Collaborative Filtering Algorithm Based on Fuzzy Equivalence and Cause-Effect Clustering

Traditional collaborative filtering algorithms recommended has brought fundamental changes in electronic retailing. But as the recommended precision and personalized demand of the users are higher and higher, the algorithm to project the drawbacks of the irrational evaluation and sparsity problems, has seriously affected the accuracy of recommendation. Based on these problems, this paper puts forward a kind of collaborative filtering recommendation algorithm based on fuzzy equivalence relation and causal clustering, an algorithm based on user - project ratings, archived, characteristic value of a matrix to calculate the similarity between them. And the establishment of the matrix is based on the project, the user has similar transitivity and causal relationship; And then through the transitive closure method a project is calculated based on fuzzy equivalence matrix optimal threshold value method to measure the similarity and construct weighted feature method based on project, user triangular fuzzy number to calculate the similarity, so as to realize precise and personalized recommendation. The experimental results show that the algorithm can improve the accuracy of the data and evaluation, improve the accuracy and personalization of the proposed algorithm.