Improved collaborative filtering recommendation based on classification and user trust

When dealing with the ratings from users,traditional collaborative filtering algorithms do notconsider the credibility of rating data, which affects theaccuracy of similarity. To address this issue, the paperproposes an improved algorithm based on classificationand user trust. It firstly classifies all the ratings by thecategories of items. And then, for each category, itevaluates the trustworthy degree of each user on thecategory and imposes the degree on the ratings of theuser. Finally, the algorithm explores the similaritiesbetween users, finds the nearest neighbors, and makesrecommendations within each category. Simulationsshow that the improved algorithm outperforms thetraditional collaborative filtering algorithms andenhances the accuracy of recommendation.