DBIECM-an Evolving Clustering Method for Streaming Data Clustering

To address the problem of the difficulty of traditional clustering methods to adapt to online clustering of streaming data and on the basis of the research on the evolutionary clustering method (ECM), this paper proposes a Davies-Bouldin index evolving clustering method for streaming data clustering (DBIECM). This method has improved the updating process of the clustering center and the radius of ECM and introduced the Davies-Bouldin Index (DBI) as the evaluation criterion for data classification. Compared with the traditional clustering method, DBIECM has better adaptability for stream data clustering. The experiments show that DBIECM has a better clustering effect on the evaluation criteria of the objective function value, DBI, as well as better accuracy and purity compared with ECM.