Rupiah Exchange Rate Prediction with Long Short-Term Memory Algorithm

Poernomo, Ayu

Syntax Literate: Jurnal Ilmiah Indonesia
Jurnal Ilmiah Indonesia

The fluctuation of the Rupiah exchange rate against foreign currencies in Asia presents a significant challenge in maintaining Indonesia’s economic stability. This study aims to forecast Rupiah exchange rates using the Long Short-Term Memory (LSTM) algorithm. Weekly exchange rate data from 2020 to 2024 were analyzed using a machine learning approach. The process involved data normalization, model training, and evaluation using Mean Absolute Per- centage Error (MAPE) and R-Squared. The results indicate that the LSTM model effectively captures non-linear patterns in time series data with high accuracy. This model implementation provides valuable insights for financial decision-makers, regulators, and academics in understanding the dynamics of foreign exchange markets.