Air pollution is caused due to surplus growth of dangerous substances in atmosphere such as gases emitted from vehicles and biochemical molecules. Continuous exposure to such gases and intake of such a substance by human causes a serious hazardous health issue like respiratory disorders, heart failure and so on. The air is heavily polluted in urban cities mainly in metropolitan cities, Particulate Matter (PM2.5) and PM10 which are minute particles which in size of micrometre can cause serious threat to human health. Therefore, it is mandatory to monitor the air quality on hour/daily basics for the living environment. In this Paper we have designed and developed a novel effective model for air quality prediction named as Competitive Swarm Political Rider Optimizer (CSPRO) based on Nonlinear Auto-Regressive Exogenous (NARX) model. Firstly, pre-processing of data is carried out by missing value imputation and then the technical indicators are extracted for prediction process. Air quality prediction is carried out using NARX model is trained by CSPRO. The proposed CSPRO based NARX has attained low Mean Absolute Prediction Error (MAPE) and minimum Mean Squared Error (MSE) of 9.22 and 0.275 respectively.

Air quality prediction, Non-linear Auto-Regressive exogenous (NARX), Rider Optimization, Political Optimizer, Relative Strength Index.