A predictive system integrating intrinsic mode functions, artificial neural networks, and genetic algorithms for forecasting S&P500 intra‐day data

• Date: 01 April 2020
• Author: Salim Lahmiri
• DOI: http://doi.org/10.1002/isaf.1470
• Published date: 01 April 2020

RESEARCH ARTICLE

A predictive system integrating intrinsic mode functions,

artificial neural networks, and genetic algorithms for

forecasting S&P500 intra-day data

Salim Lahmiri

Department of Supply Chain and Business

Technology Management, John Molson School

of Business, Concordia University, Montreal,

Canada

Correspondence

Salim Lahmiri, Department of Supply Chain

and Business Technology Management, John

Molson School of Business, Concordia

University, Montreal, QC, Canada.

Email: salim.lahmiri@concordia.ca

Summary

There is an abundant literature on the design of intelligent systems to forecast stock

market indices. In general, the existing stock market price forecasting approaches can

achieve good results. The goal of our study is to develop an effective intelligent pre-

dictive system to improve the forecasting accuracy. Therefore, our proposed predic-

tive system integrates adaptive filtering, artificial neural networks (ANNs), and

evolutionary optimization. Specifically, it is based on the empirical mode decomposi-

tion (EMD), which is a useful adaptive signal-processing technique, and ANNs, which

are powerful adaptive intelligent systems suitable for noisy data learning and predic-

tion, such as stock market intra-day data. Our system hybridizes intrinsic mode func-

tions (IMFs) obtained from EMD and ANNs optimized by genetic algorithms (GAs) for

the analysis and forecasting of S&P500 intra-day price data. For comparison pur-

poses, the performance of the EMD-GA-ANN presented is compared with that of a

GA-ANN trained with a wavelet transform's (WT's) resulting approximation and

details coefficients, and a GA-general regression neural network (GRNN) trained with

price historical data. The mean absolute deviation, mean absolute error, and root-

mean-squared errors show evidence of the superiority of EMD-GA-ANN over WT-

GA-ANN and GA-GRNN. In addition, it outperformed existing predictive systems

tested on the same data set. Furthermore, our hybrid predictive system is relatively

easy to implement and not highly time-consuming to run. Furthermore, it was found

that the Daubechies wavelet showed quite a higher prediction accuracy than the

Haar wavelet. Moreover, prediction errors decrease with the level of decomposition.

KEYWORDS

artificial neural networks, Empirical mode decomposition, forecasting intra-day data, intrinsic

mode functions, stock market, wavelet transform

1|INTRODUCTION

Time-series modelling and prediction is a fundamental task in business

research and analytics (Armstrong et al., 2015; Goodwin, 2015; Fildes

and Petropoulos, 2015). Indeed, it is receiving growing interest in dif-

ferent business applications, including retail sales (Merino and

Ramirez-Nafarrate, 2016), inventory demand (Syntetos et al., 2015),

market size (Qian and Soopramanien, 2014), and accounting (Brown

and Fernando, 2011). In particular, analysing and forecasting stock

market prices have been the subject of many empirical studies in aca-

demia since these are one of the primary inputs to a wide range of

financial applications, including portfolio optimization and derivatives

pricing. Certainly, financial investment decisions depend strongly on

the forecast of stock expected return and volatility. In general,

Received: 29 June 2019 Revised: 21 October 2019 Accepted: 22 February 2020

DOI: 10.1002/isaf.1470

Intell Sys Acc Fin Mgmt. 2020;27:55–65. wileyonlinelibrary.com/journal/isaf © 2020 John Wiley & Sons, Ltd. 55