Meteorological AQI and pollutants concentration-based AQI predictor

Over the past few decades, rapid industrialization and urbanization have severely deteriorated urban air quality. Forecasting the value of air pollutants and predicting the Air Quality Index (AQI) are crucial in improving the control measures for tackling air pollution. The current research proposes meteorological AQI and pollutants concentration-based AQI predictor. It consists of the pollutant predictor module and Air Quality Index predictor module. In this work, (a) ARIMA, ANN, RFR, XGBoost, LSTM, VAR, and ARMA are used to forecast the concentration of pollutants, (b) support vector regressor (SVR) is then used to build an AQI prediction model using forecasted air pollution concentration (in part 'a') and meteorological factors such as wind speed, temperature, and humidity, and (c) the performance of SVR with different kernel functions has been investigated. Also, the impact of considering past AQI values for predicting future AQI values has been analyzed. The dataset used in this work has been acquired from the official website of the Central Pollution Control Board, Ministry of Environment, Forest and Climate Change, Government of India. The pollutant prediction models were evaluated using MAE and RMSE as the performance parameters. It has been observed that ARIMA and ANN can be used for efficient and accurate forecasting of the concentration of air pollutants. Further, it has been found that the proposed SVR model works effectively using linear kernel function (with one lookback value), with MAE being 29 and RMSE being 44.5. This proposed predictor may help people to plan their daily activities.