Research on Personalized Topic Distillation Algorithm

This paper focuses on personalized information retrieval which is a research focus in Web information service, and itpsilas the key technology of enhance retrieval quality. As a user concept map composed of some concepts that extended from user queries based on user interests description, it´s the embodiment of user´s true intention. Each concept weight in the concept map is defined as a probability, and the whole weighted concept map can reflect the intensity distribution of user interests in the field. Thus semantic similarity and algorithm of two concepts can be defined with concept weight. In this paper, we proposes an improving approach of HITS algorithm with the topological structure of concept map. Experiment shows the algorithm is effective and feasible, and it can improve the quality of Web search engines efficiently.