Dynamic Clustering of Evolving Streams with a Single Pass

Stream data is common in many applications, e.g., stock quotes, merchandize sales record, system logs, etc.. It is of great importance to analyze these stream data. As one of the most commonly used techniques, clustering on streams can help to detect and monitor correlations among streams. Due to the unique nature of streaming data, direct application of most existing clustering algorithms fails to deliver efficient results. In this project, we introduce a novel model of stream cluster which employs a weighted distance measure. In addition, we device a novel efficient algorithm which can effectively discover all stream clusters.