Improved Collaborative Filtering Algorithm Based on Multi-dimensional Fusion Similarity
Xiaoxuan Liu · 2019
Collaborative filtering is the most commonly used technique in personalized recommendation system. In order to improve the accuracy of recommendation algorithm, the parameter is designed to improve the Pearson correlation coefficient after considering the user's interest differences in rating. In addition, the user fusion similarity is obtained by combining the user interest similarity calculated based on the user interest and the user trust similarity calculated based on the trust relationship between users. At last, the improved Pearson correlation coefficient and the user fusion similarity are weighted to obtain the multi-dimensional fusion similarity. Simulation which uses the MovieLens as dataset shows that the improved algorithm has better recommendation quality than traditional algorithms.