A Hybrid Time Series Model for the Spatio-Temporal Analysis of Air Pollution Prediction Based on  \(\textrm{PM}_{2.5}\)

The issue of air pollution represents a formidable challenge for the global community. This challenge is further compounded by the increasing levels of air pollution witnessed daily. The base pollutants which are \(\textrm{PM}_{2.5}\) , \(\textrm{PM}_{10}\) , \(\textrm{NO}_{2}\) , \(\textrm{SO}_{2}\) , NO, CO, and \(\textrm{O}_{3}\) . \(\textrm{PM}_{2.5}\) air pollution has been associated with various health complications, including respiratory diseases, heart diseases, cancer, and premature death. This attention has been brought about by the prevalence of air pollutants and their impact on human health. The author has proposed a solution by utilizing a hybrid model incorporating spatiotemporal correlation analysis. Specifically, the author suggested a model that merges CNN and LSTM. The outcomes of this model have been remarkable, with an enhancement of 64.26% (MSLR), 67.75% (MSE), 24.46% (MAPE), 35.20% (MAE), and 29.52% (RMSE) when compared to models that lack spatiotemporal analysis. Recently, there has been growing attention towards the detrimental effects of air pollution from both the government and the public. Despite the numerous efforts to curb air pollution, it continues to pose a significant threat to human health and the environment. Therefore, there is a need to continue exploring innovative solutions, such as the proposed hybrid model, to mitigate the impact of air pollution.