Proposal of a Mathematical and Computational Method for Determining the Optimal Number of Clusters in the K-Means Algorithm uri icon