In order to alleviate the sparsity problem of the collaborative filtering (CF) and improve the accuracy of the recommender system, a restaurant recommender system (RRS) based on the improved CF is proposed by analyzing the tags information in dianping.com. Firstly, the mean centering is employed to preprocess the rating in order to reduce the error caused by different personal habits. Secondly, according to the feature of the restaurant and the user rating, an improved null filling method is

more »

... posed to alleviate the sparsity of the user-restaurant matrix. And then, when calculating the user weight, the similarity of the user preference and the user trust are considered, which can improve the accuracy of the recommender system. Finally, the experiment is executed in the real data set and the experimental results show that the proposed method can obtain higher accuracy than the existing methods.