King Talal School of Business Technology

Advanced Financial Technology

:Proposal Topic
The Role of AI in Predicting Forex Market
Movements
:By
Zain Al-Omari
Anas Al-Jabi
Supervised by:
Dr. Laith Almaqableh
 1

1- Introduction

In the complex world of foreign exchange markets, where currencies are bought and sold

simultaneously, predicting movements and fluctuations affecting trade profitability within seconds poses

a challenge and an opportunity. As technology evolves, artificial intelligence (AI) has become a powerful

tool in decoding forex trading complexities. In this paper, we explore AI's role in predicting forex market

movements and exchange rate fluctuations. Currency trading is increasingly popular, with the forex

market being more reliable, secure, and regulated compared to cryptocurrency exchanges. The forex

market, or FOREX, is the largest globally, with daily trades exceeding $5.1 trillion. Its volatile and complex

nature draws comparisons to a black box, despite being available around the clock, split into four main

time zones. This landscape, influenced by geopolitical, economic, and psychological factors, often baffles

even the most seasoned traders. Stock market investments, driven by forecasts, witness billions

exchanged daily worldwide. The proliferation of prediction methods reflects their crucial role in

foreseeing future events, especially in currency markets where traditional analysis often falls short. AI's

ability to learn from past experiences is crucial, particularly in the economy, where it's widely used for

forex and gold foresight (Amat, Michalski, & Stoltz, 2018b). Testing sophisticated algorithms and machine

learning techniques against traditional methods aims to improve forecasting outcomes. AI's strength lies

in its adaptability and learning from historical data, continuously refining predictive capabilities over

time. By analyzing various factors including economic indicators, central bank policies, geopolitical

events, and market behavior, AI models identify trends and correlations influencing commodity

movements. Moreover, AI systems process information rapidly, enabling real-time analysis and decision-

making in a market where time is of the essence. This study aims to analyze Bitcoin and gold prices using

traditional and AI-based methodologies, comparing their predictive accuracies against actual market

outcomes. Through rigorous analysis and comparison against real-world data, the study highlights the
 2

strengths and limitations of both approaches in predicting price movements in cryptocurrency and

commodity markets.

This paper will answer the following questions:

1. Are AI methods more accurate in forecasting commodities and cryptocurrency prices

than the traditional method?

2. Which AI tool is more suitable to forecast the most accurate results?

2 -literature review

In the twenty-first century, the financial market holds significant sway over expanding economies,

nations, and societies (Nassirtoussi et al., 2014). Understanding this market is crucial due to the

multitude of uncertainties it faces, including economic conditions, social factors, and political events

(Adebiyi et al., 2012, 2014; Bisoi and Dash, 2014; Ding et al., 2014; Rajashree et al., 2014; Rather et al.,

2014; Lin, 2018). Among these uncertainties, the exchange rate stands out, making its parameters highly

unpredictable. Accurate exchange rate predictions greatly influence global markets and domestic

economic policy formulation, guiding both governments and businesses in assessing currency purchasing

power (Baghestani & Toledo, 2019).

Two methods dominate market prediction: fundamental and technical analysis. Fundamental

analysis encompasses various factors like a country's financial system, political stability, company

popularity, and market news, while technical analysis relies solely on historical data for price predictions.

These methods, along with machine learning, serve as primary tools for stock market predictions (Dunne,

2015). Some academics categorize them as technical and fundamental analysis (Nassirtoussi et al., 2014;

Dunne, 2015; Gyan, 2015; Prem Sankar et al., 2015; Ahmadi et al., 2018), contributing to a comprehensive
 3

valuation estimate. Both technical and fundamental analyses aim to predict stock returns, allowing

investors to buy low and sell high, albeit through different approaches, yet sharing the same goal:

evaluating stock markets.

2.1 Technical analysis

Technical analysis, a popular method for forex forecasting, was first introduced by Charles Dow

and the Dow Theory in the late 1800s. It involves examining historical pricing data to identify patterns,

trends, and indicators for forecasting future price changes. Utilizing well-informed trading decisions,

traders can assess a security's strength or weakness in relation to the market or its sectors, often

generating short-term trading signals using charting tools.

While technical analysis is widely employed in financial markets, academics have not universally

supported it despite its availability (Park & Irwin, 2009; Brock, Lakonishok, & LeBaron, 1992;

Schulmeister, 2009). This method relies on past stock price data, aiming to estimate future returns

through the analysis of historical market data, primarily stock price and volume (Chavarnakul & Enke,

2009; Park & Irwin, 2007). Integrating statistical models, intelligent system methodologies, and

conventional trading principles, studies have explored technical analysis (Bisoi & Dash, 2014).

