Hybrid Mean-Variance Mapping Optimization to determine the number of clusters in network partitioning of Hydro Québec Power Grid

Finding out the number of clusters for a particular problem statement is quite an important aspect in clustering. Many clustering techniques have been experimented in the recent past providing somewhat satisfactory results. User supplied information as well as cluster validity indices are some of the expensive techniques regarding to the computation time. A novel technique has been proposed using Mean-Variance Mapping Optimization, an emerging heuristic optimization algorithm for determination of the cluster tendency for network partitioning in Hydro Québec Power Grid, Canada. It has been shown that no user supplied information is required and the change in value of control parameters of the algorithm doesn´t even change the result thus making it a fine algorithm for determination of clusters for any data set with higher efficiency.