Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples
For two families of beta distributions, we show that the generalized Stieltjes transforms of their elements may be written as elementary functions (powers and fractions) of the Stieltjes transform of the Wigner distribution. In particular, we retrieve the examples given by the author in a previous p...
Saved in:
| Date: | 2016 |
|---|---|
| Main Author: | |
| Format: | Article |
| Language: | English |
| Published: |
Інститут математики НАН України
2016
|
| Series: | Symmetry, Integrability and Geometry: Methods and Applications |
| Online Access: | http://dspace.nbuv.gov.ua/handle/123456789/147733 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Journal Title: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Cite this: | Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples / N. Demni // Symmetry, Integrability and Geometry: Methods and Applications. — 2016. — Т. 12. — Бібліогр.: 28 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-147733 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1477332019-02-16T01:25:13Z Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples Demni, N. For two families of beta distributions, we show that the generalized Stieltjes transforms of their elements may be written as elementary functions (powers and fractions) of the Stieltjes transform of the Wigner distribution. In particular, we retrieve the examples given by the author in a previous paper and relating generalized Stieltjes transforms of special beta distributions to powers of (ordinary) Stieltjes ones. We also provide further examples of similar relations which are motivated by the representation theory of symmetric groups. Remarkably, the power of the Stieltjes transform of the symmetric Bernoulli distribution is a generalized Stietljes transform of a probability distribution if and only if the power is greater than one. As to the free Poisson distribution, it corresponds to the product of two independent Beta distributions in [0,1] while another example of Beta distributions in [−1,1] is found and is related with the Shrinkage process. We close the exposition by considering the generalized Stieltjes transform of a linear functional related with Humbert polynomials and generalizing the symmetric Beta distribution. 2016 Article Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples / N. Demni // Symmetry, Integrability and Geometry: Methods and Applications. — 2016. — Т. 12. — Бібліогр.: 28 назв. — англ. 1815-0659 2010 Mathematics Subject Classification: 33C05; 33C20; 33C45; 44A15; 44A20 DOI:10.3842/SIGMA.2016.035 http://dspace.nbuv.gov.ua/handle/123456789/147733 en Symmetry, Integrability and Geometry: Methods and Applications Інститут математики НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| description |
For two families of beta distributions, we show that the generalized Stieltjes transforms of their elements may be written as elementary functions (powers and fractions) of the Stieltjes transform of the Wigner distribution. In particular, we retrieve the examples given by the author in a previous paper and relating generalized Stieltjes transforms of special beta distributions to powers of (ordinary) Stieltjes ones. We also provide further examples of similar relations which are motivated by the representation theory of symmetric groups. Remarkably, the power of the Stieltjes transform of the symmetric Bernoulli distribution is a generalized Stietljes transform of a probability distribution if and only if the power is greater than one. As to the free Poisson distribution, it corresponds to the product of two independent Beta distributions in [0,1] while another example of Beta distributions in [−1,1] is found and is related with the Shrinkage process. We close the exposition by considering the generalized Stieltjes transform of a linear functional related with Humbert polynomials and generalizing the symmetric Beta distribution. |
| format |
Article |
| author |
Demni, N. |
| spellingShingle |
Demni, N. Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples Symmetry, Integrability and Geometry: Methods and Applications |
| author_facet |
Demni, N. |
| author_sort |
Demni, N. |
| title |
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples |
| title_short |
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples |
| title_full |
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples |
| title_fullStr |
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples |
| title_full_unstemmed |
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples |
| title_sort |
generalized stieltjes transforms of compactly-supported probability distributions: further examples |
| publisher |
Інститут математики НАН України |
| publishDate |
2016 |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/147733 |
| citation_txt |
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions: Further Examples / N. Demni // Symmetry, Integrability and Geometry: Methods and Applications. — 2016. — Т. 12. — Бібліогр.: 28 назв. — англ. |
| series |
Symmetry, Integrability and Geometry: Methods and Applications |
| work_keys_str_mv |
AT demnin generalizedstieltjestransformsofcompactlysupportedprobabilitydistributionsfurtherexamples |
| first_indexed |
2025-07-11T02:44:05Z |
| last_indexed |
2025-07-11T02:44:05Z |
| _version_ |
1837316842288316416 |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications SIGMA 12 (2016), 035, 13 pages
Generalized Stieltjes Transforms
of Compactly-Supported Probability Distributions:
Further Examples?
Nizar DEMNI
IRMAR, Université de Rennes 1, Campus de Beaulieu, 35042 Rennes cedex, France
E-mail: nizar.demni@univ-rennes1.fr
URL: https://perso.univ-rennes1.fr/nizar.demni
Received December 12, 2015, in final form April 06, 2016; Published online April 12, 2016
http://dx.doi.org/10.3842/SIGMA.2016.035
Abstract. For two families of beta distributions, we show that the generalized Stieltjes
transforms of their elements may be written as elementary functions (powers and fractions)
of the Stieltjes transform of the Wigner distribution. In particular, we retrieve the examples
given by the author in a previous paper and relating generalized Stieltjes transforms of special
beta distributions to powers of (ordinary) Stieltjes ones. We also provide further examples
of similar relations which are motivated by the representation theory of symmetric groups.
