On the Improvement of K-means Clustering Algorithm

Clustering analysis plays an important role in scientific research and commercial applica- tion. The K-means algorithm is the indirect clustering algorithm based upon comparability measure- ment between points,but it has some faults,it uses means as representative point,the point can't fig- ure the distributional structure of the mode,thus some important information is missed. An improved K-means clustering algorithm is studied. Kernel function distance is used to replace Euclidean dis- tance. Experimental results show that clustering effect is higher veracity and more steady with im- provement of K-means clustering algorithm.