Skip to Main Content
User interest is an important basis for providing users with personalized search results. In this paper, the personalized query expansion is improved by combining the notion of user interest and domain knowledge. We present an approach to personalized search that involves calculating and annotating the interest weight of terms which are implicitly derived from user browsing history, and matching them with different domain dictionaries, finally selecting the expanding terms from three aspects: ontology-based correlation degree between expanding term and query term, interest weight and domain proportion. Our experiments show that the method is effective and is better than the traditional method, especially in the case of the query term belonging to more than one field whose advantage is more obvious.