Remarkably, the power of the Stieltjes transform of the symmetric Bernoulli distribution
is a generalized Stietljes transform of a probability distribution if and only if the power is
greater than one. As to the free Poisson distribution, it corresponds to the product of two
independent Beta distributions in [0, 1] while another example of Beta distributions in [−1, 1]
is found and is related with the Shrinkage process. We close the exposition by considering
the generalized Stieltjes transform of a linear functional related with Humbert polynomials
and generalizing the symmetric Beta distribution.
Key words: generalized Stieltjes transform; Beta distributions; Gauss hypergeometric func-
tion; Humbert polynomials
2010 Mathematics Subject Classification: 33C05; 33C20; 33C45; 44A15; 44A20
1 Reminder
Let λ be a positive real number and µ be a probability measure supported in the real line
(possibly depending on λ). Its generalized Stieltjes transform is defined by
Gλ,µ(z) :=
∫
1
(z − x)λ
µ(dx)
for complex numbers z lying outside the support of µ and such that (z − x)λ is the principal
branch of the power function. In particular, G1,µ is the (ordinary) Stieltjes transform of µ
which we simply denote hereafter by Gµ. The latter has deep connections with the theory of
continued fractions, combinatorics, Padé approximations and free probability theory (see [13, 20]
and references therein). As to Gλ,µ, much less is known. For instance, this transform was
introduced in [28] for measures µ supported in the positive half-line and an inversion formula was
derived. For the same type of measures, the notion of exact order is defined in [14] and [15] and
determined for hypergeometric series. When λ is a positive integer, the random weighted average
of independent random variables gives rise to generalized Stieltjes transforms of continuous
probability measures which may be written as products of Stieltjes transforms [23, 24, 27]). In
?This paper is a contribution to the Special Issue on Orthogonal Polynomials, Special Functions and Applica-
tions. The full collection is available at http://www.emis.de/journals/SIGMA/OPSFA2015.html
mailto:nizar.demni@univ-rennes1.fr
https://perso.univ-rennes1.fr/nizar.demni
http://dx.doi.org/10.3842/SIGMA.2016.035
http://www.emis.de/journals/SIGMA/OPSFA2015.html
2 N. Demni
the recent paper [18], generalized Stieltjes transforms appear as transmutation operators between
the solutions of the hypergeometric differential equation. Motivated by a generalization of free
probability theory and relying on the characterization of ultraspherical-type generating series
for orthogonal polynomials [1, 7], we provided in [6] four examples of compactly-supported
probability measures µ = µλ for which there exist probability measures ν = νλ satisfying
Gλ,µ(z) = [G1,ν(z)]λ = [Gν(z)]λ. (1.1)
Actually, the Wigner distribution plays a central role in (Voiculescu) free probability theory and
is an instance of the symmetric beta distribution. Since any symmetric probability distribution
with finite moments of all orders arises as the weak limit of sums of self-adjoint variables in
a suitable algebra [4], it is then natural to seek a λ-deformation of free probability theory where
the symmetric beta distribution appears in the central limit theorem. At the analytic side, when
relations like (1.1) hold, they bridge between the free additive convolution and its λ-deformation
provided that the measure ν is independent of λ. As shown in [6], this last property is satisfied
by the couple of symmetric beta and Wigner distributions and this is the only example derived
there with this property. This elementary observation raises the following questions:
• Given λ > 0 and µ = µλ, what are the necessary and/or sufficient conditions ensuring (1.1)
to hold with ν being independent of λ?
• Conversely, given λ > 0 and ν independent of λ, when does the power [Gν ]λ is a generalized
Stieltjes transform of a probability distribution µ = µλ?
• Is the symmetric beta distribution the only element in the family of beta distributions
satisfying (1.1) with ν being independent of λ?
• In the same vein, find other or all examples of beta distributions satisfying (1.1) with this
property?
Most likely, it is much easier to look for partial or definitive answers to the two last questions
rather than the remaining ones. In general, if we want to derive an expression of Gλ,µ for given λ
and µ, it suffices to expand
(z, x) 7→ 1
(z − x)λ
,
in a absolutely convergent series of orthogonal polynomials with respect to µ since then Gλ,µ
is nothing else but the first term of this series. Recently, this task was achieved in [5] for the
family of Jacobi polynomials whence the generalized Stieltjes transform of any beta distribution
in [−1, 1] follows after a simple integration. Using linear and quadratic transformations of the
Gauss hypergeometric function, we retrieve the few examples of (1.1) given in [6] and prove
for two larger classes of beta distributions that their generalized Stieltjes transforms are ele-
mentary functions (powers and fractions) of the Stieltjes transform of the Wigner distribution.
