Beijing Institute of Mathematical Sciences and Applications

This course provides a rigorous, hands-on treatment of modern quantitative investing through the lens of machine learning and statistical computing. It bridges the gap between theoretical financial engineering and high-dimensional, data-driven alpha generation. Students will explore how classical financial economic frameworks are modernized, expanded, and challenged by advanced machine learning architectures. The primary focus is navigating the unique challenges of financial time series: low signal-to-noise ratios, non-stationarity, regime shifts, and market dynamics. Key topics include factor model construction and estimation, mean-variance optimization, risk modeling (including BARRA-style frameworks), rigorous backtesting methodologies, transaction cost modeling, performance attribution, and dynamic portfolio optimization. Working with industry-standard financial databases—such as CRSP, Compustat, and TAQ—students will design, implement, and backtest an original quantitative strategy.