Data warehouse design for chinese intelligent question answering system based on data mining

Abstract

According to the defects of current Chinese question answering system database, a set of scheme about data warehouse design based on data mining was put forward. An improved association rules algorithm based on keywords was applied to compute the correlation value between words in order to get the similarity of questions. After the similarity is calculated, the best answers could be extracted and the k-means clustering algorithm could be performed on the database. Experiments showed that the scheme had many merits such as intelligence, continuous self-renewing ability and improving answering executive efficiency etc. © 2007 IEEE.