We also provide a new example of beta distributions satisfying (1.1), where ν does not depend
on λ as well thereby giving a negative answer to the third question. Actually, the probability
distribution ν describes the large time behavior of the Shrinkage process and is the transition
measure of triangular diagrams [16]. Two other examples of probability measures ν occurring
in representation theory of symmetric groups are considered. In the former, ν is the symmetric
Bernoulli distribution which is the transition measure of a square Young diagram [2] and (1.1)
holds if and only if λ ≥ 1. The latter shows that the λ-power of the Stieltjes transform of
the free Poisson distribution (which arises in the decomposition of the tensor product repre-
sentation of the symmetric group [3]) is the generalized Stieltjes transform of a product of two
independent beta distributions supported in [0, 1]. Indeed, the multiplicative convolution with
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions 3
a beta distribution of special parameters provides a transformation preserving probability dis-
tributions satisfying (1.1) for any λ > 0. In the last part of the paper, we investigate a possible
generalization of the relation (1.1) holding between the symmetric beta and the Wigner distri-
butions. More precisely, we consider the generating series of Humbert polynomials which are
d-orthogonal with respect to d ≥ 1 linear functionals and integrate it with respect to the first
functional. Doing so leads to the problem of finding a suitable root of a trinomial equation of
degree d+ 1 which is known to be expressed by means of the Gauss hypergeometric function.
2 Special functions
In this paragraph, we recall the definitions and some properties of special functions occurring
in the remainder of the paper. We refer the reader to [10, 13, 21, 25]. Let Γ denote the gamma
function and recall the Legendre duplication formula:
√
πΓ(2z) = 22z−1Γ(z)Γ
(
z +
1
2
)
.
For a complex number z and a positive integer k ≥ 1, the Pochhammer symbol is defined by
(z)k = z(z + 1) · · · (z + k − 1)
with the convention (z)0 = 1. When z is not a negative integer, we can write
(z)k =
Γ(z + k)
Γ(z)
, while (−n)k =
(−1)kn!
(n− k)!
for any positive integer n ≥ k and vanishes otherwise. Next, the hypergeometric series pFq is
defined by
pFq(a1, . . . , ap, b1, . . . , bq; z) =
∑
n≥0
(a1)n · · · (ap)n
(b1)n · · · (bq)n
zn
n!
when the series converges. In particular, the Gauss hypergeometric series 2F1 converges in the
open unit disc {|z| < 1} and has an analytic extension to the complex plane cut along the half
line [1,∞). This function will play a major role in our computations and we use further the
following linear and quadratic transformations valid for | arg(1− u)| < π:
2F1(a, c+ d, c;u) =
1
(1− u)a
2F1
(
a,−d, c; u
u− 1
)
, (2.1)
2F1(a, b, 2a;u) =
1
(1− (u/2))b
2F1
(
b
2
,
b+ 1
2
, a+
1
2
;
u2
(2− u)2
)
, (2.2)
2F1
(
a, a+
1
2
, b;u
)
=
22a
(1 +
√
1− u)2a
2F1
(
2a, 2a− b+ 1, b;
u
(1 +
√
1− u)2
)
. (2.3)
In particular, (2.3) yields the closed formulas
2F1
(
a− 1
2
, a, 2a;u
)
=
22a−1
(1 +
√
1− u)2a−1
, (2.4)
and
2F1
(
a, a+
1
2
, 2a;u
)
=
1√
1− u
22a−1
(1 +
√
1− u)2a−1
. (2.5)
4 N. Demni
We shall also make use of the Euler integral representation of the 2F1: for <(c) > <(b) > 0 and
|z| < 1,
2F1(a, b, c; z) =
Γ(c)
Γ(c− b)Γ(b)
∫ 1
0
(1− uz)−aub−1(1− u)c−b−1du. (2.6)
As to hypergeometric polynomials, we denote by P
(γ,β)
n , C
(α)
n the n-th Jacobi and ultraspherical
respectively:
P (γ,β)
n (x) =
(γ + 1)n
n!
2F1
(
−n, n+ γ + β + 1, γ + 1;
1− x
2
)
, γ, β > −1,
C(α)
n (x) =
(2α)n
(α+ 1/2)n
P (α−1/2,α−1/2)
n (x), α > −1/2, α 6= 0.
3 Generalized Stieltjes transform of beta distributions
The starting point of our investigations is the following expansion proved in [5, Theorem 1]: for
any z ∈ C\(−∞, 1] on any given ellipse with foci at ±1 and any x in the interior of this ellipse,
1
(z − x)λ
=
∞∑
n=0
Γ(γ + β + n+ 1)(λ)n
Γ(2n+ γ + β + 1)
2n
(z − 1)n+λ
× 2F1
(
n+ λ, n+ γ + 1, 2n+ 2 + γ + β,
2
1− z
)
P (γ,β)
n (x).
For fixed z, this series converges uniformly in x ∈ [−1, 1] (see, e.g., [26, Chapter IX, Theo-
rem 9.1.1]) and as shown in [5], it is valid for a set of parameters λ, γ, β containing R∗+×(−1,∞)2.
