AUTOMATIC SUMMARIZATION ON TRANSLATED AL- QURAN TEXT CORPUS

• Main Author: Abdul Aziz Syafiq, Aqil
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/69187
• Institution: Institut Teknologi Bandung

Al-Quran is the life guide book for muslims around the world, the Al-Quran contains knowledge, laws, and lessons for muslims. Al-Quran consists of verses and chapters, Al-Quran language is structured differently from normal language usage and this might cause difficulties, This research focus on summarization methods found through literature study to implement extractive and abstractive summarizations. Extractive summarization will be done using transformers, and statistical methods like LSA and Textrank. Abstractive method will be done using text generation or sequence-to-sequence transformer models. This research focuses on finding the best model and parameters that can summarize Al-Quran text. The data used for evaluation will be made from the Al-Quran corpus using a reference book. Extractive method shows better performance compared to abstractive methods, with IndoBERT representing the extractive method as the overall best method. But it needs to be said that the performance of every method in this research has not reached a level which could be used effectively or realistically in learning Al- Quran.