The role of attribute selection in Deep ANNs learning framework for high‐frequency financial trading

• Published date: 01 April 2020
• Author: Monira Essa Aloud
• Date: 01 April 2020
• DOI: http://doi.org/10.1002/isaf.1466

RESEARCH ARTICLE

The role of attribute selection in Deep ANNs learning

framework for high-frequency financial trading

Monira Essa Aloud

Department of Management Information

Systems, College of Business Administration,

King Saud University, Riyadh, Saudi Arabia

Correspondence

Monira Essa Aloud, Department of

Management Information Systems, College of

Business Administration, King Saud University,

Riyadh, Saudi Arabia.

Email: mealoud@ksu.edu.sa

Funding information

OANDA and Olsen Ltd

SUMMARY

In financial trading, technical and quantitative analysis tools are used for the develop-

ment of decision support systems. Although these traditional tools are useful, new

techniques in the field of machine learning have been developed for time-series fore-

casting. This paper analyses the role of attribute selection on the development of a

simple deep-learning ANN (D-ANN) multi-agent framework to accomplish a profit-

able trading strategy in the course of a series of trading simulations in the foreign

exchange market. The paper evaluates the performance of the D-ANN multi-agent

framework over different time spans of high-frequency (HF) intraday asset time-

series data and determines how a set of the framework attributes produces effective

forecasting for profitable trading. The paper shows the existence of predictable

short-term price trends in the market time series, and an understanding of the proba-

bility of price movements may be useful to HF traders. The results of this paper can

be used to further develop financial decision-support systems and autonomous trad-

ing strategies for the financial market.

KEYWORDS

artificial intelligence, decision-support systems, deep learning in artificial neural network, FX

forecasting, genetic programming

1|INTRODUCTION

In the study and analysis of financial and other time-series data,

researchers have attempted to identify patterns in the time-series

data and hence define the statistical properties of the data. High-

frequency (HF) data refers to data recorded at a frequency that is

higher than what would be recorded on a daily basis (Dacorogna,

Gençay, Müller, Olsen, & Pictet, 2001). The availability of HF data

has provided innovative opportunities for the analysis of empirical

data and the search for new statistical properties. The challenges

associated with manual analyses have shown the need for the

development of more automated methods that enable the expert

analysis of such large amounts of HF financial data in an effort to

obtain meaningful statistics. The exploration of potentially hidden

knowledge in the market time-series data will present pathways for

financial investors to acquire practical and knowledge-driven

decisions to achieve substantial gain with a smaller degree of

investment risk. Therefore, it is now promising to pursue an empirical

and analytical approach, as well as to develop trading models bottom

up by studying and analysing financial data and searching for

investment arbitrage.

Financial analysts have developed technical analysis indicators

to examine the market time-series data and to support investors in

developing trading rules for investment decisions. Such technical

indicators are considered the most primary and traditional tools

used in financial trading (Leigh, Purvis, & Ragusa, 2002; Murphy,

1999). In the financial forecasting literature, some studies used

machine-learning (ML) techniques to develop forecasting models,

and technical analysis indicators have been used as inputs to these

models to discover the hidden patterns and the relationships

between them, in sequence, to forecast future prices movements

and thus identify the best trading indicators. The study and

Received: 30 July 2019 Revised: 5 February 2020 Accepted: 12 February 2020

DOI: 10.1002/isaf.1466

Intell Sys Acc Fin Mgmt. 2020;27:43–54. wileyonlinelibrary.com/journal/isaf © 2020 John Wiley & Sons, Ltd. 43