Abstract
The purpose of this study is to evaluate the performance of five supervised machine learning methods for the task of automated genre identification of classical Arabic texts using text most frequent words as features. We design an experiment for comparing five machine-learning methods for the genre identification task for classical Arabic text. We set the data and the stylometric features and vary the classification method to evaluate the performance of each method. Of the five machine learning methods tested, we can conclude that Support Vector Machine (SVM) are generally the most effective. The contribution of this work lies in the evaluation of the five machine learning methods for the task of genre identification for classical Arabic text using stylometric features.