Modeling Methods Of User Profile In Scientific Literature Personalized Recommendation System

Personalized recommender system is an efficient method for the user to select what he prefers from the mass information. In the information age, this technology is vastly needed in our real world, and it is an inevitable trend of the development of information service.This paper demonstrates the concept of personalized recommender system, analyses its correlated practical techniques in details, and attempts to research the modeling of user profile in the scientific literature personalized recommender system, finally prototype is realized.The user profile describes the user's personal information, major background, interest trend and previous behaviors. According to these information, the recommender system finds and predicts his demands for information so as to provide with personalized information recommendation service. The user profile is a major factor in the efficiency of the recommender system, thus modeling according to the user's interests is one of the problems which should be major concerned in personalized service.At present, the modeling and study approaches of user profiles have been already researched at home and abroad. But there is litter research in personalized recommender system in scientific literature recommendation based on individual users, combining the user's short-term and long-term interests to describe his changing demands as the information is being changed all the time. This paper describes the user profile in user-based personalized recommender system and designs the dynamic user profile. We provide dynamic user interests in the modeling process, and allow the user directly either to add or to delete the interest categories. We also provide the system with text examples to analyze the user's interest trend and combine the user's interests with the searched materials in order to find his new interests. We exactly combine the user's short-term and long-term interests, trying to accurately reflect the use's present interest.This paper includes four sections. The first one illustrates the necessity of the scientific literature personalized recommender system and applications on recommender system. The second section demonstrates the three major techniques in personalized recommender system. The third part discusses the user profile and puts forward the user profile modeling approach as well in scientific literature personalized recommender system. And in the last section it is designed and realized in details.