Research And Application Of Collaborative Filtering Algorithm Based On Social Network Model

With the progress of science and the rapid development of information technology, We have entered the era of "big data". How to obtain useful information from a large number of information has become a problem. The related technology that applied to traditional search engines has been unable to meet the demand of retrieval in the era of "big data".Recommendation system can find the user’s interests and needs based on user’s personal information, then provide active and personalized service to users. So, recommendation system is able to solve the problem of retrieval requirement. The core of recommendation system is recommendation algorithm, which is the dominant factor of system performance.At present, collaborative filtering algorithm is widely used in recommendation systems. According to the historical actions information and basic information of the users’,collaborative filtering algorithm could find the nearest neighbors which are most similar to users firstly. Then predict the user’s interest items based on the nearest neighbors. But this algorithm also has some problems, such as Cold Start, sparse matrix, extendibility and so on.Social network analysis is a research method has become more perfect and mature, and has been successfully used in other subject areas. Social network is composed of nodes and connections, which is a collection of relationship. It links the research core, and concerns the relationship between individuals and the mutual influence. Social network and collaborative filtering method focus on the relationship between individuals.This article optimizes collaborative filtering algorithm by social network analysis methods to solve the problems of traditional algorithms and improve the recommendation accuracy. The main research work and innovation points of this paper divided into the following aspects:(1)The article presents the basic concepts and model of collaborative filtering and social networksl, briefly introduces its methods and effect. Provide theoretical guidance for the following research.(2) Optimize the Collaborative Filtering Algorithm by using group dynamics model of social network analysis. The algorithm take fully individual factors and environmental factors into consideration.In order to verify the recommendation performance of improved algorithm, design experiments on realistic data sets, comparative and analysis the results with other algorithms.(3) Optimize the collaborative filtering algorithm by using graph theory model of social network analysis. This algorithm construct the interactive grid items chart, redefine the concepts and calculation methods of similarity degree between items. Designe experiment, comparative and analysis recommendation performance of the algorithm.(4) Collect tourism data information in the real travel websites, analyze and preprocess these collected data, then provide personalized recommendation services for tourists using the improved collaborative filtering algorithm proposed in paper. Finally build a personalized recommendation system.