Indonesia Conference Directory

This research aims to predict the value of the Rupiah against the US Dollar using the Backpropagation Neural Network algorithm. In order to find out the potential exchange rate of the Rupiah against the US Dollar in the future. The stages carried out in this study are by comparing the test results of 5 Backpropagation Neural network architectures in the hidden layer. The data used are secondary data which are factors that influence the exchange rate. Testing is done using Rapidminer by conducting training process training data and testing data. From the results of the training process and data testing training on the five backpropagation artificial network architectures. Produce a number of outputs in the form of charts that illustrate the level of indication of the best similarity from the results of predictions and RMSE error values as a test result on the first architecture with an average number of error values of 4.8665, in the second Architecture with an average of 5.0645, then on the third architecture evenly - average of 5,0095, the fourth architecture with an average of 258,723 and on the fifth architecture with an average of 262.26.