A Research On Model Building Of Personalized Recommendation For Sports Information Resources Based On Collaborative Filtering

In recent years,research on Personalized Recommendation Technology(PRT)has become a hot spot,not only for commercial application,like e-commerce,music,and film etc.,its related academic achievements has also increased.The hyper growth of Internet and the promotion of the"National Fitness”to a nation-wide strategy have made an exponential growth in sports information resources.The consequence is that the overloaded information makes more troubles for users to select resource.Thus,the introduction of PRT into sports information resources will greatly improve the efficiency and accuracy of users' access to resources.In this background,the aim of this paper is to explore the theory of PRT,its application in sports information resources and the characteristics of users.By introducing PRT into Sports Academic Library&Information System(SALIS),this study was designed to find out appropriate technology,algorithm,realization process and develop a personalized recommendation model for sports information resources.Moreover,it is aim to provide a theoretical reference and a basis for later research on PRT in sports information resources,and to give technical support for personalized referral services in actual system platform.The main contents of this paper are as follows:(1)Investigate the SALIS by case study and mathematical statistics methods,analyze the characteristics of sports information resources and its users.(2)Based on findings of above-mentioned studies,the collaborative filtering recommendation technology is recommended for sports information resources personalization.Common model,realization process,classification and several commonly used similarity algorithms of collaborative filtering recommendation technology are analyzed and discussed in this paper.(3)Combining the characteristics of sports information resources and their users;determining user interest model and metadata model of sports information resources by mathematical statistics and expert interview method;establishing user's appraisal rules and constructing "User-Item"scoring matrix for sports information resources.(4)Based on the "User-Item" scoring matrix,using Pearson correlation coefficient to calculate similarity degree between different sport information resources and between their users;constructing the PRT model for sports information resource based on collaborative filtering technology.