Abstract:

The traditional query expansion method  can not retrieve information according to different users’ requirement, so a new query expansion method is proposed on the basis of users’ interest and term relation. A series of terms reflecting users’ interest are obtained by mining the web log records, and the weights of original queries′  are adjusted according to the synonym relativity between these terms and the original words, then the original queries are expanded by using the ontology relevance relativity between these terms and the original words.  Experimental results show that the approach can enhance the accuracy ratio and the recall compared to the traditional query expansion method.

Key words:  query expansion; web log; user interest; synonym; ontology relevance