A Clustering Algorithm for Multiple Data Streams Based on Spectral Component Similarity

摘要

A new clustering algorithm for multiple data streams using a novel similarity metric is proposed. Based on spectral component similarity analysis, the algorithm can effectively cluster streams which show similar behavior with some unknown time delay. Auto-regressive modeling technique is exploited to measure lag correlation between data streams, which is used as the distance metrics for clustering. Based on a sliding window model, the algorithm can continuously report the most recent clustering results, and dynamically adjust the number of clusters. Our experimental results on real and synthetic datasets show that our algorithm has high clustering quality, efficiency, and scalability than other similar methods.