Traders in the FX and commodities markets favor technical analysis due to their interest in short-

term price changes (Gorgulho et al., 2011). They employ various tools and strategies, including

oscillators, moving averages, chart patterns, and Fibonacci retracements. Chart patterns like triangles

and head and shoulders provide insights into market reversals or continuations, while oscillators like the

Relative Strength Index (RSI) indicate overbought or oversold conditions (Murphy, 1999). Moving

averages aid in trend identification, while Fibonacci retracements identify potential support and

resistance levels.
 4

Technical analysts use past market prices and technical indicators to anticipate market directions

(Suresh-Kumar & Elango, 2011). Research demonstrates predictability in currency markets based on

national traits and international variables, resulting in significant carry trading profits (Filippou, Rapach,

Taylor, & Zhou, 2020). Technical analysis extends beyond the stock market, proving useful across asset

classes, though predictability tends to decrease over time (Guo, Lin, Wu, & Zhou, 2020). By analyzing

historical and current price movements, technical analysts can forecast future stock prices (Anbalagan &

Maheswari, 2014).

2.2 Fundamental Analysis

Based on a review of Fundamental Analysis literature (Fontanilla & Gentile, 2001), it entails

understanding rules and procedures to determine the intrinsic value of stocks in stock markets. This

includes studying economic forecasts, sales generation, and profitability sectors, assessing financial

company strength, management efficiency, and business opportunities from historical financial

statements and current conditions.

Fundamental analysis focuses on factors such as interest rates, GDP growth, employment data,

inflation, central bank policies, and political stability to predict currency value. Traders analyze economic

calendars to track key events and news releases that affect currency movements (Anbalagan &

Maheswari, 2014). It's essential for long-term predictions rather than short-term fluctuations (Khan,

2011).

Fundamental analysts consider various factors, including political data, natural disasters, financial

status, and corporate reports, to forecast stock prices (Tsai & Hsiao, 2010). By evaluating past, present,

and projected earnings, analysts determine the fair value of stocks (Bernard, 1994; Piotroski, 2000).
 5

Studies (Zhang et al., 2011; Chung & Kim, 2001) have shown that fundamental analysis is a

reliable indicator of stock price changes. Combining profitability, solvency, liquidity, and operating

efficiency indicators aids in creating successful investment strategies (Venkates, Madhu, & Ganesh, 2012).

Fundamental analysis components like cash flows, outside funding, and dividends help predict future

stock returns and explain stock price momentum phenomena (Nguyen, 2003).

A blend of technical and fundamental analysis can effectively predict gold and Bitcoin prices.

Technical analysis identifies short-term trends and trading opportunities in volatile cryptocurrency

markets, while fundamental research provides insights into long-term asset drivers (Anbalagan &

Maheswari, 2014). Combining both strategies can enhance prediction accuracy and understanding of

market dynamics (Venkates, Madhu, & Ganesh, 2012).

However, the unstructured nature of fundamental factors makes automating fundamental analysis

challenging. Nonetheless, advancements in machine learning enable the automation of stock market

predictions based on unstructured data, potentially improving prediction accuracy (Zhang et al., 2011).

2.3 Artificial Intelligence

In terms of technological innovation, three significant changes have transformed the foreign

exchange market. Electronic trading was the first innovation, followed by the Internet, and artificial

intelligence (AI) arrived third, primarily utilized for exchange rate prediction. AI algorithms have

demonstrated the ability to produce precise short-term projections, although long-term forecasting

remains challenging due to its ambiguity. However, modern AI programs can estimate currency prices

over a one-week period with consistent parameters.

AI supports the world's most actively traded markets by utilizing complex mathematical and

statistical models, removing geopolitical barriers, and connecting disparate market segments. Machine
 6

learning (ML) techniques have become a powerful tool for forecasting Forex trading by analyzing large

volumes of data, spotting patterns, and adjusting to shifting market conditions.

Deep learning models, particularly neural networks, have gained popularity in Forex forecasting,

effectively capturing sequential dependencies in time series data. Artificial neural networks (ANN)

replicate the human brain's learning process and possess the ability to forecast and analyze stock market

movements. LSTM, a form of recurrent neural network (RNN), excels in learning long-term dependencies,

making it suitable for financial asset price predictions.

Recent studies have proposed various models and techniques for forecasting Forex and

cryptocurrency prices using AI and ML. For instance, deep neural networks (DNNs) and convolutional

neural networks (CNNs) have been employed to predict Bitcoin price trends, while C-RNN techniques

have been used for Forex time series forecasting.