However, we shall restrict ourselves to the latter which is sufficient for our purposes and subse-
quent computations. As a matter of fact, the orthogonality of Jacobi polynomials with respect
to the beta distribution
µγ,β(dx) :=
Γ(γ + β + 2)
2γ+β+1Γ(γ + 1)Γ(β + 1)
(1− x)γ(1 + x)β1[−1,1](x)dx
readily gives1
Gλ,µγ,β (z) =
1
(z − 1)λ
2F1
(
λ, γ + 1, γ + β + 2,
2
1− z
)
. (3.1)
Note that the r.h.s. of (3.1) is not symmetric in (γ, β) unless γ = β. Nonetheless, we can use (2.1)
to transform (3.1) into
Gλ,µγ,β (z) =
1
(z + 1)λ
2F1
(
λ, β + 1, γ + β + 2,
2
1 + z
)
. (3.2)
Here, one performs the transformation for z lying in a suitable region in the complex plane (in
order to use the principal determinations of (z+1)λ, (z−1)λ, [(z+1)/(z−1)]λ) and then extends
the obtained equality analytically to C\(−∞, 1]. With (3.1) and (3.2) in hands, we start our
investigations of generalized Stieltjes transforms of two families of beta distributions.
1We could also use the Euler integral representation of the 2F1 hypergeometric function.
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions 5
3.1 Symmetric beta distributions
Take γ = β = λ− (1/2), λ > 0 in (3.1). Then (2.4) entails
1
(z − 1)λ
2F1
(
λ, λ+
1
2
, 2λ+ 1,
2
1− z
)
=
[
2
z +
√
z2 − 1
]λ
.
In the right-hand side of the last expression, we recognize the Stieltjes transform of the Wigner
distribution
GW(z) =
∫ 1
−1
1
z − x
2
√
1− x2
π
dx =
2
z +
√
z2 − 1
, z > 1.
As a result,
Γ(λ+ 1)√
πΓ(λ+ 1/2)
∫ 1
−1
(1− x2)λ−1/2
(z − x)λ
dx = [GW(z)]λ, (3.3)
which is the first example given in [6]. In order to retrieve the second example given in [6],
specialize (3.1) to γ = β = λ− (3/2), λ > 1 and use (2.5) to derive
1
(z − 1)λ
2F1
(
λ− 1
2
, λ, 2λ− 1,
2
1− z
)
=
1√
z2 − 1
[GW(z)]λ−1.
But the Stieltjes transform of the arcsine distribution reads
GAS(z) =
∫ 1
−1
1
z − x
dx
π
√
1− x2
=
1√
z2 − 1
.
Consequently, for any λ > 1,
Γ(λ)√
πΓ(λ− 1/2)
∫ 1
−1
(1− x2)λ−3/2
(z − x)λ
dx = GAS(z)[GW(z)]λ−1. (3.4)
These two examples are indeed instances of the following more general formula:
Proposition 3.1. If λ = γ + 1/2 + k ≥ 0 for some integer k ≥ 0, then the generalized Stieltjes
transform of
µγ,γ(dx) =
Γ(γ + 3/2)√
πΓ(γ + 1)
(
1− x2
)γ
1[−1,1](x)dx
may be written by means of powers and fractions of the variable GW:
Gλ,µγ,γ (z) =
4k[GW(z)]λ
[4− [GW(z)]2]k
2F1
(
k, 1− k, γ +
3
2
;
[GW(z)]2
[GW(z)]2 − 4
)
.
Proof. Specialize (3.1) to γ = β and use (2.2) to get
1
(z − 1)λ
2F1
(
λ, γ + 1, 2γ + 2,
2
1− z
)
=
1
zλ
2F1
(
λ
2
,
λ+ 1
2
, γ +
3
2
,
1
z2
)
.
Assuming γ + (1/2) ≤ λ and using (2.3) transforms the r.h.s. of the last equality to
1
zλ
2F1
(
λ
2
,
λ+ 1
2
, γ +
3
2
,
1
z2
)
= [GW(z)]λ 2F1
(
λ, λ− γ − 1
2
, γ +
3
2
;
[GW(z)]2
4
)
.
6 N. Demni
Assuming further that λ = γ + 1/2 + k for some integer k ≥ 0 and noting that
1− [GW(z)]2
4
= 2
√
z2 − 1
z +
√
z2 − 1
∈ [0, 1), z > 1,
then we can appeal to the linear transformation (2.1) and end up with
2F1
(
k, λ, γ +
3
2
;
[GW(z)]2
4
)
=
4k
[4− [GW(z)]2]k
2F1
(
k, 1− k, γ +
3
2
;
[GW(z)]2
[GW(z)]2 − 4
)
. �
Remark 3.2. Note that
1− [GW(z)]2
4
=
√
z2 − 1GW(z) =
GW(z)
GAS(z)
so that
Gλ,µλ−(1/2)−k,λ−(1/2)−k(z) = [GW(z)]λ−1[GAS(z)]k 2F1
(
k, 1− k, γ +
3
2
;
GAS(z)GW(z)
4
)
.
In the next paragraph, we derive a similar formula for a family of nonsymmetric beta distri-
butions.
