Research Of Tourism Products Recommending Method Based On The Social Network Analysis

With the technology of social network analysis we built a tourists social network, and dig out the local communities in it, then combined with the traditional collaborative filtering algorithm. This method can solve the problem of data sparsity effectively, on the other hand recommend their favorite tourism products to a group of users with social relations can reduce the recommended blindness, improve the recommendation accuracy, and the ability to provide users with more personalized services to enhance the travel experience.To solve the problems of the tourists social network building and tourists relationship classification, an approach to constructing tourists social network which based on real tourists records is presented, and a local community mining algorithm was proposed which based on center nodes expansion. This algorithm modified the PageRank algorithm for nodes ranking, and then research the method of local community mining which based on center nodes expansion. In the tourism products recommending method based on the social network analysis we calculate the direct and indirect trust value between users within the same local community, then improve the traditional collaborative filtering algorithm by cooperated with the trust value of the user into the user similarity computing.In the experiments, we use a real tourists records set for loading, conversion and noise removal. Experiments showed that the approach to constructing tourists social network which based on real tourists records could dig out the hidden local communities in the tourist social network effectively, and have a smaller time complexity. In the comparative experiments of recommendation system we use the mean absolute error MAE and accuracy as comparison parameters, the result showed that his paper's method compared with the traditional collaborative filtering algorithm, the MAE reduced 0.021 and the accuracy rate increased 2.5%.