مقالات
Automatic classification of Islamic Jurisprudence Categories
نویسندگان
Mohammad Hossein Elahimanesh, Behrouz Minaei-Bidgoli, Hossein Malekinezhad
چکیده
This paper evaluates some of text classification methods to classify Islamic jurisprudence classes. One of prominent Islamic sciences is jurisprudence, which explores the religious rules from religious texts. For this study the Islamic Jurisprudence corpus is used. This corpus consists of more than 17000 text documents covering 57 different categories. The major purpose of this paper is evaluating text to numerical vectors converting methods and evaluating different methods of calculating proximity matrix between text documents for religious text classification. The results indicate that the best classification efficacy is achieved especially when 3-grams indexing method and KNN classifier using cosine similarity measure are applied. We reported 87.3% performance for Islamic jurisprudence categories classification
کلیدواژهها
Text Classification, K-Nearest neighbor, Islamic Jurisprudence, N-grams