The geometric median (also called spatial median or L1 median) is a robust multivariate indicator of central position. This library provides fast estimation procedures that can handle rapidly large samples of high dimensional data. Function Gmedian computes the geometric median of a numerical data set with averaged stochastic gradient algorithms, whereas GmedianCov computes the median covariation matrix, a useful indicator for robust PCA. Robust clustering, based on the geometric k-medians, can also be performed with the same type of recursive algorithm thanks to kGmedian. Less fast estimation procedures based on Weiszfeld's algorithm are also available : function Weiszfeld computes the geometric median whereas WeiszfeldCov computes the median covariation matrix. These procedures may be preferred for small and moderate sample sizes. Note that weighting statistical units (for example with survey sampling weights) is allowed.
● Data Source:
CranContrib
● Keywords: Gmedian
● Alias: Gmedian-package
●
0 images
|
Fast k-medians clustering based on recursive averaged stochastic gradient algorithms. The procedure is similar to the kmeans clustering technique performed recursively with the MacQueen algorithm. The advantage of the kGmedian algorithm compared to MacQueen strategy is that it deals with sum of norms instead of sum of squared norms, ensuring a more robust behaviour against outlying values.
● Data Source:
CranContrib
● Keywords: Gmedian
● Alias: kGmedian
●
1 images
|
Computes recursively the Geometric median (also named spatial median or L1-median) with a fast averaged stochastic gradient algorithms that can deal rapidly with large samples of high dimensional data.
● Data Source:
CranContrib
● Keywords: Gmedian
● Alias: Gmedian
●
1 images
|
Computes recursively the Geometric median and the (geometric) median covariation matrix with fast averaged stochastic gradient algorithms. The estimation of the Geometric median is performed first and then the median covariation matrix is estimated, as well as its leading eigenvectors. The original recursive estimator of the median covariation matrix may not be a non negative matrix. A fast projected estimator onto the convex closed cone of the non negative matrices allows to get a non negative solution.
● Data Source:
CranContrib
● Keywords: Gmedian
● Alias: GmedianCov
●
1 images
|
Computes the Geometric median (also named spatial median or L1-median) with Weiszfeld's algorithm.
● Data Source:
CranContrib
● Keywords: Weiszfeld
● Alias: Weiszfeld
●
4 images
|
Estimation of the Geometric median covariation matrix with Weiszfeld's algorithm. Weights (such as sampling weights) for statistical units are allowed.
● Data Source:
CranContrib
● Keywords: WeiszfeldCov
● Alias: WeiszfeldCov
●
1 images
|
providing 
.