A NARX Neural Network Model To Predict One-Day Ahead Movement of The Stock Market Index of Malaysia

Abstract

Stock price prediction is one of the most challenging tasks confronting investors in the stock market. Academics and practitioners have examined a variety of techniques for predicting stock price movement. Artificial intelligence models have attracted many researchers interested in financial forecasting in the stock market. This paper investigates the use of commodity prices and exchange rates to predict the KLCI index price one day ahead using the NARX-ANN model. The methodology employs various experiments on the sample dataset using various exogenous variables and neural training algorithms. The model's prediction performance using four input variables (three commodity prices and the exchange rate) is compared with the model's prediction performance when only the KLCI price data with price lags is used as input variables. The model's performance shows promising results in terms of error rate and hit rate. However, using commodity prices and exchange rate combinations did not improve the model's short-term next-day prediction of the KLCI price.