一种改进的K-means算法

The author proposed an improved K-means clustering algorithm (CS-kmeans) to solve the problem that the current K-means algorithms are difficult to determine the number of clustering centers. First, the improved K - means algorithm will analyze the relations between the maximum inner-class distance and the minimum intra-class distance when Ideal clustering is acquired, then it starts to adjust the inner-class distance to make it smaller than the minimum interclass distance, and by the same token, to make the Inter-class distance greater than the maximum distance. Finally, by doing so, it can automatically segment and merge the number of categories so that the appropriate number of clusters can be determined. This improved K - means algorithm, as results can tell, is more effective than the old ones in determining the number of clustering centers and improving clustering qualities