On a Transformation and Re-transformation Technique for Constructing an Affine Equivariant Multivariate Median
نویسندگان
چکیده
An affine equivariant version of multivariate median is introduced. The proposed median is easy to compute and has some appealing geometric features that are related to the configuration of a multivariate data cloud. The transformation and re-transformation approach used in the construction of the median has some fundamental connection with the data driven co-ordinate system considered by Chaudhuri and Sengupta (1993, Journal of the American Statistical Association). Large sample statistical properties of the median are discussed and finite sample performance is investigated using Monte Carlo simulations.
منابع مشابه
On Strong Invariant Coordinate System (SICS) Functionals
In modern multivariate statistical analysis, affine invariance or equivariance for statistical procedures are properties of paramount interest and importance. A statistical procedure lacking such a property can sometimes acquire it if carried out not on the original data but rather on suitably transformed data, in some cases accompanied by a retransformation back to the original coordinate syst...
متن کاملSpatial Trimming, with Applications to Robustify Sample Spatial Quantile and Outlyingness Functions, and to Construct a New Robust Scatter Estimator
The spatial multivariate median has a long history as an alternative to the sample mean. Its transformation-retransformation (TR) sample version is affine equivariant, highly robust, and computationally easy. More recently, an entire TR spatial multivariate quantile function has been developed and applied in practice along with related rank functions. However, as quantile levels move farther ou...
متن کاملEquivariance and Invariance Properties of Multivariate Quantile and Related Functions, and the Role of Standardization
Equivariance and invariance issues arise as a fundamental but often problematic aspect of multivariate statistical analysis. For multivariate quantile and related functions, we provide coherent definitions of these properties. For standardization of multivariate data to produce equivariance or invariance of procedures, three important types of matrix-valued functional are studied: “weak covaria...
متن کاملTR Multivariate Conditional Median Estimation
An affine equivariant version of the nonparametric spatial conditional median (SCM) is constructed, using an adaptive transformation-retransformation (TR) procedure. The relative performance of SCM estimates, computed with and without applying the TR—procedure, are compared through simulation. Also included is the vector of coordinate conditional, kernelbased, medians (VCCMs). The methodology i...
متن کاملOutlier detection for compositional data using robust methods
Outlier detection based on the Mahalanobis distance (MD) requires an appropriate transformation in case of compositional data. For the family of logratio transformations (additive, centered and isometric logratio transformation) it is shown that the MDs based on classical estimates are invariant to these transformations, and that the MDs based on affine equivariant estimators of location and co...
متن کامل