This is the Fuzzy C Mean Clustering algorithm implemented in C, and used over IRIS dataset.
Clustering accuracy achieved: 92.67% (approx).
Instead of initializing the membership values (Uij) I initialized the cluster centroids based on the initialization method described here (Even though this paper is discusses a centroid initialization method for K-means algorithm, not Fuzzy-C-Means).
I then calculated the membership values from the cluster centroids following the standard Fuzzy-C-Means equations. The rest of the program follows the standard norms of Fuzzy-C-Means.