Implementation of recurrent neural network for the forecasting of USD buy rate against IDR

<span lang="EN-US">This study implements a recurrent neural network (RNN) by comparing two RNN structures, namely Elman and Jordan using the backpropagation through time (BPTT) programming algorithm in training forecasting process foreign exchange cases. The activation functions used are linear transfer function, tan-sigmoid function (Tansig), log-sigmoid (Logsig), which applied to hidden output layers. application of results being most appropriate for layer, while is layer. Based on USD against IDR currency, BPTT method better than method, with best iteration 4000<sup>th</sup> both. lowest root mean square error (RMSE) values produced were 0.073477 122.15 following day, yielded 0.130317 222.96 also day.</span><p> </p>