Prediction of Air Pollution using Deep Learning based LSTM and CNN Model

Abstract

Due to its detrimental impact on all creatures, air pollution is one of the most significant environmental issues in the industrialised world. Multiple authorities warn that severe air pollution occurs in many places of the globe. Taking into account the harmful impacts of air pollutants, it is essential to develop accurate models for predicting air pollution levels in order to calculate future concentrations or pinpoint pollutant sources. These models may assist governments as well as central authorities with policy implications for preventing excessive pollution levels. Even though there have been several efforts to estimate pollution levels in the past, new improvements in deep learning methods and the incorporation of additional data promise more accurate prediction outcomes. This study proposes research on the application of deep learning architectures to the UCI machine learning-compiled Beijing PM2.5 dataset. We have suggested a CNN and LSTM model with deep learning capabilities. Root Mean Square Error (RMSE) as well as Mean Absolute Error (MAE) were used to assess the performance of these models (MAE). Experimental findings reveal that our technique “CNN-LSTM” provides more precise predictions than the classic models stated and has superior predictive performance.