Travel Recommendation Method And Application Based On Passenger Profile And Trip Chain Model

With the increase of China’s economic,people’s travel demand is growing rapidly,and China’s passenger volume of civil aviation has risen to the second place in the world.The intelligent service in airport landside transportation system is required.User profile system and recommendation system are effective methods to improve passenger experience and realize rapid development of civil aviation service.By studying the model of passenger profile and trip chain,a collaborative filtering travel recommendation method based on passenger profile clustering was proposed in this paper.This method effectively solves the sparsity of feature data set,quantifies passenger travel preference and provides personalized travel recommendation.A intelligent passenger travel service system was developed,and the travel recommendation method proposed in this paper was applied to the daily travel process.The system is able to provide personalized travel recommendation service and release integrated transport status information for passengers.The study has good effect on improving the travel efficiency of passengers in the airport landside traffic hub and ensuring the ability of passenger dispersion.The main contents are as follows:(1)The passenger feature data set was analyzed based on K-means clustering algorithm and combined with the characteristics of airport business.And the results of passenger clustering were optimized by using contour coefficient method and feature dimension reduction method.From the perspective of passenger personal attributes and travel preferences,the results of passenger clustering was analyzed And a passenger profile and label system was generated in this paper.(2)The temporal and spatial characteristics of passengers’ trip chain in the airport landside transport hub was researched,and a space-time double-layer travel chain model in the airport transport hub was constructed.Based on the graphical model theory and the real airport road network data,the spatial topology model of airport road network was constructed.Based on BPR model and queuing theory,a passenger travel time prediction method was proposed and the parameters were fitted with the collected passenger flow data.(3)According to the analysis of airport scenarios and passenger characteristics,a collaborative filtering recommendation algorithm based on passenger profile characteristics analysis was designed.The sparse matrix was supplemented by the similar passenger trip data,and the feature dimension was reduced by principal component analysis.The performance of the algorithm was tested by the passenger characteristics data set.And the effectiveness of the method was verified by comparing with other recommendation algorithms.(4)The intelligent travel service system was developed on B/S architecture.The system realized the functions of passenger profile generation,travel time calculation,personalized travel recommendation and traffic status information delivery.This study simulated virtual passengers and varied traffic scenarios to verify the ability of intelligent recommendation and the visualization of traffic state information in the system.It was proved that the research can improve the passenger travel efficiency and the ability of passenger dispersion.There are 47 pictures,31 tables and 75 references.