Deep Temporal Clustering : Fully Unsupervised Learning of Time-Domain Features

Abstract

Unsupervised learning of time series data, also known as temporal clustering, is a challenging problem in machine learning. Here we propose a novel algorithm, Deep Temporal Clustering (DTC), to naturally integrate dimensionality reduction and temporal clustering into a single end-to-end learning framework, fully unsupervised. The algorithm utilizes an autoencoder for temporal dimensionality reduction and a novel temporal clustering layer for cluster assignment. Then it jointly optimizes the clustering objective and the dimensionality reduction objec tive. Based on requirement and application, the temporal clustering layer can be customized with any temporal similarity metric. Several similarity metrics and state-of-the-art algorithms are considered and compared. To gain insight into temporal features that the network has learned for its clustering, we apply a visualization method that generates a region of interest heatmap for the time series. The viability of the algorithm is demonstrated using time series data from diverse domains, ranging from earthquakes to spacecraft sensor data. In each case, we show that the proposed algorithm outperforms traditional methods. The superior performance is attributed to the fully integrated temporal dimensionality reduction and clustering criterion.

11 pages, 4 Figures, 1 Table

References in corpus (3)

• Deep Learning in Neural Networks: An Overview
• Towards K-means-friendly Spaces: Simultaneous Deep Learning and Clustering
• Self-Transfer Learning for Fully Weakly Supervised Object Localization

Cited by in corpus (14)

• CGC: Contrastive Graph Clustering for Community Detection and Tracking
• Unsupervised Representation Learning for Time Series with Temporal Neighborhood Coding
• "Layer-by-layer" Unsupervised Clustering of Statistically Relevant Fluctuations in Noisy Time-series Data of Complex Dynamical Systems
• Subtyping patients with chronic disease using longitudinal BMI patterns
• RTFN: A Robust Temporal Feature Network for Time Series Classification
• MixSeq: Connecting Macroscopic Time Series Forecasting with Microscopic Time Series Data
• TimeAutoML: Autonomous Representation Learning for Multivariate Irregularly Sampled Time Series
• Identifying botnet IP address clusters using natural language processing techniques on honeypot command logs
• Temporal clustering network for self-diagnosing faults from vibration measurements
• Clustering of Time Series Data with Prior Geographical Information
• CROCS: Clustering and Retrieval of Cardiac Signals Based on Patient Disease Class, Sex, and Age
• RTFN: Robust Temporal Feature Network
• Longitudinal patient stratification of electronic health records with flexible adjustment for clinical outcomes
• Instance Explainable Temporal Network For Multivariate Timeseries