Impact of Social Media Sentiments on Stock Market Behavior: A Machine Learning Approach to Analyzing Market Dynamics

Social media has become a valuable tool for informed decision-making. This research delves into the influence of Twitter sentiments on the stock market’s movements and price fluctuations, specifically focusing on Tesla Inc and the tweets of Elon Musk. A combination of deductive and inductive reasoning approaches is used to explore the intricate relationship between the social media platform and the stock market. Methodologically, the Twitter data undergoes rigorous processing to derive features for the machine learning predictive model, and the sentiments are extracted using the Valence Aware Dictionary and Sentiment Reasoner (VADER) tool. This study emphasizes the usefulness of social media in predictive modeling while underscoring the importance of evaluating data reliability considering challenges such as spam tweets and geographical relevance. Multiple machine learning models are tested against four distinct datasets addressing the high stock price volatility. XG Boost and Random Forest Regressor emerge as the most effective performers, particularly when moving averages are included, showing enhanced performance. This research establishes an evident correlation between social media sentiments and stock market movements, however with limited predicting power. It is also noted that integrating traditional financial metrics enriches the understanding of stock market dynamics while enhancing the model’s predictability.