Stock Volatility Prediction

My interest in finance led me to this project - I wanted to understand if machine learning could do better than traditional models at predicting stock volatility. Volatility is crucial for options pricing, risk management, and just understanding market dynamics. I built models using both Neural Networks and SVMs, testing them across seven different market sectors. The approach combined the classic Black-Scholes framework with modern ML techniques. We saw a 31% improvement in prediction accuracy compared to baseline models. What I learned is that different sectors have different volatility patterns - tech stocks behave differently than utilities, for example. The project gave me a solid foundation in quantitative finance and reinforced my interest in the intersection of AI and markets.