A Method for Mining Granger Causality Relationship on Atmospheric Visibility

• DOI: 10.1145/3447681
• Publication Date: 2021-05-30T01:59:11Z
• Authors: Bo Liu; Xi He; Mingdong Song; Jiangqiang Li; Guangzhi Qu; Jianlei Lang; Rentao Gu
• Keywords: Visibility; Gradient boosting; Haze; Endogeneity; Visibility graph

Atmospheric visibility is an indicator of atmospheric transparency and its range directly reflects the quality environment. With acceleration industrialization urbanization, natural environment has suffered some damages. In recent decades, level shows overall downward trend. A decrease in will lead to a higher frequency haze, which seriously affect people's normal life, also have significant negative economic impact. The causal relationship mining can reveal potential relation between other influencing factors, very important environmental management, air pollution control haze control. However, causality based on statistical methods traditional machine learning techniques usually achieve qualitative results that are hard measure degree accurately. This article proposed seq2seq-LSTM Granger analysis method for factors. experimental part, by comparing with such as linear regression, random forest, gradient boosting decision tree, light machine, extreme boosting, it turns out prediction accuracy model about 10% than methods. Therefore, this deeply implicit them provide theoretical support