An Improved Collaborative Filtering Recommendation Model and Method Based on Social Trust

Abstract

Trust relationship in the social business environment has a profound impact on consumers' purchase behavior and decision-making, and has become an important factor to support the development of online business activities. Collaborative filtering recommendation algorithm based on user history evaluation data usually faces the problem of data sparseness; that is, the sparse rating data leads to the decline of recommendation quality. In order to solve this problem, the combination of auxiliary data has become an inevitable trend. Therefore, with the development of social media, trust-based social recommendation algorithm has been proved to be an effective solution. However, most of the current algorithms directly use the binary trust relationship of the social network to improve the recommendation quality, without considering the difference in the trust strength of the user for each friend. In order to improve the accuracy of social recommendation algorithm, this paper calculates personal reliability and mutual identify reliability based on social data, and quantifies the social attention matrix based on mutual identify reliability and alleviates the data sparsity problem based on the idea of score matrix pre-filling. The experiment and analysis results based on the real data set of public comments show that the new collaborative filtering recommendation model and algorithm proposed in this paper further improve the recommendation accuracy.