Data Mining Analysis And Optimization Recommendation Modeling Of Vehicle And Cargo Matching

Road transportation is the most important mode of cargo transportation.With the sustained and fast growth of China’s road freight business,the supply of cargo and the number of operating vehicles are increasing.In road transportation,there has been a problem of finding goods for car and finding cars for goods,with the development of mobile Internet,to solve the problem of carrier from offline information transferred to the online platform for the car goods match,but the essence of the two sides of supply and demand information asymmetry in car problem not been solved,resulting in the low degree of standardization,car goods supply and demand information matching efficiency low and recommend the lack of stability and precision.Based on the background of highway freight logistics,this paper first defines the concept of vehicle and goods matching platform,focuses on three models of vehicle and goods matching,and sorts out the operation process of vehicle and goods matching on the platform.Then,on the basis of literature and real vehicle and cargo matching platform data,a two-level index system of vehicle source screening and cargo source screening is extracted: information screening index and matching effect index.The information screening index is used to select qualified matching objects,and the matching effect index is used to rank qualified matching objects.On index weight assignment method,innovative use of data mining and machine learning models such as Light GBM for real load of goods matching platform-a full data mining analysis,through the data cleaning and characteristics of the engineering characteristics of options and the supply of goods for processing,finally in guarantee the training model of the effect of the precision rate and recall rate reach 90%,The importance degree of each feature is obtained,that is,the index weight.Considering the single source of single sorting based on matching options,further multiple sources and overall consideration and same point matching options,which not only consider one of the options or maximize individual match supply party,but consider the same and end point of multiple options and multiple source of the maximum matching degree,cars were constructed based on matching recommended model for this problem,In addition,the combined optimization results of vehicle and cargo matching considering the overall matching degree were obtained by using Yunman platform data and Python programming solution,and the effectiveness of the matching recommendation model was verified by comparing the matching results without considering the overall matching degree.