Study On Vehicle Inspection Of Expressway Green Channel Based On Logistic Regression

Due to the imperfect management of expressway green channel vehicle and the temptation of economic interests,toll evasion situation of expressway green channel vehicle occur frequently.However,there are many problems such as low inspection efficiency and cumbersome workflow in the current inspection method of expressway green channel vehicle inspection method.How to solve these problems has become an important issue of highway management.Based on the analysis of expressway green channel vehicle toll evasion behaviors and evasion situation,puts forward the inspection model of expressway green channel vehicle based on Logistic regression model.Firstly,expressway green channel vehicle toll evasion behaviors and evasion situation are analyzed.Study the causes and characteristics of these evasion behaviors,and detailed analysis of different distribution characteristics of fake green channel vehicle and qualified green channel vehicle in free toll,vehicle tonnage and other types of range.Then put forward the design principle and design goal of the green channel vehicle inspection model.Based on the Logistic regression model,designs the green channel vehicle inspection model,and detailed analysis of the workflow of this inspection model.Secondly,based on the characteristics of green channel vehicle inspection data set and inspection business requirements,establishment of a pretreatment method for green channel vehicle inspection data set,and compared with the results of each stage.The SMOTE algorithm is used as an over sampling approach to the green channel vehicle inspection data set.The ChiMerge algorithm is chosen to discretize the attributes of the green channel vehicle inspection data.Mineralization of green channel vehicle inspection data by Min-Max standardization.So that the green channel vehicle inspection data set is suitable for Logistic regression analysis.Then using K-S algorithm to test the normality of continuous value attributes of green channel vehicle inspection data.Using the independent t-test to test the correlation between the continuity values with normal distribution of green channel vehicle inspection data and the check results of green channel vehicle.Using the Mann-Whitney U test to test the correlation between the continuity values without normal distribution of green channel vehicle inspection data and the check results of green channel vehicle.Using the Chi-square test to test the correlation between the discrepancy values of green channel vehicle inspection data and the check results of green channel vehicle.The collinearity between attributes is analyzed comprehensively by tolerance,variance expansion factor and so on.Filter attribute that have verified collinearity relationships.Based on Logistic regression,establish the model of expressway green channel vehicle inspection with attributes of the correlation test and the collinearity diagnosis.And analysis the calculation process of attributes and interpretation methods of model resultFinally,examples of expressway green channel vehicle inspection model with the data of green channel vehicle for transporting livestock meat and poultry meat.Discretization attributes of green channel vehicle inspection data,include free toll,vehicle tonnage and travel distance.Combined with the results of the correlation test and collinearity diagnosis to establish the corresponding green channel vehicle inspection model.Using training data for model training,generate the corresponding regression equation,and using the test data for model test.The sensitivity of green channel vehicle for transporting livestock meat is 85.51%,and the specificity is 87.91%.The sensitivity of green channel vehicle for transporting poultry meat is 84.24%,and the specificity is 55.77%.The test results show that the green channel vehicle inspection model is effective,and the green channel vehicle inspection model of poultry meat has over-fitting problem.