Research & Notebooks
Academic research, technical analysis, and interactive notebooks exploring quantitative finance topics and methodologies.
Dynamic Hedge Ratio Estimation Using Kalman Filtering in Pairs Trading
This paper explores the application of Kalman Filter for dynamic hedge ratio estimation in pairs trading strategies. We demonstrate improved performance over static OLS regression by allowing the hedge ratio to adapt to changing market conditions. Backtesting on US equity pairs shows a 23% improvement in Sharpe ratio compared to traditional cointegration-based approaches.
Machine Learning Approaches to Factor Timing in Equity Markets
We investigate the use of gradient boosting and neural network models for timing exposure to traditional Fama-French factors. Our analysis suggests that while factors exhibit time-varying premia, out-of-sample predictability remains limited. We propose a regime-switching framework that improves risk-adjusted returns by 15% over static factor exposure.
Deep Learning for American Option Pricing: A Comparative Study
This study compares traditional numerical methods (binomial trees, finite differences) with deep learning approaches for American option pricing. We find that neural network-based methods achieve comparable accuracy with significantly reduced computation time, making them suitable for real-time pricing applications.