Question Answering Terjemah Al Qur'An Menggunakan Named Entity Recognition

Abstract

Al Quran is the guidance of all Muslims all over the world, including Indonesian society that most of the people are moslem. the purpose of this research is to help the society for seeking the quranic proposition in the free question forms or more often known Question Answering System (QAS). In this study the algorithm used to provide the answers is Rule Based. Rule Based on previous research by indexing documents using lucene with the same set of data and same category (Who, When, and Where) getting low accuracy result (60%, 60%, and 40%). In this research, the improvement of architecture by involving Named Entity Recognition (NER) and improvement of indexing and determination of candidate answer using Vector Space Model. The results of the improved architecture are proven increased accuracy with each accuracy of each category of 90%, 80%, and 50%.