Scatter matrix explained

For the notion in quantum mechanics, see scattering matrix.

In multivariate statistics and probability theory, the scatter matrix is a statistic that is used to make estimates of the covariance matrix, for instance of the multivariate normal distribution.

Definition

Given n samples of m-dimensional data, represented as the m-by-n matrix,

X=[x1,x2,\ldots,xn]

, the sample mean is

\overline{x

} = \frac\sum_^n \mathbf_j

where

xj

is the j-th column of

X

.[1]

The scatter matrix is the m-by-m positive semi-definite matrix

S=

n
\sum
j=1

(xj-\overline{x

})(\mathbf_j-\overline)^T = \sum_^n (\mathbf_j-\overline)\otimes(\mathbf_j-\overline) = \left(\sum_^n \mathbf_j \mathbf_j^T \right) - n \overline \overline^T

where

()T

denotes matrix transpose,[2] and multiplication is with regards to the outer product. The scatter matrix may be expressed more succinctly as

S=

T
XC
nX

where

Cn

is the n-by-n centering matrix.

Application

The maximum likelihood estimate, given n samples, for the covariance matrix of a multivariate normal distribution can be expressed as the normalized scatter matrix

CML=

1
n

S.

[3]

When the columns of

X

are independently sampled from a multivariate normal distribution, then

S

has a Wishart distribution.

See also

XX\top

or X⊗X is the outer product of X with itself.

Notes and References

  1. Web site: Raghavan . 2018-08-16 . Scatter matrix, Covariance and Correlation Explained . 2022-12-28 . Medium . en.
  2. Web site: Raghavan . 2018-08-16 . Scatter matrix, Covariance and Correlation Explained . 2022-12-28 . Medium . en.
  3. Liu . Zhedong . April 2019 . Robust Estimation of Scatter Matrix, Random Matrix Theory and an Application to Spectrum Sensing . Master of Science . King Abdullah University of Science and Technology.