A Collaborative Filtering Algorithm of Selecting Neighbors Based on User Profiles and Target Item

Without considering the difference in user profiles and user rated items, traditional User-Based collaborative filtering recommendation algorithm only considers the users' score on the item when calculates the similarity between users. In order to get rid of disadvantages of traditional methods, this paper proposes a collaborative filtering algorithm of selecting neighbors based on user profiles and target item. Aiming at obtaining target user's neighbors more suitable, this paper uses a weighting coefficient to adjust the final similarity which is influences by user profiles' similarity and users' rating similarity. In the case of user's neighbor didn't rate the target item, the expanded neighbors are considered, finally predicting and recommending target items. The experimental results show that the algorithm improves the accuracy of similarity, and effectively alleviates the user rating data sparseness problem, while improving the accuracy of the prediction.