ARIMA MODEL FOR FORECASTING OF AREA, PRODUCTION AND PRODUCTIVITY OF OILSEEDS IN INDIA

Abstract

The study conducted in the year 2022-23 applies the ARIMA (Autoregressive Integrated Moving Average) model to forecast the area, production, and yield of oilseeds in India from 2022- 23 to 2031-32, based on historical data from 1951 to 2021. The data was tested for stationarity and made stationary through differencing. The ARIMA model was selected using RStudio’s Auto ARIMA function, which identified ARIMA (0,1,1) for yield and output, and ARIMA (0,1,0) for area. Diagnostic tests, including the Box-Pierce, Shapiro-Wilk, and White Neural Network tests, confirmed the model’s accuracy, with no significant autocorrelation and homoscedasticity in the residuals, though output data showed non-normality. The model forecasted a gradual increase in area (from 291 lakh hectares in 2022-23 to 314 lakh hectares in 2031-32), production (from 345 lakh tonnes to 383 lakh tonnes), and yield (from 1268 kg/ha to 1447 kg/ha), reflecting improvements in agricultural practices. While the ARIMA model is effective for forecasting, the non-normality in output suggests an area for further refinement in future studies