基于数据场的改进DBSCAN聚类算法

BSCAN algorithm is a typical density-based clustering algorithm. The algorithm can discover the arbitrary-shaped clusters. However, the clustering results depend on the two parameters Eps and MinPts which are chosen by users. And for some datasets with large density differences, either may the clustering results have the uncorrected cluster number, or may the algorithm label part of the data as noise. In this paper, a new clustering algorithm called an improved DBSCAN algorithm based on data field is proposed. As data field can nicely describe the data distribution and reflect the data relationship, the advantage is taken by the algorithm. The concept of average potential difference is introduced and each class's Eps and average potential difference are dramatically determined during the clustering process. In this way, the algorithm can receive better clustering results for some clusters with large density differences. Experimental results indicate that the proposed algorithm performs better than the DBSCAN algorithm.