Artificial Neural Network for Predicting Indonesia Stock Exchange Composite Using Macroeconomic Variables

Abstract

Stock is a high risk and high return investment. The risk-comparison scale for both losses and profits are not much different. The lure of profits temptations can be given by playing shares, sometimes make people less cautious and eventually fail to invest in stocks. To make right and profitable investment decisions, investors need to face uncertainty and fluctuating stock price movements. These phenomena cause investors to predict stock price movements for minimizing risks. The purpose of this study is to predict the Indonesian composite stock price index by using macroeconomic variables as a reflection of economic condition and as a good signal to forecast stock prices. This research is using Inflation, Interest Rates, and Exchange Rates as the macroeconomic variables. This study uses secondary data from Bank Indonesia and Indonesian Statistics Center from December 2005 to November 2017. The prediction uses Artificial Neural Network (ANN) Backpropagation method. The results gained the accuracy of 96,38% and mean-squared error of 0.0046 with the best time delay of 2 months before the predicted month. Based on the accuracy level and the error, macroeconomic variables (exchange rate, interest rate, inflation rate, and money supply M2) are the proper indicator to predict IDX Composite movement.