Study On Personalized Information Recommendation Based On Usage Mining Of Anonymous User

With the popularity of the Internet and the rapid development of the Web, people often "Lost" in the information ocean. Therefore, understanding the user's browser pattern to provide users with information they wanted, that is, personalized information recommended gradually become a hot study pot. Recent years, the personalized information services research has made a great progress. However, from previous studies, we found that there are following issues in the personalized information recommendation study, such as: the majority personalized information recommend system just concerned the register users; most personalized recommendation system considered not enough for new users; the majority of personalized system using a single user or item recommended clustering algorithm, lack of a comprehensive study of clustering.Based on the summary of existing study results, this paper focuses on following aspects to study the anonymous user personalized recommendation: firstly, discusses the steps of pretreatment of anonymous user in details, and gives an improved Maximal Forward Path identification algorithm for simplifying the process of pretreatment; secondly, based on the pretreatment of anonymous users, gives a recommendation system model and the framework; finally, puts the double cluster collaborative filtering recommendation into an emulation experiment to prove the algorithm is effective.