Exchange rate stability is one of the key indicators of macroeconomic performance and financial resilience. Understanding and forecasting the movements of a national currency are essential for designing effective monetary and fiscal policies. This study investigates the dynamics of the Uzbek soum against the U.S. dollar and provides a short-term forecast of its future trajectory using the Autoregressive Integrated Moving Average (ARIMA) model. The analysis is based on weekly data covering the period from September 2017 to October 2025, obtained from reliable financial databases.

Before model estimation, the Augmented Dickey–Fuller (ADF) test was employed to check the stationarity of the exchange rate series. The results indicated that the variable is non-stationary in levels but becomes stationary after first differencing. The autocorrelation and partial autocorrelation functions (ACF and PACF) of the differenced series were examined to identify suitable model parameters. Based on these diagnostics, the ARIMA(1,1,0) model was selected as the best-fitting specification for capturing the short-term dynamics of the weekly exchange rate.

The estimated model coefficients were statistically significant, confirming that past changes in the exchange rate have predictive power for future movements. Diagnostic tests, including the correlogram and Ljung–Box Q-statistics, verified that the model residuals are free from serial correlation, indicating that the ARIMA(1,1,0) model provides an adequate representation of the data. Using this model, forecasts were generated for ten future weekly periods to assess the potential direction of the exchange rate.

The results of the study provide important insights into the short-run behavior of the Uzbek soum and offer a data-driven basis for future exchange rate analysis and policy planning. The application of ARIMA modeling demonstrates its usefulness as a forecasting tool for monetary authorities and researchers monitoring currency stability.

Exchange rate, ARIMA model, forecasting

Lakhal, S. (2024). Forecasting the EUR/USD exchange rate using ARIMA and machine learning models. Data and Metadata. https://doi.org/10.56294/dm2024.368

Xu, M., & Li, W. (2017). Research on exchange rate forecasting model based on ARIMA model and artificial neural network model. In Proceedings of MSMEE 2017 (Vol. 123). Atlantis Press.

Berdinazarov, Z., Dodoev, K., Mamasalaev, J., & Fakhodjonov, J. (2019). Determinants of exchange rate fluctuations of Uzbek sum. [Journal name].

Escudero, P., Alcocer, W., & Paredes, J. (2021). Recurrent neural networks and ARIMA models for Euro/Dollar exchange rate forecasting. Applied Sciences, 11(12), 5658. https://doi.org/10.3390/app11125658

Perveen, N., Tariq, K., & Ahmad, H. S. (2024). Comparative studies of hybrid ARIMA and artificial neural network (ANN) techniques for predicting exchange rate in Pakistan. Bulletin of Business and Economics. https://doi.org/10.61506/01.00372

Forecasting Economics. (2021, May 22). Time series forecasting in Stata: ARIMA models (Part 1, 2, 3) [Video]. YouTube. https://www.youtube.com/

Vika, B., Sevrani, K., & Vika, I. (2016). The usefulness of artificial neural networks in forecasting exchange rates. Academic Journal of Interdisciplinary Studies, 5(1), 73. https://doi.org/10.5901/ajis.2016.v5n1p73