Visual sampling based clustering algorithm VSC

This study attempts to achieve two goals: (1) The novel visual sampling based clustering algorithm VSC is proposed, based on the visual sampling principle. The clustering algorithm VSC incorporates the visual sampling principle together with the famous Weber law such that it has two distinctive advantages: (a) it is insensitive to initial conditions and very effective for convex datasets; (b) the reasonable cluster number can be effectively determined by the new Weber-law-based clustering validity index. Our experimental results demonstrate its success. (2) The link relationship between our algorithm VSC and algorithm SCA. Both theoretical analysis and experimental results show that in many cases, our algorithm VSC here has almost the same clustering results as algorithm SCA. This fact reveals that our algorithm can be utilized to overcome the drawback of SCA, i.e., the parameter γ therein is very difficult to be well determined.