A cluster boundary detection algorithm based on shadowed set

To meet the need of extracting cluster boundary from mixed attribute data in the field of data analysis, we propose a cluster boundary detection algorithm for mixed attribute data sets in this research, named CHASM(Cluster Boundary Detection Algorithm based on Shadowed Set). Based on the structure of clusters, the CHASM defines a new objective function according to the data set which is categorized into three collections, i. e. core, exclusion and shadow. Then CHASM updates the centroid information of clusters based on the variance of contribution degree among these collections to the clusters centroids. Finally, in an iterative optimization process, the CHASM can extract its shadow sets from each cluster to form the boundary of clusters. The experimental results, on both the synthetic data and real data with mixed attributes, numerical attributes and categorical attributes, show that CHASM can effectively detect cluster boundary with higher or similar accuracy to its rival methods. Furthermore, the CHASM can eliminate noise effectively.