The aim of this work is to develop a unique multi-category classifier to identify storm type based on radar-derived attributes. Due to the quantity of attributes a principal components analysis (PCA) technique is used to reduce the multidimensional dataset to a lower dimension for analysis. The analysis consists of classifying the data into different groups; it is partitioned into subsets (clusters) where they share some common traits. Clustering can distinguish values of some features or characteristics of a situation that might have been overlooked otherwise which then can possibly lead to stratification of data. To optimize this procedure a genetic algorithm (GA) is implemented as a search technique to find exact and approximate solutions to the storm identification.