Stock Price Prediction in KLSE Using Neural Networks

Abstract

The use of neural networks in financial market prediction presents a major challenge to the design of effective neural network predictors. This paper presents a study to evaluate capabilities of five prediction approaches that use backpropagation neural networks model in pedicting stock prices of Kuala Lumpur Stock Exchange (KLSE). The approaches considered are the Standard, Volatility Data, Technical and Fundamental Data, Data Aggregation, and the Sector-Counter Approach. We found that the network that considers Sector-Counter indices is the best performing network in prediction. The complexity of the financial market data probably explains why some of the approaches cannot provide any significant improvement in the prediction.