Abstract
In recent years, uncertain data have gained increasing research interests due to its natural presence in many applications such as location based services and sensor services. In this paper, we study the problem of clustering uncertain objects. We propose a new deviation function that approximates the underlying uncertain model of objects and a new density-based clustering algorithm, U-DBSCAN, that utilizes the proposed deviation. Since, there is no cluster quality measurement of density-based clustering at present. Thus, we also propose a metric which specifically measures the density quality of clustering solution. Finally, we perform a set of experiments to evaluate the quality effectiveness of our algorithm using our metric. The results reveal that U-DBSCAN gives better clustering quality while having comparable running time compared to a traditional approach of using representative points of objects with DBSCAN.