3.2 Nonsymmetric beta distributions
In [6], two other examples of probability distributions satisfying (1.1) were derived. They cor-
respond to nonsymmetric beta distributions with parameters
(γ, β) =
(
λ− 1
2
, λ− 3
2
)
, (γ, β) =
(
λ− 3
2
, λ− 1
2
)
(3.5)
and readily follow from (3.1) together with the identities (2.4) and (2.5). As with the previous
family of symmetric beta distributions, we can derive a more general formula for generalized
Stieltjes transforms of nonsymmetric ones with parameters γ = λ− 1/2 and β = λ− k − 1/2 =
γ − k, k ≥ 1:
Proposition 3.3. For any z > 1,
Gλ,µλ−(1/2),λ−(1/2)−k(z) =
[GW(z)]λ−(k/2)
(1 + z)k/2
2F1
(
1− k, k, 2λ− k + 1;
GW(z)
GW(z) + 2
)
.
The equality extends analytically to the complex plane cut along (−∞, 1].
Proof. Using (2.3) and (2.1), we get
Gλ,µλ−(1/2),λ−(1/2)−k(z) =
1
(z − 1)λ
2F1
(
λ, λ+
1
2
, λ+ β +
3
2
,
2
1− z
)
= [GW(z)]λ 2F1
(
2λ, λ− β − 1
2
, λ+ β +
3
2
;−GW(z)
2
)
=
2k[GW(z)]λ
[2 +GW(z)]k
2F1
(
1− k, k, 2λ− k + 1;
GW(z)
GW(z) + 2
)
.
From the relation
1 +
[GW(z)]2
4
= zGW(z),
it follows that(
1 +
GW(z)
2
)2
= (z + 1)GW(z)
and the proposition follows. �
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions 7
Remark 3.4. Interchanging the roles of γ and β with the help of (3.2), similar computations
yield
Gλ,µλ−(1/2)−k,λ−(1/2)
(z) =
2k[GW(z)]λ
[2−GW(z)]k
2F1
(
1− k, k, 2λ− k + 1;
GW(z)
GW(z)− 2
)
=
[GW(z)]λ−(k/2)
(z − 1)k/2
2F1
(
1− k, k, 2λ− k + 1;
GW(z)
GW(z)− 2
)
.
4 Generalized Stieltjes transforms as powers
of Stieltjes transforms: further examples
So far, we dispose of four couples (µ, ν) of probability distributions satisfying (1.1): those
displayed in (3.3) and (3.4) and those corresponding to the two couples of parameters specified
in (3.5). However, only (3.3) has the property that ν (the Wigner distribution in this case)
does not depend on λ. In this section, we derive three more examples enjoying this property
which, like the Wigner distribution, appeared as transition distributions of Young or continuous
diagrams (see [16] for more details).
4.1 The square diagram
The symmetric Bernoulli distribution
ν =
1
2
[δ−L + δL], L ∈ N\{0},
is the basic example of transition distribution of a Young diagram. Indeed, it corresponds to
the square diagram of width L [2]. To simplify, take L = 1 so that
Gν(z) =
z
z2 − 1
.
Then we shall prove
Proposition 4.1. If ν is the symmetric Bernoulli distribution, then
Gλ,µ(z) = [Gν(z)]λ
for every λ ≥ 1, where
µ(dx) = µλ(dx) =
1
2λ
[
[δ1 + δ−1](dx) +
1√
|x|
h(
√
|x|)dx
]
and
h(x) = hλ(x) :=
λ(λ− 1)
4
xλ−1 2F1
(
λ
2
+ 1,
λ+ 1
2
; 2; 1− x
)
1[0,1](x).
Proof. Let λ > 0 and assume (1.1) holds for some probability measure µ = µλ, namely,∫
1
(1− wx)λ
µ(dx) =
1
(1− w2)λ
=
∑
k≥0
(λ)k
k!
w2k, w = 1/z ∈ (0, 1).
Hence, µ is symmetric and its even moments are given by∫
x2kµ(dx) =
(2k)!(λ)k
k!(λ)2k
.
8 N. Demni
Using Legendre duplication formula, these moments may be written as
Γ(k + λ)Γ(k + 1/2)
2λ−1Γ(k + (λ/2))Γ(k + (λ+ 1)/2)
=
(λ)k(1/2)k
(λ/2)k((λ+ 1)/2)k
.
According to [9, Theorem 6.2], there exists a probability distribution χ = χλ supported in (0, 1)
such that∫ 1
0
xkχ(dx) =
(λ)k(1/2)k
(λ/2)k((λ+ 1)/2)k
if and only if λ ≥ 1. This is an instance of the so-called G-distributions [8, 9] and its Lebesgue
decomposition is given by
χ(dx) =
Γ((λ+ 1)/2)Γ(λ/2)
Γ(1/2)Γ(λ)
×
[
δ1(dx) +
λ(λ− 1)
4
xλ−1 2F1
(
λ
2
+ 1,
λ+ 1
2
; 2; 1− x
)
1[0,1](x)dx
]
=
1
2λ−1
[
δ1(dx) +
λ(λ− 1)
4
xλ−1 2F1
(
λ
2
+ 1,
λ+ 1
2
; 2; 1− x
)
1[0,1](x)dx
]
.
Noting ξ is the push forward of µ under the square map x 7→ x2, we are done. �
Remark 4.2. Some parts of Theorems 1 and 2 in [8] as well as Theorem 6.2 in [9] appear also
in [14].
