Classification of Bulughul Maraam Categories: Prohibitions, Recommendations, and Information Using Extreme Learning Machine and Fasttext

Abstract

Hadith is the second source of Islamic law after the Quran. After the hadiths were compiled, Imam of Hadith created collections of hadiths, one of which is Imam Bukhari who compiled the book Bulughul Maraam, which is considered to have the highest level of authenticity. Digital collections of hadiths can now be found in the form of e-books and web pages, which help in the search for hadiths. The classification of hadiths is necessary to organize them by category, making it easier to search for hadiths based on their categories. Text mining is needed to classify hadiths because it can identify patterns in unstructured text. This research aims to improve the accuracy of classifying recommended, prohibited, and informational hadiths using a dataset of 7008 hadiths, which consists of primary data taken from the book Bulughul Maraam in the Indonesian language. Previously, similar research was conducted in 2017 that classified recommended, prohibited, and obligatory hadiths with an accuracy of 85%, but only for Sahih Bukhari hadiths. In this research, the same classification categories will be examined, proposing a different method, namely the Extreme Learning Machine method and Word2vec Fasttext for text representation with a larger dataset. The results of this research show a model accuracy of 86.31%, 86% precision, and 87% recall, indicating that the proposed model performs well in classifying hadiths.