Machine Learning Algorithms for Forecasting and Categorizing Euro-to-Dollar Exchange Rates

Aug 4, 2026

Upload by Eliana

Forecasting changes in foreign exchange rates remains a central problem in finance, motivating extensive work using machine learning for market analysis and prediction. This study applies multiple machine-learning methods, including AdaBoost, logistic regression, gradient boosting, random forest classifier, bagging, Gaussian naïve Bayes, XGBoost classifier, and decision-tree classifier, and also tests a combined approach using logistic regression, random forest, and Gaussian naïve Bayes. Technical indicators are integrated into the training dataset to improve predictive precision. Experimental results show the proposed method achieves strong accuracy (0.948), enabling investors to identify advantageous times for EUR/USD buying and selling.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Abstract

Introduction

Foreign exchange market overview

Machine learning in financial time-series prediction

Related work and prior studies

Motivation and research objective

Received 13 March 2024, accepted 17 May 2024, date of publication 23 May 2024, date of current version 3 June 2024. Digital Object Identifier 10.1109/ACCESS.2024.3404824 Machine Learning Algorithms for Forecasting and Categorizing Euro-to-Dollar Exchange Rates MOHAMED EL MAHJOUBY 1, MOHAMED TAJ BENNANI1, MOHAMED LAMRINI 1, BADRE BOSSOUFI2, THAMER A. H. ALGHAMDI3,4, AND MOHAMED EL FAR1 1Department of Computer Science, Laboratory (LPAIS), Faculty of Science Dhar El Mahraz, Sidi Mohamed Ben Abdellah University, Fez 30003, Morocco 2LIMAS Laboratory, Faculty of Sciences Dhar El Mahraz, Sidi Mohamed Ben Abdellah University, Fez 30003, Morocco 3Wolfson Centre for Magnetics, School of Engineering, Cardiff University, CF24 3AA Cardiff, U.K. 4Electrical Engineering Department, School of Engineering, Al-Baha University, Al Bahah 65779, Saudi Arabia Corresponding authors: Mohamed El Mahjouby ([mohamed.elmahjouby@usmba.ac.ma](mohamed.elmahjouby@usmba.ac.ma)) and Badre Bossoufi ([badre.bossoufi@usmba.ac.ma](badre.bossoufi@usmba.ac.ma)) ABSTRACT Forecasting changes in foreign exchange rates is a well-explored and widely recognized area within finance. Numerous research endeavors have delved into the utilization of methods in machine learning to analyze and predict movements in the foreign exchange market. This work employed several machine-learning techniques such as Adaboost, logistic regression, gradient boosting, random forest classifier, bagging, Gaussian naïve Bayes, extreme gradient boosting classifier, decision tree classifier, and our approach (we have combined three models: logistic regression, random forest classifier, and Gaussian naive Bayes) . Our objective is to predict the most advantageous times for purchasing and selling the euro about the dollar. We integrated a range of technical indicators into the training dataset to enhance the precision of our techniques and strategy. The outcomes of our experiment demonstrate that our approach outperforms alternative methods, achieving superior prediction performance. Our methodology yielded an accuracy of 0.948 . This study will empower investors to make informed decisions about their future EUR/USD transactions, helping them identify the most advantageous times to buy and sell within the market. INDEX TERMS Foreign exchange, prediction, logistic regression, random forest, Naïve Bayes. I. INTRODUCTION Foreign exchange, sometimes shortened to ‘‘forex’’ or ‘‘FX,’’describes the international exchange market. Participants in this decentralized market, which includes banks, financial organizations, businesses, governments, and individual traders, purchase and sell several global currencies. Because it allows one currency to convert into another, the foreign exchange market is essential to international trade and investment. Several variables, such as interest rates, market sentiment, geopolitical developments, and economic indicators, affect exchange rates in the foreign exchange market. To reduce risk or pursue speculative opportunities, players in the foreign exchange market try to profit from currency swings. The associate editor coordinating the review of this manuscript and approving it for publication was Yongming Li . Recently, machine learning methods and algorithms—which have shown to be very successful in several fields, including biomedicine, image and speech analysis, and machine translation [1]— have been used in the examination of time series financial [2], [3] . The prevailing machine learning algorithms utilized in financial prediction encompass support vector machines [4] and diverse architectures of Artificial Neural Networks (ANNs) [5] . Several researchers have utilized various approaches for forecasting time series data. In reference [6], an LSTM model was developed to predict stock prices. The proposed method forecasts agricultural productivity using RNN, CNN, and LSTM [7] . Research has shown that machine learning is employed to predict B

FAQs

What is the main goal of the study for EUR/USD trading?

The study aims to predict the most advantageous times to buy and sell the euro against the dollar using machine-learning methods.

Which machine-learning algorithms are evaluated in the work?

The work evaluates AdaBoost, logistic regression, gradient boosting, random forest classifier, bagging, Gaussian naïve Bayes, extreme gradient boosting classifier, decision tree classifier, and a combined model that merges three classifiers.

How are technical indicators used to improve model performance?

Technical indicators are integrated into the training dataset so the models can learn patterns that enhance prediction precision.

Aug 4, 2026

Upload by Eliana