4.2 The Shrinkage process
While the Wigner distribution is the limiting transition distribution of Young diagrams drawn
from the Plancherel measure, the arcsine distribution describes in this case the limiting shape
of these diagrams (it is referred to as the Rayleigh measure [17]). In turn, the latter is the
transition distribution whose Rayleigh measure is the symmetric Bernoulli distribution and
belongs to a more general family of distributions related with triangular diagrams and with the
Shrinkage process [16]. The analogue of example (3.4) is derived as follows. Consider (3.1) with
λ = γ + β + 2:
1
(z − 1)λ
2F1
(
λ, γ + 1, λ,
2
1− z
)
=
1
(z − 1)λ
(
1− z
−z − 1
)γ+1
=
1
(z − 1)λ−γ−1
1
(z + 1)γ+1
for z > 1. Write γ + 1 = λp for some 0 < p < 1, then
∫
1
(z − x)λ
µpλ−1,(1−p)λ−1(dx) =
[
1
(z − 1)1−p(z + 1)p
]λ
=
[∫
1
z − x
µp−1,−p(dx)
]λ
.
In particular, if p = 1/2 then
Γ(λ)
2λ−1[Γ(λ/2)]2
∫ 1
−1
1
(z − x)λ
(
1− x2
)(λ/2)−1
dx =
[∫ 1
−1
1
z − x
dx
π
√
1− x2
]λ
.
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions 9
4.3 The free Poisson distribution
Let µ be a (non necessarily symmetric) probability distribution supported in [−1, 1] and de-
note κλ the probability distribution with beta density
Γ(λ)
[Γ(λ/2)]2
[x(1− x)](λ/2)−11[0,1](x).
Denote κλ ? µ the multiplicative convolution of κλ and µ, that is, the probability distribution of
the product of two independent random variables with probability distributions κλ and µ. By
definition,∫
f(x) (κλ ? µ) (dx) =
∫∫
f(uv)κλ(du)µ(dv)
for any bounded measurable function f . Then, for |z| > 1, the integral representation (2.6)
yields∫
1
(z − x)λ
κλ ? µ(dx) =
Γ(λ)
[Γ(λ/2)]2zλ
∫ (∫ 1
0
1
(1− (v/z)u)λ
[u(1− u)](λ/2)−1du
)
µ(dv)
=
1
zλ
∫
2F1
(
λ,
λ
2
, λ;
v
z
)
µ(dv) =
1
zλ/2
∫
1
(z − v)λ/2
µ(dv).
In particular, if µ = µλ satisfies (1.1) with the exponent λ/2∫
1
(z − x)λ/2
µλ/2(dx) = [Gν(z)]λ/2
for some probability measure ν independent of λ, then∫
1
(z − x)λ
κλ ? µ(dx) =
[
[Gν(z)]1/2
z1/2
]λ
at least for z > 1. Since =
(
[Gν(z)]1/2/z1/2
)
< 0 for =(z) > 0, then the last equality extends
analytically to the upper half-plane. Moreover,
lim
y→∞
(iy)
[Gν(iy)]1/2
(iy)1/2
= 1
so that z 7→ [Gν(z)]1/2/z1/2 is a Nevannlina–Pick function and may be represented as the
Stieltjes transform of a probability measure (see for instance [22, Lemma 2.2]). This elementary
observation allows to derive further examples of probability distributions (µ, ν) satisfying (1.1).
For instance,
Proposition 4.3. For any λ > 0 and z in the upper half-plane,∫
1
(z − x)λ
κλ ? κλ+1(dx) =
2λ
[z +
√
z(z − 1)]λ
.
Proof. Let µ = κλ+1, then for z in a suitable region in the upper half-plane,∫
1
(z − x)λ
κλ ? κλ+1(dx) =
Γ(λ+ 1)
[Γ(λ+ 1/2)]2
1
zλ/2
∫
1
(z − x)λ/2
[x(1− x)](λ+1)/2−1dx
=
1
zλ
2F1
(
λ
2
,
λ+ 1
2
, λ+ 1;
1
z
)
=
2λ
[
√
z(
√
z +
√
z − 1)]λ
=
2λ
[z +
√
z(z − 1)]λ
,
10 N. Demni
where the third equality follows from (2.4). But, the map
z 7→ 2
z +
√
z(z − 1)
= 2
z −
√
(z − (1/2))2 − (1/4)
z
is the Stieltjes transform of the free Poisson distribution with parameters (1, 1/4) (see for instance
[20, p. 204]) whose density reads
2
π
√
1− x
x
1[0,1](x).
Therefore, the proposition follows by analytic continuation. �
Remark 4.4. The free Poisson distribution appears the limiting transition measure of random
diagrams arising from the decomposition of tensor product representations of the symmetric
group [3]. On the other hand, the density of κλ ? κλ+1 is given by [8]
Γ(λ)Γ(λ+ 1)
Γ(λ/2)Γ((λ+ 1)/2)Γ(λ+ (1/2))
x(λ/2)−1(1− x)λ−1/2
× 2F1
(
λ− 1
2
,
λ+ 1
2
, λ+
1
2
, 1− x
)
1[0,1](x).
