An improved C-means clustering algorithm

C-means algorithm needs the number of clusters and it is also sensitive to the initial partition and the input sequence. In order to overcome these disadvantages, an improved C-means algorithm with two-step is proposed. This algorithm assumes that the instances follow normal distribution. To resolve the sensitivity to the input sequence of the instances, these instances are sorted on their densities before clustering, and the proper instances are selected as the initial clustering centers with a definite process. New algorithm employs the theory of gravity to distribute the instances. Experimental results and comparisons are given to illustrate the performance of the new algorithm over that of C-means.