Text Classification of Translated Qur'anic Verses Using Supervised Learning Algorithm

Abstract

The Quran, comprising Allah's absolute divine messages, serves as guidance. Although reading the Quran with tafsir proves beneficial, it may not offer a comprehensive understanding of the entire message conveyed by the Al-Quran. This is due to the Quran addressing diverse topics within each surah, necessitating readers to reference interconnected verses throughout the entire chapter for a holistic interpretation. However, given the extensive and varied verses, obtaining accurate translations for each verse can be a complex and time-consuming endeavor. Therefore, it becomes imperative to categorize the translated text of Quranic verses into distinct classes based on their primary content, utilizing Fuzzy C-Means, Random Forest, and Support Vector Machine. The analysis, considering the obtained Davies-Bouldin Index (DBI) value, reveals that cluster 9 emerges as the optimal cluster for classifying QS An-Nisa data, exhibiting the lowest DBI value of 4.30. Notably, the Random Forest algorithm demonstrates higher accuracy compared to the SVM algorithm, achieving an accuracy rate of 66.37%, while the SVM algorithm attains an accuracy of 50.56%.