Research On Scenic Spot Passenger Flow Analysis And Prediction System Based On Deep Learning

With the vigorous promotion of smart city construction by national policies,smart tourism has ushered in unprecedented development opportunities.At present,deep learning technology and intelligent big data system have been widely used in the fields of meteorology,transportation,energy and so on.More and more cities in China have begun to pay attention to the information construction of scenic spots,providing the tourism industry with a large amount of data with high timeliness.But at the same time,the increasing number of tourists has also brought a series of challenges to the development of the tourism industry.The passenger flow and resource allocation of tourist attractions have gradually appeared the imbalance of time and space,which has a negative impact on the management of scenic spots and the travel experience of tourists.Therefore,this thesis provides passenger flow analysis and long-term and fine passenger flow forecast services for all scenic spots in Shanghai and Zhejiang Province,and visually displays the results,so that tourists can plan their travel itinerary in advance according to the long-term and fine forecast results,so as to achieve the effect of staggering peak travel.Based on big data processing,deep learning,visualization and other technologies,this thesis studies and develops the scenic spot passenger flow analysis and prediction system.The main research contents and work are as follows:1.In view of the poor timeliness and single information of the existing scenic spot passenger flow data sets,according to the actual use of the system,the scenic spot data of Shanghai and Zhejiang Province are continuously collected to establish the scenic spot passenger flow data sets.At the same time,the scenic spot big data streaming cloud computing platform is built,and the Structured Streaming real-time streaming computing engine is used to complete the real-time big data analysis of all scenic spots in the province,and the scenic spot multi-dimensional analysis indicators are obtained.2.In view of the lack of application of LSTF(Long Sequence Time-series Forecasting)in scenic spot passenger flow forecasting,a long time series passenger flow forecasting method based on self-attention mechanism is proposed.Based on the transformer architecture,this method constructs a GRN-Informer long-time series passenger flow prediction model to predict the future passenger flow of scenic spots in provinces in a long series and low time granularity.The experimental results show that the passenger flow forecasting model is better than other forecasting models,and the error of passenger flow forecasting results increases slowly with the length of forecasting.3.A tourist flow analysis and prediction system based on deep learning is designed and implemented,and the analysis and prediction results are visualized interactively in real time.The system mainly includes: user login and authorization,data collection,big data real-time flow processing,scenic spot passenger flow forecast,data visualization and other functions.Design according to the system requirements to achieve various functions,and test the system from two aspects of function and performance to verify that the realization of the system meets the expectations.