<![CDATA[Foreign exchange trading is one of the largest financial markets in the world, therefore the movement of foreign exchange prices is always fluctuating every day. Therefore, an accurate prediction system is needed to determine the price that will occur. Predicted prices can help traders in determining the right position in running this business. Predicted price can be done by doing analysis on data formed in the past. One way that can be done is to perform technical analysis by using artificial neural network. In this study neural network will be used to predict the price of the euro against the US dollar on daily movement. The artificial neural networks used are trained with supervised learning methods, which use the input and output data sets. Through network learning can recognize patterns between input and output. The algorithm used is backpropagation, this algorithm uses error value to change network weight at step back propagation. While the error value is obtained through advanced browsing steps, which do first. The performance of artificial neural networks is measured using the value of MSE (Mean Square Error) and the MAPE (Mean Absolute Percentage Error) value. From the results of the research, it was found that the network with 6 input neurons and 11 hidden layer neurons became the best network with MSE value of 0.000001 and MAPE value of 0,59%.]]>
Copyrights © 2017
