Clustering Research Based on Density Gradient

Cluster analysis, intending to help users understand the natural grouping or structure in a data set, has received a lot of attention in the last few years. Aimed to solve difficult problems in clustering with irregularly distributed data set, a new clustering algorithm based on density gradient is provided. With analysis of density of each point and its neighbors, the algorithm searches points with the maximum density and takes them as centers of original clusters. Then it combines some little clusters into larger clusters according to the distribution of boundary points between neighbor clusters. Experimental results show that the new algorithm has better performance than DBSCAN.