Research On Prediction Of Typical Vehicle Emission Factors Based On Data Mining

With the rapid development of economy and the continuous expansion of city scale,the exhaust emission from transportation has become an important source of air pollution.In order to adopt effective road network vehicle emission control strategy,reduce the pollution caused by traffic exhaust to the environment,and build a green traffic system,it is necessary to carry out scientific quantitative evaluation of vehicle emissions according to vehicle emission factors.However,most of the related research on the emission factor calculation scheme focuses on the field of transportation,but there is a blank in the research on using computer technology to calculate the emission factor.In this paper,the feasibility of machine learning algorithm in solving the problem of vehicle emission factor calculation and prediction in the field of transportation is studied.By using the relatively mature machine learning algorithm in data mining,the collected vehicle emission data is analyzed and modeled,and good prediction effect is finally achieved.The main contents and contributions of this paper are as follows:(1)By analyzing and modeling the collected motor vehicle emission data,this paper finally constructed seven prediction models to realize the prediction of motor vehicle emission factors,and gave the specific construction process and the effect evaluation of each model.Compared with the effect of the MOVES model,the feasibility of the machine learning algorithm in solving the calculation and prediction of vehicle emission factors is proved,which provides a new idea for the calculation and prediction of vehicle emission factors in the field of transportation.(2)This paper evaluates the effects of the seven prediction models built,and finally finds that the random forest regression model can achieve good results in solving the calculation and prediction of vehicle emission factors,and gives the specific parameters used in the construction of the random forest regression model in this data set.(3)In this paper the influence of vehicle emission factor is the most important five characteristics,namely,motor vehicle instantaneous velocity and instantaneous acceleration of motor vehicles,motor vehicle vehicle weight and engine exhaust of motor vehicles and motor vehicle fuel types,and gives the proportion to their relative importance,for the follow-up in the calculation and prediction research of vehicle emission factor provides the reference to adjust each feature weights.