Research On Vehicle Cargo Matching Of Network Freight Platform Based On Intelligent Recommendation

Highway transportation is a very important part of China’s freight logistics.In recent years,its business demand has shown a rising trend,and highway freight transportation presents the characteristics of scattered geographical sources of vehicles and goods,resulting in the widespread situation of difficult to find goods and goods.Therefore,the network freight platform specially matching the supply and transportation capacity came into being,and the matching ability has become the most important competitiveness of each network freight platform.Taking LHB network freight platform as the background,this paper studies the vehicle cargo matching problem.Through the analysis of the current situation of vehicle cargo matching on the platform,it is found that there are irregularities in the process of vehicle cargo matching,which are mainly reflected in two aspects: first,LHB network freight platform adopts manual matching mode and does not recommend intelligent vehicle cargo matching.Second,the "acquaintance mode" leads to a narrow selection range of vehicle sources and goods sources,and the matching of uncertain and unstable manual experience,which jointly leads to the low efficiency of vehicle and goods matching on the platform,which is not conducive to the development of online freight platform.Based on the problems of LHB network freight platform,this paper constructs a vehicle cargo matching recommendation model of network freight platform based on deep learning.The model is divided into three stages: in the first stage,a vehicle cargo matching recommendation model based on back propagation(BP)neural network is constructed.BP neural network is used to extract the features of the owner and owner information input by the recommendation system.Compared with the traditional recommendation algorithm,it can better learn the data features.At the same time,using the advantages of neural network in fitting nonlinear data,a regression prediction model based on BP neural network is constructed to replace the traditional recommendation model,which multiplies the vehicle owner’s characteristics and cargo owner’s characteristics by vectors,so as to better fit the scoring data.In the second stage,because convolutional neural networks(CNN)is conducive to extracting more complex deep-seated features and makes up for the feature information that BP neural network cannot extract to a certain extent,a vehicle cargo matching recommendation model of BP neural network convolutional neural network(bp-cnn)with double tower structure is designed,BP neural network and one-dimensional convolutional neural network(1-dcnn)are used to extract the input data of the recommendation system at the same time to form a double tower extraction structure.The third stage: Based on the consideration of the practicability of the model,design the recommendation display page to intuitively reflect the recommendation results and facilitate the operation of the platform business personnel.Finally,using the relevant data obtained from LHB network freight platform,the vehicle cargo matching recommendation model of network freight platform based on deep learning is trained,analyzed and optimized.