COMP7409 Leveraging Clustering and Twitter Sentiment Analysis for Machine Learning-Driven Trading Strategies with Deep Learning Approach

In modern financial markets, integrating social media sentiment with advanced clustering techniques has emerged as a transformative approach to trading strategies. This study explores the fusion of Twitter sentiment analysis and clustering-driven insights, leveraging deep learning to identify patterns and predict market movements. By analyzing engagement metrics and technical indicators, we group stocks into clusters that reflect shared characteristics, such as volatility and momentum. Simultaneously, Twitter sentiment provides real-time market sentiment signals. Combining these approaches, we develop machine learning-driven portfolios optimized for performance and adaptability, offering a novel framework for decision-making in complex, dynamic trading environments. This is a group project, and the project report is not available for public viewing. Part of code is available in the GitHub repository. The result of the prediction is not accurate and not suitable for real-world trading. This is just a study project.