Automatic Question-Answering Based on Wikipedia Data Extraction

Abstract

The question-answering (QA) system plays a vital role in artificial intelligence. The goal of automatic QA is to find out correct answers to the natural language questions raised by users from some specified datasets. Data on the Web is about everything and contains almost all the answers we needed. Wikipedia is a collaboratively edited, multilingual, free Internet encyclopedia which contains more than 30 million articles and can be considered to be a huge dataset for us to extract answers from. In this paper, we propose a method to integrate Wikipedia data extraction with automated question answering, which allows us to extract answers to questions from Wikipedia pages in real time. Experimental results show that the QA system based on our proposed method achieves good precision while answering questions.