The key challenge in text classification techniques is the overall performance. In this paper we discuss the need of using parallel/distributed data mining algorithms for text classification. Three classification algorithms: k-NN Classifier, Centroid Classifier and Naive Bayes Classifier are considered. As data is growing the need of Hadoop, Map Reduce and Spark models is more important in data science. The overall outcome of the paper is the actual findings on the algorithms' efficiencies, such as accuracy of correct classification and the speed of execution. The empirical results show that Centroid Classifier is the most accurate classifier in this case with up to 95% accuracy compared to k-NN which is 92% accurate and Naive Bayes with 91.5%.