Professional Documents
Culture Documents
net/publication/303993136
CITATIONS READS
2 169
1 author:
Thomas Royen
Fachhochschule Bingen
25 PUBLICATIONS 293 CITATIONS
SEE PROFILE
All content following this page was uploaded by Thomas Royen on 24 June 2016.
Abstract. The p - variate gamma distribution in the sense of Krishnamoorthy and Parthasarathy exists for all
positive integer degrees of freedom and at least for all real values p 2, p 2. For special structures
of the “associated“ covariance matrix it also exists for all positive . In this paper a relation between central and
non-central multivariate gamma distributions is shown, which implies the existence of the p - variate gamma
distribution at least for all non-integer greater than the integer part of ( p 1) / 2 without any additional
assumptions for the associated covariance matrix.
1. Introduction
The p-variate chi-square distribution (or more precisely: “Wishart-chi-square distribution“) with degrees of
freedom and the “associated“ covariance matrix (the p2 ( , ) - distribution) is defined as the joint distribu-
tion of the diagonal elements of a Wp ( , ) - Wishart matrix. Its probability density (pdf) has the Laplace trans-
form (Lt)
The p - variate gamma distribution in the sense of Krishnamoorthy and Parthasarathy [4] with the associated
covariance matrix and the “degree of freedom” 2 (the p ( , ) - distribution) can be defined by the
Lt
| I p T | (1.2)
of its pdf
g ( x1,..., x p ; , ). (1.3)
For values 2 this distribution differs from the p2 ( , ) - distribution only by a scale factor 2, but in
this paper we are only interested in positive non-integer values 2 for which | I p T | is the Lt of a pdf and
not of a function also assuming any negative values. These values are called here “admissible values”. The
admissibility of all values 2 p 1 follows from the existence of the Wp (2 , ) - distribution, and the ad-
missibility of 2 ( p 2, p 1) follows from formula (1.4) below. Smaller values of are admissible at least
with some additional assumptions for . Sufficient and necessary conditions for entailing infinite divisibility
Key words and phrases: multivariate gamma distribution, non-central multivariate gamma distribution, Wishart
distribution
2010 Mathematics Subject Classifications: 60E05, 62E10
-2-
of the Lt | I p T |1 (i.e. all 0 are admissible) are found in [1] and [3]. According to [1], | I p T |1 is
infinitely divisible if and only if there exists any signature matrix S diag (s1 ,..., s p ),s j 1, for which S 1S
is an M-matrix. For the infinite divisibility of a more general class of multivariate gamma distributions see [2].
At least all values 2 m 1,m ,m p, are admissible for “ m - factorial” covariance matrices
W 2 AAT ( W W I p BBT ,B WA with rows b j ) with a suitable matrix W diag (w1 ,..., wp ),
wj 0, and a real ( p m) - matrix A of the lowest possible rank m. This follows from the representation
g ( x1 ,..., x p ; , ) E p
j 1
w2j g (w2j x j , 12 b j Sb jT (1.4)
g ( x, y) e y n0 g n ( x) n!
yn
is the non-central gamma density with the non-centrality parameter y and the central gamma densities g n ,
and the expectation refers to the Wm (2 , I m ) - Wishart matrix S . With W 2 I p , where is the lowest
eigenvalue of , it follows m p 1. A special case – entailing infinite divisibility – is a one-factorial
W 2 aaT with a real column a.
In the following section it will be shown that all values 2 [( p 1) / 2] (the integer part of ( p 1) / 2 ) are
admissible without any further assumptions for . This result is obtained as a corollary of a relation between
central and non-central multivariate gamma densities given in theorem 1. This relation was already derived in a
similar form for integer values 2 in [6], but the non-integer values require a different proof by means of
the Lt.
For 2 max( p1 1, p2 1) the Wishart Wp2 (2 , 22 ) - pdf and the non-central Wp1 (2 , 0 , ) - pdf exist
where the symmetrical positive semi-definite ( p1 p1 ) - matrix is any “non-centrality matrix” of rank k p1 .
The diagonal of a random matrix Z has a non-central p1 ( , 0 , ) - distribution if 2Z has a Wp1 (2 , 0 , ) -
distribution, which exists (apart from integer values 2 with rank() min( p1, 2 ) ) for all ( p1 1) / 2
with rank() p1 (see [5]).
-3-
g ( x1 ,..., x p1 ; , 0 , ) (2.3)
has the Lt
with T1 diag (t1 ,..., t p1 ). (In the literature the “non-centrality matrix” is frequently defined in a different non-
symmetrical way.) Let p2 denote the set of all ( p2 p2 ) - correlation matrices C and let
g ( x1 ,..., x p ; , ) ( p2 ( )) 1 | 22 |
Remark. The simple special case with p2 1,C 1 (and 22 1 ) was already given by theorem 2 in [11].
