Personalized Search Recommendation Based on Gradual Forgetting Collaborative Filtering Strategy

The existing search engines are always lack of the consideration of personalization and display the same search results for different users despite their differences in interesting and purpose. So through analyzing the dynamic search behavior of users, the paper introduces a new method of using a keyword query graph to express the personalized search behavior of the user, and constructs a dynamic and personalized search behavior profile for each user according to their search records. In order to reflect the dynamic changes with time of the user's preference, the paper introduces non-lineal gradual forgetting collaborative filtering strategy into the personalized search recommendation model. By calculating the similarity between every two users, the model can do the recommendation based on neighbors and be used to construct the personalized search engine.