Aggregate Versus Disaggregate Data in Artificial Neural Network for Stock Market Forecasting

Authors

  • Edward Tang Chun Hoe Universiti Tun Hussein Onn Malaysia
  • Maria Elena Nor

Keywords:

Data Disaggregation, Artificial Neural Network, MFE, MAPE, Trend Change Error

Abstract

The stock market serves as a place for the public to earn profits in a short period of time and hence, the idea of stock market forecasting was then surfaced as the reference for those traders to avoid losses and gain profits. There are many types of methods available for forecasting, and this study aims to apply data disaggregation on the stock closing prices and utilize them to make forecasts through ANN. Specifically, it investigates whether the application of disaggregate series can provide better forecast performance than using the traditional aggregate series. Historical stock closing prices from Malayan Banking Berhad was used for this study, and the forecast made from the ANNs by using aggregate and disaggregate series were measured by using MFE, MAPE, and trend change error. According to the result, the aggregate forecast had outperformed the disaggregate forecast as it obtained lesser error in MFE, MAPE, and trend change error, and can be concluded that the use of aggregate series for ANN forecasting was the superior option.

Downloads

Published

03-08-2022

Issue

Section

Articles

How to Cite

Edward Tang Chun Hoe, & Nor, M. E. (2022). Aggregate Versus Disaggregate Data in Artificial Neural Network for Stock Market Forecasting. Enhanced Knowledge in Sciences and Technology, 2(1), 143-152. https://publisher.uthm.edu.my/periodicals/index.php/ekst/article/view/5256