Forecasting of PM10 Concentrations in Indian Medium-Sized City Using New Combined Grey Model

Indian medium-sized city, Kolhapur, has witnessed consistent growth in ambient PM10 concentrations over the last few years, and an accurate forecast is essential. However, for any forecast model, its prediction accuracy and capacity are prerequisites. Accordingly, a combined grey model was developed by combining the grey model [GM (1,1)] and the simple exponential smoothing (SES) model. The new model grey simple exponential smoothing model, abbreviated as GSESM, is newly utilized for PM10 concentrations. To demonstrate the applicability of the proposed GSESM, the GM (1,1) and SES models are also introduced. The PM10 concentrations data from 2008 to 2019 and 2020 were used for in-sample and out-of-sample calculations, respectively. With the relatively lowest mean average percentage error (MAPE = 2.89%) value in the out-of-sample case and reasonable development coefficient (–a = 0.0068) value, the GSESM proves it is not only highly accurate but also it can be used for mid-long term forecasts. Finally, the GSESM was utilized to forecast PM10 concentrations (2021–2025).