A study of light pollution risk level assessment based on multiple linear regression

Abstract

Light pollution is affecting modern people’s lives, and it may damage people’s health and safety. To develop an applicable metric to identify the light pollution risk level of a location, this paper first analyzes the correlation of factors affecting light pollution, and obtains that the correlation coefficient with light pollution mainly includes population density, per capita income, GDP, green coverage, the proportion of the tertiary industry and employment rate. Select these explanatory variables which have high linear correlation with the degree of light pollution to establish a multiple linear regression model. Then, carry out heteroscedasticity and multicollinearity tests on the model, and next, after removing the three variables of population density, per capita income and the proportion of tertiary industry, the optimal linear regression model is determined through OSL. Finally, different levels of light pollution are classified by DN value, which represents the night light of a location, to achieve the assessment of light pollution risk level.