Learning· Multivariate Anomaly Detection with Self-learning Graph Convolutional Networks· p. 168
Related questions
AI for Time Series_ Volume 1_ Unlocking Patterns with Deep Learning
Min Wu;Emadeldeen Eldele;Zhenghua Chen;Shirui Pan;Qingsong Wen;Xiaoli Li;
First edition · CRC Press