Analysis On The Impact Of China’s Economic Policy On The Environment

The purpose of this study is to analyze the impact of economic policy factors on environment of 31 provinces.The data collected from 18 aspects including the three major industries of economy,political and legal system and science and technology investment,including per capita GDP,the proportion of added value of secondary industry,urbanization,import and export volume,human capital and fiscal decentralization from 2003 to2016.Firstly,this thesis established SVM two classification and multi-classification models based on different penalty functions.The penalty functions considered include SCAD,elastic net and elastic SCAD penalty function.The study explored the environmental impact factors according to the variable selection results under different models.The conclusion is that the SVM two classification and multi-classification models under elastic SCAD penalty have the best prediction accuracy,in which 22 significant variables are screened by the two classification model and 15 significant variables are screened by the multi-classification model;Secondly,based on the least square method,the logistic regression method and elastic SCAD SVM method were combined in parallel to calculate the weight coefficient of the combined model.The combined model has better prediction accuracy in both two classification problems and multi-classification problems,the two classification combination model screened 12 significant variables,and the multi-classification combination model screened 7 significant variables;Finally,all models are integrated to screen out the following variables that have the most significant impact on the environment: per capita GDP,population density,human capital,fiscal decentralization,degree of marketization,etc.From the perspective of economic policy,this thesis explored the environmental impact factors of China’s economic policy.The elastic SCAD SVM method was applied to the statistical analysis of environmental impact factors,considering the two classification model and multi-classification model in detail.Based on the screening of active variables,the environmental quality of various provinces in China is predicted.Finally,combining elastic SCAD SVM model of logistic regression to further improve the prediction accuracy and the accuracy of variable selection.