Development of statistical, artificial neural network and hybrid models, forecasting wheat area and production in pakistan

Agriculture plays a vital role in the economic development of Pakistan, as it is the main largest source of economy to achieve the targets of economic growth. The objective of this research was to forecast the area and production of wheat crop taking historical data from 1947 to 2021 for Economic Survey, Ministry of Finance, Government of Pakistan. Initially a linear model ARIMA, through Box-Jenkin's (1976) methodology was applied to forecast the wheat area and production. Then we compare it with ETS, TBATS, Artificial Neural Network (ANN), and ARIMA-ETS, ARIMA-TBATS, and ARIMA-ANN hybrid models. It was observed that the ARIMA-ANN model show the lowest values for RMSE (214.2673, 921.9103) and MAE (175.470, 710.7325) for both wheat area and production. Forecasting of wheat area and production has been performed till 2030. Ten year average forecast for wheat area under the ARIMA-ANN was 9058.10 and yearly expected wheat area was increased by 2.87 percent each year. Change was computed and revealed that the area of the wheat crop is expected to increase by about 2.87 percent. The average wheat production for the ten-year forecast from the ARIMA-ANN model was 25878.03 and the yearly expected percentage change in the production of wheat was about 2.34 percent to increase. These forecast estimates for wheat crop will be important for the Government in formulating their policies to fulfil the food necessities of the nation, trade, support prices and planning the industrial sector.