A Collaborative Filtering Algorithm Combined with User Rating Credibility and Similarity
Liu Sheng-zon · Journal of Chinese Computer Systems · 2014
Traditional recommendation algorithm based on trust is susceptible to the impact of differences in users' rating preferences, especially,the accuracy of the algorithm is volatile in the presence of deception score data. To solve these problems,this paper proposes a collaborative filtering recommendation algorithm combined with user rating credibility and similarity. This algorithm combined user ratings accuracy,recognition and ratings count weighting factors,and analyze the impact of these factors to the user ratings credibility,then establish a calculation relations between user ratings credibility and these factors. This article carries comparative experiment with two kinds of data set,one is in the presence of deception data and the other is not. In the absence of deception data. The experimental results show that this algorithm improves the recommend accuracy in the absence of deception data,and improves the recommend accuracy and robustness in the presence of deception data.