RESUMO
BACKGROUND: Air pollution is a significant issue for a developing country like India and the air quality index (AQI) forecasting helps to predict air quality levels in advance and allows individuals to take precautionary measures to protect their health. OBJECTIVES: The study aimed to forecast the AQI for an industrial area (SIDCUL, Haridwar City) using a time series regression model. MATERIALS AND METHODS: Three years of existing AQI data points (post-COVID-19) were collected from the Uttarakhand Pollution Control Board for the SIDCUL area of Haridwar City and tried to know the status of AQI values for the following 12 months. Trend and seasonality components were seen through the decomposition process. Further, the augmented Dickey-Fuller test was applied to check the stationarity of the series before finalizing the best-suited time series model for forecasting the AQI values. RESULTS: With the help of autocorrelation function (ACF)/partial ACF plots, a seasonal autoregressive integrated moving average (ARIMA) (0,1,0) (1,0,0)[12] model was selected with the minimum akaike information criterion (253.143) and mean absolute percentage error (17.42%). The AQI values have also been forecasted for this industrial area (SIDCUL) for the following year. CONCLUSION: The seasonal ARIMA (0,1,0) (1,0,0)[12] model may be helpful to forecast the AQI values for a nonstationary time series dataset. Research indicates that the air of the SIDCUL area will become moderately polluted and may cause breathing discomfort to asthma patients' health. The scientists might apply this model to other polluted regions of the country so that the public and the government can take preventive measures in advance.