In this project I aim to improve estimation of the higher order comoment matrices currently implemented in the R packages PerformanceAnalytics and PortfolioAnalytics in three ways. First, the implementation of the existing estimators will be optimised, resulting in a significant speed gain. This will make the methods usable for higher dimensions, as required by many practitioners. Second, I will add a full range of state-of-the-art higher order comoment estimators, currently not available for the community. Since these estimators greatly improve the accuracy of estimating the skewness and kurtosis of linear combinations of random variables, all functions in PerformanceAnalytics and PortfolioAnalytics depending on these moments will be more reliable. This is in particular important for risk or performance measures depending on higher order moments in the package PerformanceAnalytics, as well as certain objective functions or constraints in the package PortfolioAnalytics.

Student

cdries

Mentors

  • Brian Peterson
  • kboudt
  • Peter Carl
close

2017