A Hybrid Model for GAFAM – Stock price Movement Forecasting and Correlation

This research introduces a model for predicting stock prices of major technology companies—Google (Alphabet), Apple, Facebook (Meta), Amazon, and Microsoft (GAFAM)—by integrating financial data and sentiment analysis. The model comprises three components: (1) a Long Short-Term Memory (LSTM) network for forecasting stock prices using historical financial data, (2) a sentiment analysis module leveraging the RoBERTa Transformer-based model to evaluate news article sentiment and incorporate it into predictions, and (3) a correlation analysis to compute a matrix that identifies interdependencies among the GAFAM companies' stock prices. The proposed model predicts stock prices for the next five days based on the provided dataset, achieving an average accuracy of 64.8%. It evaluates the performance within a tolerance range of 1.5% to 1.8%, accounting for inherent market volatility, and highlights variability in prediction accuracy across different companies. Incorporating real-time data through APIs ensures the model remains current and adaptive to changing market conditions. The proposed approach not only forecasts individual stock prices but also examines the impact of market sentiment on price movements and explores the interconnections among GAFAM stocks. By combining deep learning for time-series forecasting with advanced sentiment analysis, the model offers a robust tool for investors and analysts, supporting smarter investment strategies and promoting economic growth aligned with Sustainable Development Goal 8 (SDG-8).