ClusterPortfolios is an R package for constructing portfolios based on statistical clustering techniques. Clustering financial asset returns and allocating capital along cluster boundaries can increase robustness, decrease sampling sensitivity, improve diversification and enhance portfolio performance.
Methods implemented in this package:
- Hierarchical Risk Parity (Lopez de Prado, 2016)
- Constrained HRP (Pfitzinger & Katzke, 2019)
- Nested Clusters Optimization (Lopez de Prade, 2019)
- Clustered Minimum Variance
- Clustered Equal Weights
- Hierarchical filters of the covariance matrix
devtools::install_github("https://github.com/jpfitzinger/ClusterPortfolios")
library(ClusterPortfolios)
data("Industry_10")
rets <- Industry_10
sigma <- cov(rets)
HRP(sigma, UB = 0.15, tau = 0.5)
Lopez de Prado, M. (2016). Building Diversified Portfolios that Outperform Out-of-Sample. SSRN Electronic Journal.
Lopez de Prado, M. (2019). A Robust Estimator of the Efficient Frontier. SSRN Electronic Journal.
Pfitzinger, J., Katzke, N. (2019). A Constrained Hierarchical Risk Parity Algorithm with Cluster-Based Capital Allocation. Stellenbosch University, Department of Economics. Working Paper 14/2019.