Prediksi Harga Indeks Dollar pada Tahun 2019-2024 Menggunakan Backpropagation Jaringan Saraf Tiruan

Abstract

This study aims to predict the United States Dollar Index (USDX) using an Artificial Neural Network (ANN) model with a backpropagation approach. Historical USDX data from 2019 to 2021 was used to train the model, and predictions were made for USDX values in 2024. The ANN model, consisting of two layers with 50 neurons in the hidden layer, was trained over 200 epochs. The prediction results indicate that the model can estimate USDX values with a relatively low Mean Squared Error (MSE) and an acceptable percentage error based on the actual average value for 2024. The visualized results demonstrate that the ANN model can follow historical data patterns well, although there is room for improvement, particularly in capturing seasonal variations.