Classification of Quranic Topics Using SMOTE Technique

Abstract

This paper aims to classify the Quranic topics that differ in their number of verses by applying the SMOTE technique. SMOTE is used to rebalance samples of minority classes in these Quranic topics. Moreover, SMOTE is combined with many classifiers to choose the best technique for the Quranic classification. Also, the k-values of SMOTE were studied to select the best values for the Quranic datasets and improve the performance of imbalanced classification. The SMOTE was implemented with many classifiers to choose the best one. The results showed that the Voted Perceptron classifier was the best technique when implemented with the SMOTE method to classify the Quranic topics. Also, it is concluded that the best range of K numbers in SMOTE method is [1, 10], to obtain the higher performance of Quranic classification.