Research On Tourist Attractions Recommendation Based On Frequent Sequence Mining And Global Relevance Graph

With the improvement of income level,tourism has become an important way to enrich spiritual life.Tourism has developed rapidly,and the personalized travel pattern comes into being.The establishment of personalized travel itinerary involving many factors such as user hobbies,popularity of tourist attractions,geographic location is complex and time and effort consuming.Travel websites and mobile applications include a large number of tourists' historical travel records,reflecting the characteristics of scenic spots and users' preferences for tourist attractions.The effective use of these data to recommend paths for tourists can greatly reduce the workload of tourists and enhance user experience.The traditional recommendation of tourist attractions,mainly focused on individual tourist attractions,ignore the context and sequence characteristics of tourists behavior,and can not recommend a series of successive sequence according to the user's current position.In addition,there are also shortcomings of historical data loss and statistical incomplete.This paper studies the tourist attractions recommendation based on frequent sequence mining and global relevance graph.1.We propose a successive travel sequence recommendation based on frequent sequence mining(SeqRem),which uses frequent sequence mining to construct a historical sequence relevance graph.We model the division of the existing sequence as the TOP-K maximum point weight of the relevance graph,and propose a heuristic calculation method.According to the existing travel schedule,we recommend a successive travel sequence for users.The results of the SeqRem show that the precision of the SeqRem method is better than LORE and FPMC-L method in the successive single spot and sequence recommendation.2.We propose a travel itinerary recommendation method based on the global attractions relevance graph.With the frequency of tourist attractions and migration,the method of calculating the experience income of the nodes and edges in the global relevance graph is constructed,and the path search algorithm based on the ant colony optimization is proposed to calanlate the highest K income for the user.The result of our travel itinerary recommendation shows that the experience income of the recommended path is proportional to the recommended accuracy rate in a certain range,and the effectiveness of the model is illustrated by the recommended precision rate.In addition,the travel itinerary recommendation in thiesis is better than PTIR method.