Sentiment analysis, which determines the emotion behind a series of words, has gained traction in

financial analysis, particularly through social media platforms. Twitter sentiment has been correlated with

trading volume and returns in major international financial markets. Moreover, sentiment analysis from

news stories has been integrated with technical analysis indicators to create prediction algorithms.

The combination of neural networks with expert systems has improved prediction accuracy and

performance in financial forecasting. However, compared to equities, the forex market is more complex

due to its high volatility, nonlinearity, and irregularity. Nonetheless, AI and ML methods offer effective

means for Forex forecasting, considering both quantitative and qualitative factors that influence exchange

rate fluctuations.
 7

Methodology-3

Based on the comparison provided, determining the "best" method for market analysis depends

on various factors, including the trader's goals, time horizon, and expertise level. Technical analysis,

focusing on historical price data and chart patterns, is particularly effective for short-term trading due to

its immediate trading signals and ability to identify trends and market reversals. However, its predictive

power may decrease over time, and it does not account for broader economic or company-specific

factors.

On the other hand, fundamental analysis, centered around the intrinsic value of assets based on

economic indicators and company performance, is more suitable for long-term investment decisions. It

provides a deep understanding of underlying asset value and is useful for identifying mispriced securities

and long-term growth potential. Nonetheless, its reliance on complex and unstructured data makes it less

effective for short-term price movements.

Machine learning and AI techniques offer a promising approach to market analysis by automating

and enhancing prediction models using both structured and unstructured data. These methods, such as

neural networks and decision trees, demonstrate high predictive accuracy, especially in handling

complex and nonlinear data. However, they require significant computational resources and expertise,

and their opaque nature can make interpretation challenging.

Contribution-4

The introduction sets the stage for exploring the role of artificial intelligence (AI) in predicting

movements in foreign exchange (forex) markets. It portrays the forex market as a complex and volatile

environment where predicting fluctuations is both challenging and rewarding. By contrasting forex

trading with cryptocurrency exchanges, the introduction emphasizes the reliability, security, and
 8

regulation of forex trading. The significance of the forex market, with its massive daily trading volume, is

highlighted, along with its divided trading hours across major time zones. Furthermore, the introduction

underscores the challenges posed by geopolitical, economic, and psychological factors in forex trading. It

also introduces the relevance of AI in overcoming these challenges, leveraging its ability to analyze vast

amounts of data, learn from historical patterns, and adapt to changing market conditions. The paper's

objective to analyze the accuracy of AI-based methodologies in predicting commodity and

cryptocurrency prices compared to traditional methods is outlined, setting the research agenda for the

subsequent sections. Finally, it presents key questions the study aims to address, focusing on the

comparative effectiveness of AI methods and traditional approaches in forecasting market movements.

5-Conclusion

In-depth discussion of forex trading and artificial intelligence's (AI) capacity to forecast market

movements is provided in this article. It talks about the difficulties caused by geopolitical, economic, and

psychological variables as well as the complexities of the foreign exchange market and how reliable it is

in comparison to cryptocurrency exchanges. The study emphasizes AI-driven methods as effective means

of enhancing prediction accuracy and evaluating enormous volumes of data. In predicting commodities

and cryptocurrency values, it contrasts conventional and AI-based approaches and comes to the

conclusion that AI techniques have potential to improve prediction accuracy, particularly in volatile

markets like cryptocurrencies. The use of artificial intelligence (AI) technologies into forex forecasting is

considered a noteworthy progression, providing traders with unparalleled perspectives and prospects

for well-informed decision-making.

 9

References

1. Levinson, M. (2014). The Economist Guide to Financial Markets: Why They Exist and How They Work.
London, UK: The Economist.

2. Ozturk, M., Toroslu, I. H., & Fidan, G. (2016). Heuristic based trading system on Forex data using technical
indicator rules. Applied Soft Computing, 43, 170–186.

3. Anastasakis, L., & Mort, N. (2009). Exchange rate forecasting using a combined parametric and
nonparametric self-organizing modeling approach. Expert Systems with Applications, 36, 12001–12011.

4. Cartea, A., Huang, R. D., & Masulis, R. W. (2019). FX spreads and dealer competition across the 24-h
trading day. Review of Financial Studies, 12, 61–93.

5. Masry, S., Dupuis, A., Olsen, R., & Tsang, E. (2013). Time zone normalization of FX seasonality.
Quantitative Finance, 13, 1115–1123.

