Triyanna Widiyaningtyas, Muhammad Rakha Pinggala, Heni Vidia Sari, Aisyah, Denny Widhiyanuriyawan, Wahyu Caesarendra
The United States Dollar (USD) serves as the official currency of the United States. The reason many use the USD as their medium of exchange is due to the strength of the American economy; the United States has the largest and most diverse economy in the world. This study evaluates the performance of the Prophet algorithm to predict the USD-IDR exchange rate using univariate and multivariate approaches. The Prophet algorithm, introduced by Facebook (now META), is the chosen method for this investigation. Additional data, such as inflation, Bank Indonesia interest rate, and Federal Reserve interest rate, are used to improve prediction accuracy. The metrics employed in this study are Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE). The evaluation results show that the multivariate Prophet gives better results, with a high accuracy of MAPE 1.14% for training and 3.92% for testing. Additionally, the RMSE scores of multivariate models yield better scores compared to univariate models. The results also show that the Federal Reserve interest rate most influences USD exchange rate fluctuations with a correlation of 0.67. © 2025 IEEE.
Universitas Negeri Malang, Department of Electrical Engineering and Informatics, Malang, Indonesia; Universitas Brawijaya, Department of Mechanical Engineering, Malang, Indonesia; Curtin University Malaysia, Department of Mechanical and Mechatronics Engineering, Sarawak, Malaysia
