statsArb
Several implementations of statistical arbitrage strategies.
- Model Returns residuals as an OU process with fixed entry and exits detailed in "Statistical Arbitrage in the U.S. Equities Market" by Marco Avellaneda and Jeong-Hyun Lee 2008, defactoring can be either OLS or Ridge Regressions
- defactoring single stock returns with the same industry ETF (which holds the stock as well) returns
- defactoring single stock returns with 15 PCA decomposed market eigen-portfolios
- defactoring with 5 factor ETFs (small-cap, value, low volatility, growth and momentum)
- Model Mean-reversal of mutual information as OU process
- SVM or XGB classifer of widened spread reversal
- Logistic Regression on rather trade or not