5 Further developments
The Humbert polynomials
(
H
(α,d)
n
)
n≥0 of parameters α > −1/2, α 6= 0, d ∈ N\{0} and degrees
n ∈ N, are defined by their generating series:∑
n≥0
H(α,d)
n zn =
1
(1− (d+ 1)xz + zd+1)α
,
and reduce to ultraspherical polynomials
(
C
(α)
n
)
n
when d = 1. They cease to be orthogonal as
soon as d ≥ 2 and are rather d-orthogonal in the following sense (see, e.g., [19]): there exist d
linear functionals Γ
(α,d)
0 , . . . ,Γ
(α,d)
d−1 on the space of polynomials such that
Γk
(
H(α,d)
n H(α,d)
r
)
= 0, r > dn+ k, Γk
(
H(α,d)
n H
(α,d)
nd+k
)
6= 0,
for all n ∈ N, 0 ≤ k ≤ d− 1. In particular, Γ
(α,d)
0
(
H
(α,d)
0 H
(α,d)
r
)
= 0 for any r > 0 so that∫
1
(1− (d+ 1)xz + zd+1)α
Γ0(dx) = 1.
Hence, we can find a suitable region outside the compact support of Γ
(α,d)
0 for which∫
1
(fα(z)− x)α
Γ0(dx) = [(d+ 1)z]α, where fα(z) :=
1 + zd+1
(d+ 1)z
.
Now, the trinomial equation fα(z) = y has (d + 1) roots which may be expressed through the
Gauss hypergeometric function [11, 12] and we seek the one which tends to zero when y tends
to infinity in the upper half-plane. For instance, when d = 2 then the polynomial equation
z3 − 3yz + 1 = 0, =(y) > 0,
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions 11
may be transformed by setting z =
√
−yw into
w3 + 3w +
1
(−y)3/2
= 0.
From [12, p. 265] the sought root is given by
z = − (−y)1/2
3(−y)3/2
2F1
(
1
3
,
2
3
,
3
2
;− 1
4((−y)3/2)2
)
=
1
3y
2F1
(
1
3
,
2
3
,
3
2
;
1
4y3
)
,
where the last equality is valid for y lying in some sector in the upper half-plane. Using the
Euler integral representation
2F1
(
1
3
,
2
3
,
3
2
;
1
4y3
)
=
41/3Γ(3/2)
Γ(2/3)Γ(5/6)
∫ 1
0
y
(y3 − x)1/3
x−1/3(1− x)−1/6dx,
we get the following equality∫
1
(y − x)α
Γ
(α,2)
0 (dx) =
{
41/3Γ(3/2)
Γ(2/3)Γ(5/6)
∫ 1
0
1
(y3 − x)1/3
x−1/3(1− x)−1/6dx
}α
.
More generally, the solution to the equation
zd+1 + (d+ 1)z +
1
(−y)(d+1)/d
= 0, =(y) > 0,
tending to zero as y →∞ may be derived along the same lines written in [12, p. 266], and may
be expressed through the hypergeometric series
dFd−1
(
i
d+ 1
, 1 ≤ i ≤ d, i+ 1
d
, 1 ≤ i ≤ d, i 6= d− 1;
(−1)d
ddyd+1
)
.
Thus, the reasoning above applies and leads to the generalized Stieltjes transform of Γ
(α,d)
0 as
a α-power of this hypergeometric series.
Remark 5.1. An anonymous referee pointed out to the author that the functional Γ
(α,d)
0 co-
incides up to the variable change x 7→ ddxd+1 with the representative measure of the hyper-
geom̄etric function d+1Fd as a generalized Stieltjes transform [15, Theorem 2]. Actually, Γ
(α,d)
0
is expressed in terms of the Meijer G-function [19, Theorem 2.4]
1
x
Gd+1,0
d+1,d+1
(
ddxd+1
∣∣∣(α+ 1)/d, . . . , (α+ d)/d, 1
1/(d+ 1), . . . , d/(d+ 1), 1
)
which reduces to
1
x
Gd,0d,d
(
ddxd+1
∣∣∣(α+ 1)/d, . . . , (α+ d)/d
1/(d+ 1), . . . , d/(d+ 1)
)
,
while Theorem 2 in [15] entails
d+1Fd
(
α, a1, . . . , ad, b1, . . . , bd,
1
y
)
=
d∏
i=1
Γ(bi)
Γ(ai)
∫ 1
0
yα
(y − x)α
Gd,0d,d
(
x
∣∣∣b1, . . . , bd
a1, . . . , ad
)
dx
x
.
for y ∈ C\(−∞, 1].
12 N. Demni
Acknowledgements
The author would like to thank H. Cohl and C. Dunkl for their helpful remarks on earlier
versions of the paper. He also greatly appreciates the suggestions of the anonymous referees
which considerably improved the presentation.
References
[1] Al-Salam W.A., Verma A., Some sets of orthogonal polynomials, Rev. Técn. Fac. Ingr. Univ. Zulia 9 (1986),
83–88.
[2] Biane P., Representations of symmetric groups and free probability, Adv. Math. 138 (1998), 126–181.
