Enhancing Stock Market Prediction with LSTM-based Ensemble Models and Attention Mechanism

Abstract

Forecasting the stock market remains a difficult task due to the influence of several factors, including financial performance and market sentiment. To address this problem, we propose a new research approach that combines LSTM-based ensemble models with attention mechanisms for enhanced prediction accuracy and performance. Additionally, we leverage sentiment analysis and meta-learning techniques to further refine our predictions. The proposed methodology is effective in capturing market sentiment and effectively modeling the complex dynamics of stocks, achieving high forecasting accuracy and outperforming traditional LSTM models.