<![CDATA[This study aims to compare the performance of the Gated Recurrent Unit (GRU) and Convolutional Neural Network (CNN) methods in predicting the exchange rate of the Argentine Peso (ARS) against the United States Dollar (USD). Using historical exchange rate data from January 2017 to December 2022, both models were trained and evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R² Score metrics. The results showed that the GRU model outperformed the CNN model in all evaluation metrics with MSE of 1.907 compared to 3.273 for CNN, RMSE of 1.381 compared to 1.809 for CNN, MAE of 1.063 compared to 1.433 for CNN, and R² Score of 0.996 compared to 0.994 for CNN. This study shows that GRU is more effective in capturing temporal patterns in currency exchange rate data compared to CNN, which highlights the advantages of recurrent architecture for financial time series prediction problems.]]>
Copyrights © 2024
