Text categorization is the process of sorting text documents into one or more predefined categories or classes of similar documents. Differences in the results of such categorization arise from the feature set chosen to base the association of a given document with a given category. This process is challenging mainly because there can be large number of discriminating words which render many of the current algorithms unable to complete this. For most of these tasks there exist both relevant as well as irrelevant features. The objective here is to bring about a text classification on the basis of the features selected and also pre-processing to bring down the dimensionality and increase the accuracy of classification of the feature vector. Here the most commonly used methods are meta-heuristic algorithms in order to facilitate selection. Artificial fish swarm algorithm (AFSA) takes the underlying intelligence of the behaviour of fish swarming to combat the problems of optimization as well as the combinatorial problems. This method has been greatly successful in diverse applications but does suffer from certain limitations like not having multiplicity. Therefore, a modification has been proposed to AFSA which is MAFSA that has a crossover in its operation in order to bring about an improvement in the text classification selection. SVM or Support Vector Machine, Adaboost classifiers and naïve bayes are all used here. MAFSA has proved itself to be superior to AFSA in terms of precision and also the selected feature numbers.