Historical data show that the unemployment rate in Palestine remains high, leading to significant challenges. Predicting unemployment rates would assist decision-makers in economic and financial planning, enabling them to implement preventive measures to mitigate their economic and social consequences. Research indicates that classical time series models may fail by producing a white noise error, given that unemployment rates are nonstationary and nonlinear in nature. This study empirically investigates the behavior of the Palestinian quarterly unemployment rate by addressing the research question: How efficient are neural network auto-regression (NNAR) models in forecasting the short-term Palestinian unemployment rate? Using quarterly unemployment rate data from 2001 to 2023, the study applies the NNAR model to determine the short-run state of the Palestinian unemployment rate. The model’s performance is evaluated using accuracy measures, including mean absolute error (MAE), mean absolute percentage error (MAPE), mean absolute scaled error (MASE), and root mean square error (RMSE). For comparison purposes, the study applies seasonal auto-regressive moving average (SARIMA), Holt-Winter’s (HW) additive, and HW’s multiplicative models. Findings indicate that NNAR [1, 1, 10] 4 is the optimal model, outperforming conventional models. The results also reveal that the Palestinian unemployment rate is expected to remain high, fluctuating between
تاريخ النشر
15/12/2024
الناشر
Electronic Journal of Applied Statistical Analysis
محسن حسين عيّاش | Mohsen Ayyash | 850