Topic Classification of Islamic Questionand Answer Using Naive Bayes Classifier

Abstract

Topic classification is one of the most important components in an automatic Islamic question-answering system, which is capable of automatically providing the most relevant answers given a question about the Islamic issue. In our research, the Islamic question-answering system to be built collects existing Islamic questions and answers from trusted online Islamic consultation websites. To speed up the search for finding the appropriate answers, each Q & A entry should be classified into a topic. However, the question-answering system cannot directly adopt the topic classes provided by the online Islamic consultation websites, because different websites use different classifications. Since the number of Q & A entries could reach tenth thousands, an automatic topic classification method is required. In this paper, a naive Bayes classifier is implemented to classify Q & A entries. The classifier gives a satisfying result with 0.88 precision.