The Research And Design On Personal Model Based On Mining User's Interests

The increasing Internet information gave birth to the newly developed way of service—personal service. And as the core of this new service, personal modeling has also gained increased attention among researchers. After the study and analysis of the related technologies in personal modeling, a personal model based on mining user's interests is proposed in this paper. The model has been applied in PMSE system to provide references for personal analysis.This paper provides a detail design of the client-based model. Without much participation of the user, through mining the web pages visited by the user, the user's interests are automatically deduced from the user's implicit feedback. In the model, the user's interests are recorded and managed according to the time period and the type and level of the interests involved. According to the humans' forgetting rule, the model is regularly updated and optimized through selectively and unevenly forgetting the interests, thereby to appropriately render the characteristics of the user's interests.In this paper, the discussion is mainly focused on the following four aspects:(1) mine user's interests and extract the topics of user's interests;(2) conclude a calculable model from the user's interests;(3) forget and update the personal model automatically; and(4) analyze the queries in personal search using the model. Finally, experiments are conducted to verify the efficacy of themodel .The characteristics of the model are summarized, and further improved assumption towards the model is suggested.