摘要： 针对时间序列数据流的增量聚类研究较少的现状，采用多维时态子空间聚类对数据流的增量聚类进行探究。多维时态子空间聚类是指在连续一段时间内，数据流中的值的距离小于2α，它的另一个要求是最后的聚类结果必须包含一定数量的数据流。聚类结果随时间的演变能持续增量地更新，这个更新机制采用滑动窗口的形式，把最早时刻的数据删除后，添加入新到达的数据。采用股票数据对算法进行测试与验证，实验证明，该算法效果较好。

关键词: 数据流, 增量聚类, 多维时态子空间聚类

Abstract: There is only a little research on incrementally cluster temporal data streams. This paper focuses on the problem of clustering temporal data streams based on the sliding window. Temporal Multiple-dimension Subspace α-Cluster(TMSC) is adopted. It consists of data streams, whose distance is less than 2α, and must contain a number of streams which are predefined. The result of the clustering is updated through time evolving. Sliding window mechanism is adopted to realize it. The earliest data is removed, and the new coming data is added. Stock data is used to test the algorithm, and the result is quite well.

Key words: data stream, incrementally clustering, Temporal Multiple-dimensional Subspace α-Cluster(TMSC

中图分类号: 

• TP311