Improving the Scoring Process of Question Answering System in Indonesian Language Using Fuzzy Logic

Basically, Question Answering System (QAS) show the answer from sources with several candidates. The key success factor is how system knows that the candidate answer is exactly the correct answer. To determine the correct answer, this paper aims at applying the fuzzy logic to improve the score of the candidate answers for the QAS using Indonesian language. The methodology is preprocessing, question analyzer, passage retrieval, passage scoring, scoring for similar text, measuring keyword and candidate answer, fuzzy logic controller, rules and extraction answer. There are four fuzzy variables applied namely Lucerne's score, Wijono's algorithm score, a similarity between passage and query text, the distance between keywords and text entity. This research focuses on applying the fuzzy logic on the scoring process of QAS and succeeded in increasing the accuracy level of QAS.