Forecasting Egyptian Stock Market Trends: A Comparative Framework Using Machine Learning, Time Series Analysis, and Deep Learning

Forecasting stock prices is challenging, especially in volatile and data-constrained markets such as Egypt. This complexity is driven by economic fluctuations, limited historical data, and market-specific irregularities, which often hinder the predictive performance of traditional forecasting models. Existing research frequently overlooks the unique temporal patterns and local dynamics of the Egyptian Exchange (EGX30). To bridge this gap, this study proposes a comprehensive forecasting framework that integrates classical machine learning, time series analysis, and deep learning techniques. Using a tailored multilingual dataset covering ten major EGX30-listed companies, the framework incorporates meticulous data cleaning, structured preprocessing, and domain-specific feature engineering to enhance model reliability. Model performance was assessed using standard regression metrics, with results showing that the ARIMA model consistently delivered the most accurate forecasts, outperforming both traditional and deep learning approaches. These findings emphasize the value of classical time series models in emerging markets and their potential in supporting data-driven investment decisions.