Abstract: Stock market prediction is a challenging task due to the market’s complexity and volatility. Recent literature has turned to sentiment analysis – extracting opinions or emotions from news and social media – as a complementary signal for forecasting stock movements. This survey reviews existing approaches that combine sentiment analysis with machine learning to predict stock prices or trends. We outline the spectrum of sentiment analysis techniques (from lexicon-based to deep learning-based methods) and the variety of predictive models (regression, SVM, neural networks, etc.) employed. We summarize key findings from prior studies, which largely indicate that incorporating sentiment features can improve predictive accuracy ([1010.3003] Twitter mood predicts the stock market) ([The impact of microblogging data for stock market prediction: Using ...]), while also highlighting inconsistencies and mixed results. The survey further discusses practical deployments in industry – including hedge funds and financial data services leveraging sentiment – and examines the persistent challenges (noisy data, alignment of sentiment signals with price movements, market non-stationarity, and model interpretability) that limit performance. We conclude by identifying gaps and suggesting future research directions to develop more robust, interpretable, and effective sentiment-enhanced stock prediction models.

Keywords: Stock Market Prediction; Sentiment Analysis; Machine Learning; Natural Language Processing; Financial News; Social Media; Deep Learning; Predictive Modeling.

| DOI: 10.17148/IARJSET.2025.12225