As a corollary we obtain
Theorem 2. The function g ( x1 ,..., x p ; , ) with the Lt | I p T | from (1.2) is a p ( , ) - pdf at least
for 2 ([( p 1) / 2], ).
Proof of theorem 1. The equation in theorem 1 will be verified by the Lt of both sides. The left side has the Lt
| I p T | and we get with the Schur complement for determinants
1
| I 2 22T2 | | I1 0T1 1222 21T1 12T2 ( I 2 22T2 )1 21T1 |
1
| I 2 22T2 | | I1 0T1 1222 (I 2 22T2 (I 2 22T2 )1 )21T1 |
1
| I 2 22T2 | | I1 0T1 1222 ( I 2 22T2 )1 21T1 |
1
| I1 0T1 | | I 2 22T2 | | I1 1222 ( I 2 22T2 )1 21T1 ( I1 0T1 )1 | . (2.7)
With (2.4), (2.5) and (2.6) we find for the right side the Lt
1
| I1 0T1 | | I 2 22T2 | | I1 1222 ( I 2 22T2 )1 21T1 ( I1 0T1 )1 |,
which coincides with formula (2.7), where the last identity follows from the general equation
I1 A12
| I 2 B21 A12 || I1 A12 B21 |.
B21 I2
Some further remarks. If W 2 AAT is m - factorial, where A – and consequently B WA – may con-
tain any mixture of real or pure imaginary columns, then the function with the representation from (1.4) has
again the Lt | I p T | and it is a p ( , ) - pdf at least for all values 2 m 1 [( p 1) / 2]. For
special structures of smaller values of 2 are possible, e.g. with an m - factorial W 2 AAT with a
real matrix A and m 1 [( p 1) / 2] . Furthermore, let be, e.g., p2 p1 ,0 W02 A0 A0T with a real
( p1 m0 ) - matrix A0 of rank m0 and m12 rank(12 ) p2 . Then, at least 2 max(m0 m12 1, p2 1) is
admissible and max(m0 m12 1, p2 1) [( p 1) / 2] is possible for low values of m0 and m12 .
On the other hand it is at present an open question if there exist some ( p p) - covariance matrices for
which 2 is inadmissible for some values 2 ([( p 3) / 2],[( p 1) / 2]), p 5.
for the p - variate cumulative p ( , ) - distribution function G p . This inequality was proved for 2 and
for all values 2 p 2 in [9] (see also [10]), and it implies the famous Gaussian correlation inequality for
2 1. Now, this inequality can be extended to all non-integer values 2 [( p 1) / 2] without any further
assumptions for .
References
[1] Bapat, R. B., Infinite divisibility of multivariate gamma distributions and M-matrices, Sankhyā Series A 51
(1989), 73-78.
[2] Bernardoff, P., Which multivariate gamma distributions are infinitely divisible?, Bernoulli 12 (2006),169-189.
[3] Griffiths, R. C., Characterization of infinitely divisible multivariate gamma distributions, J. Multivarate Anal.
15 (1984), 13-20.
[4] Krishnamoorthy, A. S. and Parthasarathy, M., A multivariate gamma type distribution, Ann. Math. Stat. 22
(1951), 549-557.
[5] Letac, G. and Massam, H., Existence and non-existence of the non-central Wishart distribution, (2011),
arXiv:1108.2849.
[6] Miller, K. S. and Sackrowitz, H., Relationships between biased and unbiased Rayleigh distributions, SIAM
J. Appl. Math. 15 (1967), 1490-1495.
[7] Royen, T., On some central and non-central multivariate chi-square distributions, Statistica Sinica 5 (1995),
373-397.
[8] Royen, T., Integral representations and approximations for multivariate gamma distributions, Ann. Inst.
Statist. Math. 59 (2007), 499-513.
[9] Royen, T., A simple proof of the Gaussian correlation conjecture extended to some multivariate gamma
distributions, Far East J. Theor. Stat. 48 (2014), 139-145. (preprint: arXiv:1408.1028)
[10] Royen, T., Some probability inequalities for multivariate gamma and normal distributions, Far East J.
Theor. Stat. 51 (2015), 17-36.
[11] Royen, T., Non-central multivariate chi-square and gamma distributions, (2016), arXiv:1604.06906.
[12] Siotani, M., Hayakawa, T. and Fujikoshi, Y., Modern Multivariate Statistical Analysis: A Graduate Course
and Handbook, American Science Press, Inc. (1985), Columbus, Ohio, U.S.A.
View publication stats