Deep Temporal Clustering : Fully Unsupervised Learning of Time-Domain Features
This addresses the problem of temporal clustering for researchers and practitioners in fields with time series data, though it appears incremental as it combines existing components like autoencoders with a novel clustering layer.
The authors tackled unsupervised learning of time series data by proposing Deep Temporal Clustering (DTC), an end-to-end framework that integrates dimensionality reduction and clustering, and demonstrated it outperforms traditional methods across diverse domains like earthquakes and spacecraft sensor data.
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.