A Multi-label Classification on Topic of Hadith Verses in Indonesian Translation using CART and Bagging

Hadith is a source of law for Muslims after the al-qur'an, in which there are instructions form words, actions, attitudes, and others. must be studied practiced by Muslims, then used as way life al-qur'an. Classifying hadith to make it easier learn looking at text pattern translation Bukhari based on three classes or categories suggestions, prohibitions, information. The classification carried out multi-label classification. process uses N-gram TF-IDF feature extraction, CART bagging methods, hamming loss evaluation methods. Bagging cover shortcomings CART, namely, model less stable, which, if slight change training data, will have significant effect resulting learning model. Several testing methods were obtain best hammer value this study. Based several tests that been out, 0.1914 80.86%. These results indicate use can help increase accuracy 5%.