Optimized Collaborative Filtering Algorithm Based on Item Rating Prediction

Abstract

Collaborative filtering recommendation algorithm is currently the most widely used personalized recommendation algorithm. Sparsity problem of user rating data led to the recommendation quality of traditional collaborative filtering algorithms are far from ideal. To solve the problem, the paper first cloud model and project characteristic attributes to calculate the similarity between the project has taken into consideration in computing project similarity scores were similar between the project and consider the project between the characteristic attribute similarity, and then to predict ungraded items rated. Finally, the cloud model to calculate the similarity between users to obtain the target user's nearest neighbor. Experimental results show that the algorithm improves the accuracy of the similarity of the calculated project, and effectively solve the problem of data sparsity, and improve the quality of the recommendation system recommended.