Research And Implementation Of User's Preference Learning System Based On Route Selection Behavior

In the daily transportation,the preferences of users'travel and routing will directly affect the state change of the transportation,which in turn affects the control decisions of the intelligent transportation system.At the same time,changes in the transportation will also affect the user's routing decisions.Therefore,the learning of user routing preferences plays an important role in improving the accuracy and effectiveness of traffic scheduling and travel planning.At present,the traffic data contains a large amount of travel trajectory information.By mining and analyzing these data,the user's travel preferences can provide a basis for vehicle networking services and intelligent traffic control.Aiming at the difficulty of target user behavior pattern recognition caused by the uneven distribution of time and space of vehicle trajectory data,this paper proposes a routing preference generation method based on Generatie Adversarial Networks,using traffic situation data and user historical behavior data,through deep learning.The network extracts the feature map to mine the behavior characteristics of the target user,and fuses the macro situation feature with the micro user behavior feature,and combines the generation of the anti-learning algorithm based on the maximum information entropy to improve the robustness of the model.Finally,through the situation prediction method based on routing aggregation,the factors of user routing preference are integrated into the prediction of traffic situation to improve the performance of situation prediction.On the basis of the user-based behavior preference learning method based on generation confrontation,combined with the functional requirements of the routing preference learning system,the paper analyzes the requirements in detail,and then builds the overall framework of the system according to the requirements analysis.The static functions and dynamic interactions of each module are designed in detail,and the key implementation process descriptions are given.Finally,the function implementation of the system was tested to verify the effectiveness of the system.