关键词: 数据流, k-means, 偏向最近, 倾斜时间窗口, 聚类分析

Abstract: A recent-biased clustering algorithm of data stream based on tilted-time window is proposed. First, the algorithm segments sliding window equal in length to form no overlap data blocks(basic window), then extracts feature of every data block through Haar wavelet transform, and preserves detailed feature of recent data by varying number of wavelet coefficients of data block, namely more recent data block, more wavelet coefficient preserved, and vice versa. Finally, by defining recent-biased distance of data stream, the recent-biased clustering algorithm of data stream based on tilted-time window is implemented. Remarkably faster computational speed and higher efficiency are achieved by this algorithm. Experiments on real validate the proposed algorithm.

Key words: data stream, k-means, recent-biased, tilted-time window, clustering analysis