Study on question-answering system based on Meta search engine

Meta search engine which can provide rich information is obviously an ideal source of answers to many types of questions. This paper describes a question answering system based on web, and puts forward to a new method to search information based on Meta search engine and new word discovery. In the information retrieval module, N-gram model and web pages which get from search engine are used to find new words. Experiments show that the addition of new words discovery not only greatly improves the precision of documents retrieval in the module of information retrieval, but also improves the accuracy of answer extraction.