Hybrid machine learning for stock price prediction in the Moroccan banking sector - Slideshare

Aug 3, 2026

Upload by Violet

Hybrid machine learning is applied to stock price prediction in Morocco’s banking sector using historical market data from Bank of Africa. The study optimizes a simulated prediction model with machine-learning regression algorithms and feature-selection methods, comparing linear regression, extreme gradient boosting, random forest, L2-regularized least squares, and support vector regression variants. Results show that hybrid feature-selection and model combinations based on top score percentiles, univariate linear regression, and linear SVR improve performance under RMSE and R2-Score metrics. Feature reduction addresses high-dimensionality issues and enhances accuracy.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

1. INTRODUCTION

Machine learning for stock forecasting

Automated trading with regression

Motivation and challenges

Regression performance advantages

Hybrid machine learning for stock price prediction in the Moroccan banking sector Bouzgarne Itri1, Youssfi Mohamed1, Bouattane Omar1, El Madani Latifa1, Moumoun Lahcen2, Oualid Adil2 1Laboratory Computer Science, Artificial Intelligence and Cyber Security (2IACS), ENSET Mohammedia, Hassan II University of Casablanca, Casablanca, Morocco 2Laboratory Mathématiques, Informatique et Sciences de l’Ingénieur (MISI), FST Settat , Hassan 1st University in Settat, Settat, Morocco | Article history:
Received Jun 19, 2023 Revised Nov 11, 2023 Accepted Jan 18, 2024 | Analyzing historical stock market data using machine-learning techniques is crucial for data scientists and researchers to optimize stock price prediction models. This study uses machine learning regression algorithms and featureselection methods to optimize a simulated stock price prediction model using real historical data from Bank of Africa, a Moroccan bank. The approach compares multiple supervised regression algorithms, such as linear regression, extreme gradient boosting, ordinary least squared, random forest regressor, a linear least-squares L2-regularized, epsilon-support vector regression, and linear support vector regression. Each of these algorithms is associated with different feature selection algorithms to improve the performance of the prediction model. The analysis results revealed that hybridizing algorithms between the highest score percentiles, univariate linear regression, and linear support vector regression perform better according to the root mean squared error and R2-Score measures. This approach overcomes the problems associated with high-dimensional data by reducing the number of features and improving prediction accuracy.
This is an open access article under the CC BY-SA license.
| | --- | --- | | Keywords:
Banking stock market Feature selection Linear support vector regression
Machine learning Regression algorithm | | Corresponding Author: Bouzgarne Itri Laboratory Computer Science, Artificial Intelligence and Cyber Security (2IACS), ENSET Mohammedia, Hassan II University of Casablanca Casablanca, Morocco [Email: bouzgarne.itri@gmail.com](Email: bouzgarne.itri@gmail.com) Article Info ABSTRACT 1. INTRODUCTION Many investors place their money in the stock market in the hopes of achieving high returns. They may use various methods, such as news articles or moving averages, to understand market trends and make trading decisions. It is also important to learn the fundamentals that will help build a long-term investment strategy. However, it is difficult to achieve good returns due to the non-linearity and randomness of the stock market. The deteriorating economic climate, due in part to inflation, should encourage beginner investors in the stock market to prepare for a potential recession and to educate themselves in order to create a portfolio suited for all market situations. To accurately predict trends, many investors turn to machine learning methods, which allow for a more in-depth analysis of the many external factors influencing stock prices, such as policy, economy, social issues, and investment sentiment [1] . Automated trading based using regression methods is one of the machine learning techniques that involve using historical market data to train a machine learning model to predict future price movements [2] . These predictions are then used to automatically execute trades or provide indicators that can help with decision-making. A multitude of researchers has employed diverse machine learning models to forecast stock market trends, yielding promising outcomes, such as support vector regression (SVR) [3] and random forecast (RF) [4] and multiple algorithms in comparison: random forest, extreme gradient boosting (XGBoost), adaptive boosting (AdaBoost), support vector machine regression, k-nearest neighbors (KNN), and artificial neural network (ANN) [5] . Machine learning regression techn

FAQs

What problem does the study address?

The study addresses the difficulty of predicting stock price movements due to the stock market’s non-linearity and randomness, aiming to improve prediction accuracy using hybrid machine learning methods.

Which modeling approach and algorithms are compared?

The study uses regression algorithms and feature-selection techniques, comparing methods such as linear regression, extreme gradient boosting (XGBoost), random forest, L2-regularized least squares, and support vector regression variants (including epsilon-SVR and linear SVR).

What did the results show about the best-performing hybrid approach?

The analysis found that hybridizing algorithms based on the highest score percentiles with univariate linear regression and linear support vector regression performed better, evaluated using RMSE and R2-Score metrics, while feature reduction helped manage high-dimensional data.

Aug 3, 2026

Upload by Violet