In multivariate statistics , if
ε
{\displaystyle \varepsilon }
vector of
n
{\displaystyle n}
random variables , and
Λ
{\displaystyle \Lambda }
n
{\displaystyle n}
symmetric matrix , then the scalar quantity
ε
T
Λ
ε
{\displaystyle \varepsilon ^{T}\Lambda \varepsilon }
quadratic form in
ε
{\displaystyle \varepsilon }
It can be shown that[1]
E
[
ε
T
Λ
ε
]
=
tr
[
Λ
Σ
]
+
μ
T
Λ
μ
{\displaystyle \operatorname {E} \left[\varepsilon ^{T}\Lambda \varepsilon \right]=\operatorname {tr} \left[\Lambda \Sigma \right]+\mu ^{T}\Lambda \mu }
where
μ
{\displaystyle \mu }
Σ
{\displaystyle \Sigma }
expected value and variance-covariance matrix of
ε
{\displaystyle \varepsilon }
trace of a matrix. This result only depends on the existence of
μ
{\displaystyle \mu }
Σ
{\displaystyle \Sigma }
normality of
ε
{\displaystyle \varepsilon }
not required.
A book treatment of the topic of quadratic forms in random variables is that of Mathai and Provost.[2]
Since the quadratic form is a scalar quantity,
ε
T
Λ
ε
=
tr
(
ε
T
Λ
ε
)
{\displaystyle \varepsilon ^{T}\Lambda \varepsilon =\operatorname {tr} (\varepsilon ^{T}\Lambda \varepsilon )}
Next, by the cyclic property of the trace operator,
E
[
tr
(
ε
T
Λ
ε
)
]
=
E
[
tr
(
Λ
ε
ε
T
)
]
.
{\displaystyle \operatorname {E} [\operatorname {tr} (\varepsilon ^{T}\Lambda \varepsilon )]=\operatorname {E} [\operatorname {tr} (\Lambda \varepsilon \varepsilon ^{T})].}
Since the trace operator is a linear combination of the components of the matrix, it therefore follows from the linearity of the expectation operator that
E
[
tr
(
Λ
ε
ε
T
)
]
=
tr
(
Λ
E
(
ε
ε
T
)
)
.
{\displaystyle \operatorname {E} [\operatorname {tr} (\Lambda \varepsilon \varepsilon ^{T})]=\operatorname {tr} (\Lambda \operatorname {E} (\varepsilon \varepsilon ^{T})).}
A standard property of variances then tells us that this is
tr
(
Λ
(
Σ
+
μ
μ
T
)
)
.
{\displaystyle \operatorname {tr} (\Lambda (\Sigma +\mu \mu ^{T})).}
Applying the cyclic property of the trace operator again, we get
tr
(
Λ
Σ
)
+
tr
(
Λ
μ
μ
T
)
=
tr
(
Λ
Σ
)
+
tr
(
μ
T
Λ
μ
)
=
tr
(
Λ
Σ
)
+
μ
T
Λ
μ
.
{\displaystyle \operatorname {tr} (\Lambda \Sigma )+\operatorname {tr} (\Lambda \mu \mu ^{T})=\operatorname {tr} (\Lambda \Sigma )+\operatorname {tr} (\mu ^{T}\Lambda \mu )=\operatorname {tr} (\Lambda \Sigma )+\mu ^{T}\Lambda \mu .}
Variance in the Gaussian case [ edit ] In general, the variance of a quadratic form depends greatly on the distribution of
ε
{\displaystyle \varepsilon }
ε
{\displaystyle \varepsilon }
does follow a multivariate normal distribution, the variance of the quadratic form becomes particularly tractable. Assume for the moment that
Λ
{\displaystyle \Lambda }
var
[
ε
T
Λ
ε
]
=
2
tr
[
Λ
Σ
Λ
Σ
]
+
4
μ
T
Λ
Σ
Λ
μ
{\displaystyle \operatorname {var} \left[\varepsilon ^{T}\Lambda \varepsilon \right]=2\operatorname {tr} \left[\Lambda \Sigma \Lambda \Sigma \right]+4\mu ^{T}\Lambda \Sigma \Lambda \mu }
[3] In fact, this can be generalized to find the covariance between two quadratic forms on the same
ε
{\displaystyle \varepsilon }
Λ
1
{\displaystyle \Lambda _{1}}
Λ
2
{\displaystyle \Lambda _{2}}
cov
[
ε
T
Λ
1
ε
,
ε
T
Λ
2
ε
]
=
2
tr
[
Λ
1
Σ
Λ
2
Σ
]
+
4
μ
T
Λ
1
Σ
Λ
2
μ
{\displaystyle \operatorname {cov} \left[\varepsilon ^{T}\Lambda _{1}\varepsilon ,\varepsilon ^{T}\Lambda _{2}\varepsilon \right]=2\operatorname {tr} \left[\Lambda _{1}\Sigma \Lambda _{2}\Sigma \right]+4\mu ^{T}\Lambda _{1}\Sigma \Lambda _{2}\mu }
[4] In addition, a quadratic form such as this follows a generalized chi-squared distribution .
