A Research On Personalized Tourism Information Intelligent Push Method Based On User Interest Modeling

With the explosive growth of the tourism information and the complexity and diversity of tourism demand,how to conduct filtering and disorderly information,and the users are most concerned and most interesting information show,has become one of the most important challenges of the era of information explosion.Based on this,the personalized recommendation system were introduced,the system is different from the conventional search engine users need active search keywords to match information,it is to get the user current preference through the analysis of user interest and initiative to provide interesting information for users,which not only improve the user experience,but also a corresponding increase in business return.Recommendation system from the proposed to the application,especially the arrival of the era of intelligent tourism,visitors need to solve some of the problems.Intelligent Tourism personalized push services,mainly for the comprehensive analysis of the basic information of the user,behavior preference information and situational features three,build up interest model for each user,and then push the corresponding personalized service.Although the personalized recommendation service development has made some progress,but still faces many challenges,for example,the current recommendation system modeling technology is not comprehensive and accurate description of user preferences,but can not integrate elements of user scenarios,to push the contents of the recommendation algorithm is not accurate enough,the practical application need not adapt to the user.The pressure on the above problems,this paper focuses on the combination of user basic attributes and domain knowledge,recommendation system modeling problems of personalized travel services;and expand the user neighbor selection strategy to improve the diversity of recommendations and novel problems.This paper studies from three aspects,the main work is as follows:(1)This paper studies the method of user modeling in personalized recommendation system,and proposes a user modeling method based on extended vector space model.The use of the user's demographic information,the user's interaction with the goods,while filtering the redundant information,more accurately describe the user preferences,and the cold start of the system has improved.(2)This paper puts forward a method for modeling the tourist interest when the context is abundant,and designs a personalized recommendation algorithm.Based on the analysis ofthe tourist behavior data,this paper makes a detailed analysis of the tourists' behavior data,and finds out the spatial and temporal correlation between the tourist attractions.Based on this discovery,presents the integration of contextual information tourist interest representation will have similar travel preferences(such as the tourist season and place)of tourists is mapped to the hidden space are similar to that of comparative interest and visitors can realize the package contents.Based on the interest model and package price constraints,the paper designs and implements the recommendation algorithm of the tourism package,which can be used as a personalized travel package recommendation service for tourists.(3)The user preference model combining diversity and multidimensional information,by example,to build a personalized travel recommendation,through data acquisition,data analysis and processing,user modeling,recommendation,recommendation of diversification of services for tourists.Finally,according to the algorithm proposed in this paper is the design of a tourist attractions recommendation system,this system can be combined with the user basic information and data,to provide personalized recommendation service for users,and solve the problem of the cold start recommendation.