6. Wanjawa, B. W., & Muchemi, L. (2014). ANN Model to Predict Stock Prices at Stock Exchange Markets.

7. Ghaznavi, A., Aliyari, M., & Mohammadi, M. R. (2016). Predicting stock price changes of Tehran artmis
company using radial basis function neural networks. International Research Journal of Applied Basic Sciences,
10(8), 972–978.

8. (Ticknor, 2013; Chen et al., 2014; Metghalchi et al., 2014; Rather et al., 2014; Prem Sankar et al., 2015;
Wanjawa, 2016; Agarwal et al., 2017; Tsai & Wang, 2017; Lin, 2018; Zhou et al., 2018).

9. Amat, Christophe, Michalski, T., & Stoltz, G. (2018). Fundamentals and exchange rate forecastability with
simple machine learning methods. Journal of International Money and Finance, 88, 1–24.

10. Nassirtoussi, A. K., Aghabozorgi, S., Wah, T. Y., & Ngo, D. C. L. (2014). Text mining for market
prediction: A systematic review. Expert Systems with Applications, 41, 7653–7670.

11. Adebiyi et al. (2012, 2014); Bisoi and Dash (2014); Ding et al. (2014); Rajashree et al. (2014); Rather et al.
(2014); Lin (2018).

12. Baghestani, H., & Toledo, H. (2019). Oil prices and real exchange rates in the NAFTA region. The North
American Journal of Economics and Finance, 48, 253-264.

13. Vanstone, B., & Finnie, G. (2009). An empirical methodology for developing stockmarket trading systems
using artificial neural networks. Expert Systems with Applications, 36(3), 6668-6680.

14. Murphy, J. J. (1999). Technical Analysis of the Financial Markets: A Comprehensive Guide to Trading
Methods and Applications. New York, NY, USA: New York Institute of Finance.

15. (Suresh- Kumar and Elango, 2011; Wei et al., 2011; Suthar et al., 2012; de Oliveira et al., 2013; Ballings et
al., 2015; Gaius, 2015; Su and Cheng, 2016).
 10

16. Filippou, I., Rapach, D., Taylor, M. P., & Zhou, G. (2020, September 10). Exchange Rate Prediction with
Machine Learning and a Smart Carry Portfolio.

17. Guo, X., Lin, H., Wu, C., & Zhou, G. (2020, December 2). Extracting Information from Corporate Bond
Yields: A Machine Learning Approach.

18. Fontanills, A. G., & Gentile, T. (2001). The Stock Market Course. John Wiley & Sons, Inc.

19. (Bernard, 1994; Fischer and Jordan, 1995; Penman, 1991; Bauman, 1996; Rielly and Brown, 2002; Jones,
2007; Bodie et al., 2009).

20. Reilly, F., & Brown, K. (2002). Investment Analysis and Portfolio Management (7th Ed.). South-Western.

21. Chung, H. Y., & Kim, J. B. (2001). A structured financial statement analysis and the direct prediction of
stock prices in Korea. Asia Pacific Financial Markets, 8(2), 87-117. Venkates, R., Madhu, S. C., & Ganesh, L.
S. (2012).

22. Brock, W. A., Lakonishok, J., & LeBaron, B. (1992). Simple technical trading rules and the stochastic
properties of stock returns. Journal of Finance, 47, 1731–1764.

23. Nguyen, P. (2003). Fundamental analysis and stock returns: Japan 1993-2003. Tokyo: WBP Financial
Integrator.

24. Ball, R., Kothari, S. P., & Watts, R. L. (1993). Economic determinants of the relation between earnings
changes and stock returns. The Accounting Review, 68(3), 622-638.

25. Piotroski, J. (2002). Value investing: The use of historical financial statement information to separate
winners from losers. Journal of Accounting Research.

26. Chavarnakul, T., & Enke, D. (2009). Parisi & Vasquez, & Roberts. In Journal Title (p. 3517).

27. Mahmoud, A., & Sakr, S. (2012). The predictive power of fundamental analysis in terms of stock return and
future profitability performance. International Research Journal of Finance and Economics.

28. Chung, H. Y., & Kim, J. B. (2001). A structured financial statement analysis and the direct prediction of
stock prices in Korea. Asia-Pacific Financial Markets, 8, 87–117.

29. Reilly, F., & Brown, K. (2002). Investment Analysis and Portfolio Management (7th ed.). South-Western.

30. Bernard, V. L. (1994). Accounting-based valuation methods, determinants of market-to-book ratios, and
implications for financial statements analysis (Working paper No. 9401). University of Michigan, Business
School, Faculty Research.
11