A Density-Based Clustering Algorithm for Uncertain Data

As the development of the data acquisition technology, the research of the uncertain data has been the center of people´s attention, and at the same time, the cluster of the uncertain data has been in use widely. In this paper, after studying the cluster of the uncertain data, considering characters of the uncertain data, we propose a improved algorithm. We called it En-DBSCAN. In order to adapt the request of uncertain data´s clustering we add probability factors and the theory of information entropy. The algorithm brings in a conception of probability radius to adjust uncertain data´s scope of EPS neighborhood and information entropy to reduce center point´s indeterminacy. Besides, this paper gives an analysis and confirmation about the algorithm.