An Explainable Hybrid Artificial Intelligence Framework for Stock Market Prediction Using Financial Time-Series and Market Sentiment

This dissertation presents an explainable hybrid artificial intelligence framework for stock market prediction using financial time-series information and market sentiment. The study investigates the integration of historical market observations, technical indicators, sequential deep-learning models, financial sentiment analysis and explainable artificial intelligence within a unified forecasting framework. The methodology addresses data preprocessing, temporal alignment, feature engineering, model development, multimodal feature integration and evaluation. Particular emphasis is placed on preventing information leakage through chronological data handling and on improving the interpretability of model predictions. The dissertation provides a structured research framework for investigating whether financial sentiment provides complementary information to numerical market features in stock-market forecasting.