Abstract
This study focuses on predicting stock prices in the Egyptian market, particularly the EGX30, an influential financial hub in the Middle East. While most research targets global stocks, there is a growing need to understand stock trends in developing countries like Egypt. This study compares various machine learning models for forecasting EGX30 trends, covering both short and long-term predictions.
Utilizing historical EGX30 data, the study evaluates models based on metrics such as root mean squared error, Mean Absolute Percentage Error, and coefficient of determination. The models assessed include K-Nearest Neighbours (KNN), Random Forest, eXtreme Gradient Boosting (XGBoost), Long Short-Term Memory networks (LSTM), and Gated Recurrent Unit networks (GRU).
The goal is to identify the most effective models for predicting EGX30, considering Egypt's unique market dynamics. Insights from this study aim to assist investors in making informed decisions.
Results indicate that the Gated Recurrent Unit (GRU) outperformed other models in one-week, one-month, and two-month predictions, while the eXtreme Gradient Boosting (XGBoost) model excelled in one-day predictions, highlighting their utility in predictive analysis for financial markets. The study also emphasizes the importance of ensemble techniques, particularly for long-term predictions, achieving results that were five times better than GRU for two-month forecasts.
Additionally, the surprisingly strong performance of K-Nearest Neighbours (KNN) in long-term predictions suggests its enduring relevance and potential future applications in fintech domains.
Blogger's Review: This article provides valuable insights for investment decisions in the Egyptian stock market by comparing various machine learning models, particularly emphasizing the superior performance of GRU and XGBoost. It highlights the significance of selecting appropriate models under specific market conditions, offering practical guidance for future research in fintech.