针对常规算法不能适应数据流的动态环境问题,提出一种基于网格和分形维数的数据流聚类算法,有效地完成了对数据流的分析任务,克服了传统聚类方法的不足,把整个挖掘过程分为在线进程和离线进程,最终实现数据流的聚类.
Considering coventional methods can not adapt to the dynamic enviroment of data streams,a clusterng method is proposed based on grid and fractal dimension.It completes the analysis of data stream effectively,and overcomes deficiency of the traditional clustering algorithm.With this method,the mining procedure is divided into gric-clustering and gral-clustering.Finally,this method can be used to get final clusters for data stream.