A Signal Filter Based Clustering Algorithm

In this paper, we introduce SFCLUS (signal filter based clustering algorithm) an effective and efficient approach to the clustering problem. Using the concept of signal filter to reduce the noise level and find the approximate locations of clusters. A new mathematics morphological clustering operator is designed to discover clusters around those locations. The combination of signal filtering and mathematics morphology can achieve high accurate clustering result and insensitive to the grid size. In contrast to existing approaches, SFCLUS is able to detect arbitrarily shaped clusters; it is very efficient with a complexity of O(N); it can distinguish clusters of different density; it is insensitive to large amounts of noise; it has a determinate result, insensitive with respect to input data; it is not sensitive to the grid size.