查看文章

The thesis puts forward a new vector space retrieval algorithm based on association diagram extension of key words. By using key words and the related words appearing simultaneously in a large scale, the algorithm allows generating the association diagram which indicates the simultaneous relationship between key words. The degree of association between any two key words is represented by mutual information. In addition, the algorithm derives the weight of key words in retrieval vector via association diagram, thus the vector space retrieval based on association diagram extension of key words is realized. As proved by the experiment, this algorithm has a good retrieval effect and satisfactory recall and precision rate in case of relatively small expense.[PUBLICATION ABSTRACT]