Abstract:
A search engine aims to retrieve full documents whereas a question answering system aims to extract the exact answer. A question answering system involves the process of accepting a NL (Natural Language) question, analyzing, and processing to match against a knowledge base to generate the right answer from documents for users. For the Holy Quran this involves accepting the NL question and processing it to retrieve the right verse or verses from our Quran knowledge base. Question answering systems can use two types of algorithms: rule based techniques and/or AI (Artificial Intelligence) based techniques. Question Answering systems have three main components: question classification, information retrieval and answer extraction. We present a rule-based system for the Holy Quran that retrieves the right verse(s) from the Holy Quran instead of generating NL answers. We use a java program to extract the answer from a MS-Access database which contains our knowledge base for our Quran question answering system. We find that the system gives better results for the question after improving the system by removing stop words.
