random vector


A random vector is a finite-dimensional formal vector of random variablesMathworldPlanetmath. The random vector can be written either as a column or row of random variables, depending on its context and use. So if X1,X2,โ€ฆ,Xn are random variables, then

๐—=(X1X2โ‹ฎXn)=(X1,X2,โ€ฆ,Xn)T

is a random (column) vector. Similarly, one defines a random matrix to be a formal matrix whose entries are all random variables. The size of a random vector and the size of a random matrix are assumed to be finite fixed constants.

The distributionPlanetmathPlanetmath of a random vector ๐—=(X1,X2,โ€ฆ,Xn) is defined to be the joint distributionPlanetmathPlanetmath of its coordinatesPlanetmathPlanetmath X1,โ€ฆ,Xn:

F๐—โข(๐ฑ):=FX1,โ€ฆ,Xnโข(x1,โ€ฆ,xn).

Similarly, the distribution of a random matrix is the joint distribution of its matrix components.

Let ๐—=(X1,X2,โ€ฆ,Xn) be a random vector. If Eโก[Xi] exists (<โˆž) for each i, then the expectation of X, called the mean vector and denoted by ๐„โข[๐—], is defined to be:

๐„โข[๐—]:=(Eโก[X1],Eโก[X2],โ€ฆ,Eโก[Xn]).

Clearly ๐„โข[๐—]T=๐„โข[๐—T]. The expectation of a random matrix is similarly defined. Note that the definitions of expectations can also be defined via measure theory. Then, using Fubiniโ€™s Theorem, one can show that the two sets of definitions coincide.

Again, let ๐—=(X1,X2,โ€ฆ,Xn)T be a random vector. If ๐=๐„โข[๐—] is defined and Eโก[XiโขXj] are defined for all 1โ‰คi,jโ‰คn, then the varianceMathworldPlanetmath of X, denoted by ๐•๐š๐ซโข[๐—], is defined to be:

๐•๐š๐ซโข[๐—]:=๐„โข[(๐—-๐)โข(๐—-๐)T].

It is not hard to see that ๐•๐š๐ซโข[๐—] is an nร—n symmetric matrixMathworldPlanetmath and it is equal to the covariance matrixMathworldPlanetmath of the Xiโ€™s.

:

  1. 1.

    If X is an n-dimensional random vector with A a mร—n constant matrix and ๐œถ an m-dimensional constant vector, then

    ๐„โข[๐€๐—+๐œถ]=๐€๐„โข[๐—]+๐œถ.
  2. 2.

    Same set up as above. Then

    ๐•๐š๐ซโข[๐€๐—+๐œถ]=๐€๐•๐š๐ซโข[๐—]โข๐€T.

    If the Xiโ€™s are iid (independent identically distributed), with variance ๐ˆ2, then

    ๐•๐š๐ซโข[๐€๐—+๐œถ]=๐ˆ2โข๐€๐€T.
  3. 3.

    Let ๐— be an n-dimensional random vector with ๐=๐„โข[๐—], ๐šบ=๐•๐š๐ซโข[๐—]. ๐€ is an nร—n constant matrix. Then

    ๐„โข[๐—Tโข๐€๐—]=trโก(๐€โข๐šบ)+๐Tโข๐€โข๐.
Title random vector
Canonical name RandomVector
Date of creation 2013-03-22 14:27:20
Last modified on 2013-03-22 14:27:20
Owner CWoo (3771)
Last modified by CWoo (3771)
Numerical id 15
Author CWoo (3771)
Entry type Definition
Classification msc 62H99
Classification msc 15A52
Defines random matrix
Defines distribution of a random vector
Defines distribution of a random matrix
Defines mean vector