Multisource Information Fusion Algorithm for Personalized Tourism Destination Recommendation

In this paper, the existing scenic spot recommendation algorithms ignore implicit trust and transmission of users when dealing with user relationships, lack historical browsing behavior data in new city scenes leads to an inaccurate recommendation. a personalized method combining relationship tag preference is proposed. Firstly, degree introduced quality poor only considering similarity users. By mining users, problem that research cannot make recommendations direct difficult obtain solved, sparsity cold start problems are effectively alleviated. Secondly, process interest analysis, between spots tags extended among tags, users’ preferences decomposed into long-term for different which alleviates tour records lacking. The tourism proposed paper integrates many features social networks sparseness feature learning based on by using vectorization deep technology. Its has very important usage scenarios commercial value industry. This model can efficiently mine association rules multisource information data. experimental results show correlation selected tourists provide effective decision-making.