SHORT BIO RESEARCH TEACHING
 
 
Robust localisation and scatter estimates

Robust estimation

Explanations

Compute classical statistical estimators (mean and covariance) and robust alternatives. The robust estimators take into account that outliers/noise could be observed in the data set and try to:
  • Reduce the impact of the outliers (try to "clean" the data set before computing the estimators)
  • Detect the outliers or problems with the data set
More details:
Research paper about this robust algorithm (RelaxMCD)
An application to financial portfolio selection

Generation of a data set

Number of observations:
(max 200)
Dimension:
Include point contaminated observations:
...or type data in the area below (dim separator: space, observation separator: newline, decimal: dot)

Coverage:
(Coverage=number of observations - max number of contaminated ones)

 
 
  Michaël Schyns - Management Information Systems
Useful links: University of Liège - HEC-Management School - QuantOM - Chaudfontaine
Email: M.Schyns