- Leveraged stock market data from yfinance package and a provided database to create Markowitz minimum variance portfolios using the Factor Graphical Lasso (FGL) Algorithm suggested by Lee and Seregina
- FGL utilizes graphical methods for inverse covariance (precision) matrix estimation
- Created a 6 year rolling window trading strategy with yearly updates to portfolio weights with appropriate transaction fees
- The resulting portfolio outperformed the S&P500 and a balanced (equally weighted) portfolio in returns and volatility
- Portfolio Returns are seen below:
