Recommendation Based on Collaborative Filtering by Convolution Deep Learning Model Based on Label Weight Nearest Neighbor

Sparseness of user-to-item rating data is one of the major factors that deteriorate the quality of the recommender system. To handle the sparsity problem, this paper introduces the parameters such as tag information and user information, and improves the recommendation precision by obtaining the nearest neighbor set that has the greatest impact on the target user. A convolution deep learning model based on label weight nearest neighbors (LWNCDL) is proposed. First, the nearest neighbor set that has the greatest impact on the target user is found by similarity. Then, the item feature vector is studied by convolution neural network. Finally, the prediction matrix is decomposed by the probability matrix. The results show that the proposed method is more efficient for recommendation accuracy compared with the traditional recommendation algorithm.