A Density Based k-means Clustering Algorithm
-
Abstract
For k-means algorithm to the initial cluster centers,sensitive to outliers shortcomings,we propose a density-based improved k-means algorithm.The algorithm introduces entropy and weighted distance,starting from the neighbor density,remove the isolated points on the algorithm while determining the initial cluster centers,making the cluster center is relatively stable.Experimental results show that the algorithm in terms of accuracy,operating efficiency has a very good improvement.
-
-