A STATISTICAL MODEL FOR FORECASTING FOREIGN DIRECT INVESTMENT IN EGYPT USING THE HYBRID APPROACH OF ANN AND ARIMA MODELS
This study aims to use the hybrid method between artificial neural networks (ANN) and autoregressive integrated moving average (ARIMA) models for forecasting foreign direct investment (FDI) in Egypt from Q1 2019 to Q4 2021. The ANN, ARIMA and hybrid models were compared to choose the best model in terms of explanatory ability by using multiple determination coefficient R2, as well as predictive ability by using square root of the mean squared errors (RMSE) and mean absolute error (MAE) criteria. The statistical result indicates that the optimal model was the hybrid model ARIMA-ANN (3, 1, 0) (5-1-4) where its explanatory and predictive capacity was higher than the other models. The result shows the most important economic factors affecting foreign direct investment, including imports (IMP), gross domestic production (GDP), exports (EXP), consumer price index (CPI) and interest rate (IR).
foreign direct investments (FDI), artificial neural networks (ANN), autoregressive integrated moving averages (ARIMA).