关键词: 模糊，粒度，K-means，文本聚类，归一化距离函数

Abstract: The traditional K-means is very sensitive to initial clustering centers and the clustering result will wave with the different initial input. To remove this sensitivity, a new method was proposed to get initial clustering centers. This method is as follows; provide a normalized distance function in the fuzzy granularity space of data objects, then use the function to do a initial clustering work to these data objects who has a less distance than granularity试，then get the initial clustering centers. The test shows this method has such advantages on increasing the rate of accuracy and reducing the program times.

Key words: Fuzzy,Uranular computing,K-means,Text cluster,Normalized distance function