Computing the variance in the non-symmetric case [ edit ] The case for general
Λ
{\displaystyle \Lambda }
ε
T
Λ
T
ε
=
ε
T
Λ
ε
{\displaystyle \varepsilon ^{T}\Lambda ^{T}\varepsilon =\varepsilon ^{T}\Lambda \varepsilon }
so
ε
T
Λ
~
ε
=
ε
T
(
Λ
+
Λ
T
)
ε
/
2
{\displaystyle \varepsilon ^{T}{\tilde {\Lambda }}\varepsilon =\varepsilon ^{T}\left(\Lambda +\Lambda ^{T}\right)\varepsilon /2}
is a quadratic form in the symmetric matrix
Λ
~
=
(
Λ
+
Λ
T
)
/
2
{\displaystyle {\tilde {\Lambda }}=\left(\Lambda +\Lambda ^{T}\right)/2}
Λ
{\displaystyle \Lambda }
Λ
~
{\displaystyle {\tilde {\Lambda }}}
In the setting where one has a set of observations
y
{\displaystyle y}
operator matrix
H
{\displaystyle H}
residual sum of squares can be written as a quadratic form in
y
{\displaystyle y}
RSS
=
y
T
(
I
−
H
)
T
(
I
−
H
)
y
.
{\displaystyle {\textrm {RSS}}=y^{T}(I-H)^{T}(I-H)y.}
For procedures where the matrix
H
{\displaystyle H}
symmetric and idempotent , and the errors are Gaussian with covariance matrix
σ
2
I
{\displaystyle \sigma ^{2}I}
RSS
/
σ
2
{\displaystyle {\textrm {RSS}}/\sigma ^{2}}
chi-squared distribution with
k
{\displaystyle k}
λ
{\displaystyle \lambda }
k
=
tr
[
(
I
−
H
)
T
(
I
−
H
)
]
{\displaystyle k=\operatorname {tr} \left[(I-H)^{T}(I-H)\right]}
λ
=
μ
T
(
I
−
H
)
T
(
I
−
H
)
μ
/
2
{\displaystyle \lambda =\mu ^{T}(I-H)^{T}(I-H)\mu /2}
may be found by matching the first two central moments of a noncentral chi-squared random variable to the expressions given in the first two sections. If
H
y
{\displaystyle Hy}
μ
{\displaystyle \mu }
bias , then the noncentrality
λ
{\displaystyle \lambda }
RSS
/
σ
2
{\displaystyle {\textrm {RSS}}/\sigma ^{2}}
^ Bates, Douglas. "Quadratic Forms of Random Variables" (PDF) . STAT 849 lectures . Retrieved August 21, 2011 . ^ Mathai, A. M. & Provost, Serge B. (1992). Quadratic Forms in Random Variables . CRC Press. p. 424. ISBN 978-0824786915 ^ Rencher, Alvin C.; Schaalje, G. Bruce. (2008). Linear models in statistics (2nd ed.). Hoboken, N.J.: Wiley-Interscience. ISBN 9780471754985 OCLC 212120778 . ^ Graybill, Franklin A. Matrices with applications in statistics (2. ed.). Wadsworth: Belmont, Calif. p. 367. ISBN 0534980384
![{\displaystyle \operatorname {E} \left[\varepsilon ^{T}\Lambda \varepsilon \right]=\operatorname {tr} \left[\Lambda \Sigma \right]+\mu ^{T}\Lambda \mu }](https://wikimedia.org/api/rest_v1/media/math/render/svg/baad183f5bdae8ceea0ab20ebb804d7767187c36)
![{\displaystyle \operatorname {E} [\operatorname {tr} (\varepsilon ^{T}\Lambda \varepsilon )]=\operatorname {E} [\operatorname {tr} (\Lambda \varepsilon \varepsilon ^{T})].}](https://wikimedia.org/api/rest_v1/media/math/render/svg/75d12897773f98f644462e4aad2cbeee4a24e538)
![{\displaystyle \operatorname {E} [\operatorname {tr} (\Lambda \varepsilon \varepsilon ^{T})]=\operatorname {tr} (\Lambda \operatorname {E} (\varepsilon \varepsilon ^{T})).}](https://wikimedia.org/api/rest_v1/media/math/render/svg/a125f6229a870a6ab67fbc927e076fb256a5d4d2)
![{\displaystyle \operatorname {var} \left[\varepsilon ^{T}\Lambda \varepsilon \right]=2\operatorname {tr} \left[\Lambda \Sigma \Lambda \Sigma \right]+4\mu ^{T}\Lambda \Sigma \Lambda \mu }](https://wikimedia.org/api/rest_v1/media/math/render/svg/c9ef7078cfbe86a82e759e2caa85cbe9baaa6e54)
![{\displaystyle \operatorname {cov} \left[\varepsilon ^{T}\Lambda _{1}\varepsilon ,\varepsilon ^{T}\Lambda _{2}\varepsilon \right]=2\operatorname {tr} \left[\Lambda _{1}\Sigma \Lambda _{2}\Sigma \right]+4\mu ^{T}\Lambda _{1}\Sigma \Lambda _{2}\mu }](https://wikimedia.org/api/rest_v1/media/math/render/svg/549bc7cb9f3da34683dadfff4e73d8cb41110989)
![{\displaystyle k=\operatorname {tr} \left[(I-H)^{T}(I-H)\right]}](https://wikimedia.org/api/rest_v1/media/math/render/svg/96fc47f619bea6a72ba7b8b4bb0bc32bd436c4b6)
