Using a Hybrid Model (EVDHM-ARIMA) to Forecast the Average Wheat Yield in Iraq

Purpose: Using an appropriate analytical method to deal with time series containing linear and nonlinear compounds and minimizing nonstationary in order to obtain good modeling and more reliable forecasts. Theoretical framework: Many methodologies have been developed in the past to perform time series forecasting, including those presented by the two scientists (Box and Jenkins) and known as (ARIMA) models (Box et al., 2015), It gives more reliable forecasts when analyzing time series for linear compounds, but they are less suitable when dealing with nonlinear compounds that characterize real-world problems. This causes an increase in forecasting error, so it has recently been demonstrated use of hybrid models that compounds linear and nonlinear the best forecast is obtained. Design/methodology/approach: Using the hybrid model (EVDHM-ARIMA) and comparing with single model (ARIMA) and comparing the two models using the (RMSE) criterion to forecast the average yield per dunum of the harvested area for the wheat crop in Iraq for the period (2024-2033), given the strategic importance of this crop in providing food security for the country. Findings: The showed, value of (RMSE) for the estimates obtained from the hybrid model (EVDHM-ARIMA), is the best for forecasting of the research data. Research, Practical & Social implications: We suggest generalizing the research idea for forecasting in different fields and comparing it with other forecasting methods. Originality/value: Adopting the forecasting values obtained from the proposed method in the annual agricultural plans and developing proactive solutions to meet the citizen’s need for his daily sustenance.