Research And Application Of Tourism Service Recommendation Based On User Influence In Social Network And Time Series

With the widespread use of mobile smart devices and the continuous improvement of mobile communication technologies,in the mobile network,users can share their trip stories at any time and place,which greatly enhances the user's enthusiasm for commenting and increased the immediacy of user comments.The massive user data can provides an opportunity to mining users' travel preferences.Tourism service recommendation not only can help users identify favorite scenic spot,but also can help tourism service providers acquire more target customers and increase business revenue.Therefore,both academia and industry are constantly adopting more advanced and effective methods to improve the efficiency of the tourism recommendation system.Tourism recommendation has been studied by lots of researchers,but most of the existing research methods without considering the complexity of travel,and ignoring the influence of various factors on user decision-making.A large number of scenic spots,especially those with natural scenery,only have a few months of every year are best to visit.In addition,compared with watching movies,buying books,users travel less.The existing tourism recommendation method does not take into account the seasonality of tourism service and the user's travel rules.In addition,not take full advantage of multi-dimensional social information in social network also leads to the poor efficiency of tourism recommendation system.In order to solve these problems,this paper propose a personalized tourism service recommendation methods fusing multiple factors,the main work is listed as follows:(1)We analyze the research background,state-of-the-art,related theories of tourism service recommendation and their limitations.(2)Aiming at the seasonal fluctuation of tourism service popularity and score,this paper combined genetic algorithm(GA)and deep learning algorithm(LSTM network),and an algorithm for predicting the popularity and score of tourism service is proposed and applied to the real data set,the experimental results show that the GA-LSTM network performs well.(3)Integrating the predicted popularity and predicted scores of the tourism service,and calculating the similarity score between the tourism service popularity based on the predicted score and the user travel law(GA-LSTM_Rec).Considering the multi-dimensional social information of users in social networks,this paper proposed the Explicit Social Influence(ExSoInf)and the Implicit Social Influence(ImSoInf)of users,and they are integrated into SVD algorithm(SoInf_SVD).Calculating user preferences for tourism service categories using matrix factorization technique(CaMF).And than,the calculation results of the three components are weighted and summed to obtain a fusion tourism service recommendation algorithm(GA-LSTM_CSInf).Finally,we conduct a comprehensive performance evaluation for our fusion recommendation algorithm using real-world data sets collected from Yelp.And the result shows that our method has significant improvement in recommendation quality compared to other advanced recommendation methods.(4)According to the general form of the current tourism service recommendation APP,the tourism service recommendation algorithm proposed in this paper is integrated into the online travel application,and a MyTravel tourism service recommendation application prototype is designed and implemented.MyTravel application prototype provides rating and reviewing,personalized recommendation and other functions.