Research On Key Techniques Of Spatio-temporal Continuity Intelligent Recommendation Of Tourism Information

With the rapid development of the tourism industry and the Internet industry,more and more people share their travel experiences in social networks,including many geo-tagged photos.These geo-tagged photos contain a wealth of spatio-temporal data that provide an important database for people to analyze and mine visitor behavior and provide travel information recommendations.At present,in the relevant research on the recommendation of tourism information based on spatio-temporal data,the time is usually classified into a time period,and the position is mapped to the city,and the appropriate recommendation is given according to the city,time period and so on.This recommendation method is not sensitive to the user's current time and position changes,and is suitable for helping the user to develop a travel plan in advance,and is not suitable for providing real-time travel information recommendation.In order to meet the demand of real-time tourism information,we proposes a spatio-temporal continuity personalized recommendation algorithm for tourism attractions based on the attractions data in Ctrip tourism community and the geo-tagged photos in Flickr.It mainly includes the following research work:(1)Spatio-temporal data processing: It's hard to get the name and other semantic information of the attractions by getting attractions using clustering method on geo-tagged photos,and it's hard to get the hierarchical relationship between attractions too.To avoid these problems,we build an attractions information tree by crawling attractions information from the tourist community,and using the tree to match the attractions for geo-tagged photos.The attractions information tree can represent the hierarchical relationship between attractions effectively,and also can be used to filter the recommended attractions based on the location of user.(2)Spatio-temporal continuity personalized recommendation algorithm for tourism attractions: We presents a spatio-temporal continuity personalized recommendation method for tourism attractions based on distance-weighted PageRank.The recommendation is based on the the travel history of similar users of the target user in the similar time and the distance between target user and attractions.The similar user corresponds to the personalized recommendation,the similarity time corresponds to thecontinuity of time,the distance corresponds to the spatial continuity.In this paper,we use the type tags of attractions in user travel history to get the user characteristics,and use clustering method to get similar users.We use the time sequence to build time groups,and use the groups to get the travel history of similarity time.By adding a distance weight to the PageRank model,we achieve spatial continuity recommendation.The test results show that the spatio-temporal continuity personalized recommendation algorithm for tourism attractions based on distance-weighted PageRank have a higher recommendation accuracy than other algorithms.