Technical Research on Missing Value Imputation for Atmospheric Quality Data Based on Improved GAIN

With the rapid development of China's economy, environmental issues have also become a key focus of attention. Environmental prediction can promptly detect potential pollution hazards and nip them in the bud. However, due to the interference of uncontrollable factors during the collection and storage of environmental data, data loss occurs. Addressing the issue of missing environmental data, this article makes various improvements to the existing GAIN filling network model. Finally, an improved GAIN network model is adopted, which uses dual generators to iteratively capture features and adds an attention module to obtain potential features. Compared with traditional filling models, the improved GAIN network model exhibits better performance in filling missing environmental data.