부산대학교 도서관

With the continuous improvement of global economic level, the increasingly serious light pollution has gradually become the focus of discussion. This paper focuses on the influence of various factors from different aspects on the risk level of light pollution, and establishes a complete and systematic comprehensive evaluation model to provide a basis for rating and preventing the risk of light pollution in a certain area. This paper uses hierarchical analysis to calculate the weights of relevant factors affecting light pollution, and the main influencing factors of light pollution can be obtained by comparing the magnitude of the weights. A comprehensive evaluation model of light pollution severity can be calculated by using principal component analysis, and its reliability can be verified by conducting sensitivity analysis on the model. The spectral clustering algorithm can be used to calculate the grading evaluation criteria for areas with different light pollution risk levels. This paper determines the correlation between each influencing factor by calculating the Spearman correlation coefficient, divides them into three groups, and proposes three intervention measures. Finally, a gray prediction model is used to predict the values of the relevant influencing factors of light pollution in the next five years, and a comprehensive evaluation model is used to calculate the light pollution risk grade evaluation values without and with three interventions, respectively, to obtain the changes of the evaluation values in five years.