Modeling and Forecasting the Euro/Lek Exchange Rate Using Advanced Time Series Models and Artificial Intelligence
Modeling and Forecasting the Euro/Lek Exchange Rate Using Advanced Time Series Models and Artificial Intelligence

Author(s): Katerina Zela, Dorjan Zela
Subject(s): Economy, National Economy, Financial Markets, ICT Information and Communications Technologies
Published by: Университет за национално и световно стопанство (УНСС)
Keywords: Euro/ Lek; time series; ARIMA; artificial neural network; financial forecasting
Summary/Abstract: The Euro/ Lek exchange rate is a key indicator for the financial stability and economic development of Albania, directly affecting foreign trade, remittances and monetary policies. This study aims to model and forecast the performance of the Euro/ Lek exchange rate using combined statistical and artificial intelligence approaches. Initially, the structure of historical data is analyzed through classical time series methods such as ARIMA and SARIMA, which provide the basis for capturing linear features and seasonality. Then, to address the complexity and non-linear behavior of the market, artificial neural network (ANN) and Long Short-Term Memory (LSTM) models are integrated, building a hybrid ARIMA-ANN approach. The comparison of the models' performance is carried out through error indicators such as RMSE, MSE and MAPE, demonstrating that hybrid models provide more accurate results in short- and medium-term forecasting. The study findings suggest that the use of advanced artificial intelligence methods, in combination with traditional models, represents a powerful tool for banks, businesses and policymakers in risk management and in making strategic decisions on the exchange rate.

• Details
• Contents