Backpropagation Neural Network Application in Predicting The Stock Price of PT Bank Rakyat Indonesia Tbk

Abstract

Investors often make mistakes when making stock transactions even though having chosen good company stocks. The thing that needs to be considered in making stock transactions is to see the movement of stock prices. The movement of the stock price in PT Bank Rakyat Indonesia Tbk has changed in the form of a decrease or increase. An increasing stock price will provide benefits for investors by selling stocks. But, investors actually decide to make stock purchases. The existence of stock purchase transactions causes investors to take a high risk because stock prices fluctuate. To anticipate the occurrence of high risk to investor, stock price predictions are made using a Backpropagation Neural Network (BPNN). BPNN can adapt quickly and is able to predict nonlinear data such as stock prices and produce a high level of accuracy. The results of this study obtained the best BPNN model, namely the BP(5,3,1) model with a Mean Absolute Percentage Error (MAPE) of 0,8193%. These results show that the model has good network performance so that it can predict stock prices well because it gets a small prediction error.