In the proposed algorithm, they extend the K-Means clustering process to calculate a weight for each dimension in each cluster and use the weight values to identify the subsets of important ...