Wolfram Research

Function Repository Resource:

MultivariateKurtosis

Source Notebook

Find the kurtosis coefficient for multivariate data

Contributed by: Wolfram Research

ResourceFunction["MultivariateKurtosis"][matrix]

gives a multivariate kurtosis coefficient for matrix.

Details and Options

ResourceFunction["MultivariateKurtosis"] is a univariate measure of kurtosis for multivariate data.
ResourceFunction["MultivariateKurtosis"][matrix] is equivalent to where matrix={x1,x2,,xn}, Mean[matrix] and Σ is the estimated population covariance matrix.
For a matrix with p columns, a value of the multivariate kurtosis coefficient close to p (p+2) indicates approximate multinormality.

Examples

Basic Examples

Multivariate kurtosis for bivariate data:

In[1]:=
ResourceFunction[
 "MultivariateKurtosis"][{{1, 3}, {3, 4}, {5, 7}, {8, 10}}]
Out[1]=

Requirements

Wolfram Language 11.3 (March 2018) or above

Resource History

License Information