Weekly maintenance every saturday 03:00 - 06:00 UTC
- Mean historical return.
- Exponentially weighted mean historical return.
- CAPM estimate of returns.
- Fix non-positive semidefinite matrices.
- Sample covariance.
- Semicovariance.
- Exponentially weighted covariance.
- Shrunk covariance matrices:
- manual shrinkage
- Ledoit Wolf shrinkage
- Oracle Approximating shrinkage
Mean-Variance optimization.
- minimum of volatility
- maximum of Sharpe ratio
- maximum of the quadratic utility, given some risk aversion
- maximum of return for a given target risk
- minimum of risk for a given target return
Semivariance optimization.
- minimum of semivariance
- maximum of the quadratic utility, given some risk aversion
- maximum of return for a given target risk
CVaR and CDaR optimization.
- minimum of CVaR or CDaR
- maximum of return for a given CVaR or CDaR
- minimum of CVaR or CDaR for a given target return
- All of these methods support L2 regularization also.
To start the portfolio optimization refer to the Quickstart page.
If you have a questions or bug reports, feel free to contact me either with contact email or redirecting by the link "Ask a question", which is located on the left sidebar (but you should register on github for this).