Traffic Flow Forecast Of Scenic Spots And Research On Tourist Itinerary Planning Technology

Self service travel has become a popular way of tourism,tourism platform has produced a variety of tourism data: travel notes,comments,pictures and so on.At the same time,the real-time GPS data generated by the portable positioning equipment also supplement the tourism data.These multi-source heterogeneous tourism data contain tourists’ travel time-space information and behavior information,which provides rich data support for understanding the real-time popularity of scenic spots and travel schedule.This paper focuses on the key issues of tourism recommendation: scenic spot flow prediction and tourism itinerary planning.In order to better arrange travel itinerary for users,this paper proposes a scenic spot traffic flow prediction algorithm based on deep learning and a travel planning method based on time-space prediction and time frame.The main research work of this paper is as follows:(1)A traffic flow prediction algorithm based on multi-source heterogeneous data and Multi-Graph Convolution Network Gate Recurrent Unit(M-GCNGRU)is proposed.In this method,the effective features of multi-source heterogeneous tourism data are fused by constructing multiple topological maps,and the spatial features of scenic spots are captured by Graph Convolution Network(GCN)and the temporal features of traffic flow are captured by Gate Recurrent Unit(GRU),so as to predict the future traffic flow of scenic spots.The experimental results on real data sets show that the proposed method is superior to other comparison algorithms in terms of the measurement index of the prediction algorithm,especially in the medium and long-term prediction.(2)This paper proposes a method of Reconstructing Historical itinerary based on travel notes.Firstly,the scenic spots in travel notes are automatically identified according to the text similarity,and then a trip reconstruction probability model is constructed based on the first-order Markov,prior knowledge and the spatial distribution of scenic spots,and then the historical trips of tourists are extracted from travel notes.Through the experimental verification on real data sets,this method is more accurate and comprehensive than other methods to identify the command entity of scenic spots,and more accurately restore the real travel itinerary of tourists.(3)This paper proposes a travel planning method based on time-space prediction and time frame.This method divides the time period that can be used for planning through the time frame.On the premise of meeting the constraints of users and maximizing the tourism experience,and taking into account the important factors such as the opening hours of scenic spots,popularity,travel time,tickets,travel time,travel expenses and traffic flow,a tourism itinerary planning model is constructed.Experiments on real data sets show that this method can save travel time and cost more than other travel planning methods.To sum up,the proposed scenic spot traffic flow prediction method effectively integrates multiple features of multi-source heterogeneous tourism data;the proposed travel planning method comprehensively considers the historical static factors and real-time dynamic factors that affect the tourism experience,and has very important use scenarios and commercial value in the tourism industry.