Comparison of Bagging and Boosting in Imbalanced Multilabel of Al-Quran Dataset

This paper address multi-label text classification in the case of Al-Quran in English translation. This research built from two models which are Bagging and Boosting in comparison. Those models are chosen because previous research gives a good performance, especially in the imbalanced dataset. Imbalanced data is one problem in classification and this research focuses on that problem. This research started by preprocessing data, then TF-IDF is used for feature extraction. To measure the prediction of classification, this research uses the Hamming Loss to see the performance. Based the several experiments, the best performance result is obtained by Random Forest with a Hamming Loss value of 0.1019. Then, For the fastest computational time is given by Extreme Gradient Boosting (XGBoost) with 224.76 seconds for running the system. Thus, Bagging and Boosting have their advantage in classification.