[3] Biane P., Approximate factorization and concentration for characters of symmetric groups, Int. Math. Res.
Not. 2001 (2001), no. 4, 179–192, math.RT/0006111.
[4] Cabanal-Duvillard D., Ionescu V., Un théorème central limite pour des variables aléatoires non-commuta-
tives, C. R. Acad. Sci. Paris Sér. I Math. 325 (1997), 1117–1120.
[5] Cohl H.S., Fourier, Gegenbauer and Jacobi expansions for a power-law fundamental solution of the poly-
harmonic equation and polyspherical addition theorems, SIGMA 9 (2013), 042, 26 pages, arXiv:1209.6047.
[6] Demni N., Generalized Cauchy–Stieltjes transforms of some beta distributions, Commun. Stoch. Anal. 3
(2009), 197–210, arXiv:0902.0054.
[7] Demni N., Ultraspherical type generating functions for orthogonal polynomials, Probab. Math. Statist. 29
(2009), 281–296, arXiv:0812.3666.
[8] Dufresne D., The beta product distribution with complex parameters, Comm. Statist. Theory Methods 39
(2010), 837–854.
[9] Dufresne D., G distributions and the beta-gamma algebra, Electron. J. Probab. 15 (2010), no. 71, 2163–2199.
[10] Erdélyi A., Magnus W., Oberhettinger F., Tricomi F.G., Higher transcendental functions, Vol. I, Bateman
Manuscript Project, McGraw-Hill Book Co., New York, 1953.
[11] Glasser M.L., Hypergeometric functions and the trinomial equation. Higher transcendental functions and
their applications, J. Comput. Appl. Math. 118 (2000), 169–173.
[12] Hille E., Analytic function theory, Vol. 1, Introduction to Higher Mathematics, Ginn and Company, Boston,
1959.
[13] Ismail M.E.H., Classical and quantum orthogonal polynomials in one variable, Encyclopedia of Mathematics
and its Applications, Vol. 98, Cambridge University Press, Cambridge, 2005.
[14] Karp D., Prilepkina E., Generalized Stieltjes functions and their exact order, J. Class. Anal. 1 (2012), 53–74.
[15] Karp D., Prilepkina E., Hypergeometric functions as generalized Stieltjes transforms, J. Math. Anal. Appl.
393 (2012), 348–359, arXiv:1112.5769.
[16] Kerov S.V., Transition probabilities of continual Young diagrams and the Markov moment problem, Funct.
Anal. Appl. 27 (1993), 104–117.
[17] Kerov S.V., Interlacing measures, in Kirillov’s Seminar on Representation Theory, Amer. Math. Soc. Transl.
Ser. 2, Vol. 181, Amer. Math. Soc., Providence, RI, 1998, 35–83.
[18] Koornwinder T.H., Fractional integral and generalized Stieltjes transforms for hypergeometric functions as
transmutation operators, SIGMA 11 (2015), 074, 22 pages, arXiv:1504.08144.
[19] Lamiri I., Ouni A., d-orthogonality of Humbert and Jacobi type polynomials, J. Math. Anal. Appl. 341
(2008), 24–51.
[20] Nica A., Speicher R., Lectures on the combinatorics of free probability, London Mathematical Society Lecture
Note Series, Vol. 335, Cambridge University Press, Cambridge, 2006.
[21] Rainville E.D., Special functions, The Macmillan Co., New York, 1960.
[22] Shohat J.A., Tamarkin J.D., The problem of moments, American Mathematical Society Mathematical Sur-
veys, Vol. 1, American Mathematical Society, New York, 1943.
[23] Soltani A.R., Homei H., Weighted averages with random proportions that are jointly uniformly distributed
over the unit simplex, Statist. Probab. Lett. 79 (2009), 1215–1218.
http://dx.doi.org/10.1006/aima.1998.1745
http://dx.doi.org/10.1155/S1073792801000113
http://dx.doi.org/10.1155/S1073792801000113
http://arxiv.org/abs/math.RT/0006111
http://dx.doi.org/10.1016/S0764-4442(97)88716-2
http://dx.doi.org/10.3842/SIGMA.2013.042
http://arxiv.org/abs/1209.6047
http://arxiv.org/abs/0902.0054
http://arxiv.org/abs/0812.3666
http://dx.doi.org/10.1080/03610920902802599
http://dx.doi.org/10.1214/EJP.v15-845
http://dx.doi.org/10.1016/S0377-0427(00)00287-9
http://dx.doi.org/10.1017/CBO9781107325982
http://dx.doi.org/10.1017/CBO9781107325982
http://dx.doi.org/10.1016/j.jmaa.2012.03.044
http://arxiv.org/abs/1112.5769
http://dx.doi.org/10.1007/BF01085981
http://dx.doi.org/10.1007/BF01085981
http://dx.doi.org/10.3842/SIGMA.2015.074
http://arxiv.org/abs/1504.08144
http://dx.doi.org/10.1016/j.jmaa.2007.09.047
http://dx.doi.org/10.1017/CBO9780511735127
http://dx.doi.org/10.1017/CBO9780511735127
http://dx.doi.org/10.1016/j.spl.2009.01.009
Generalized Stieltjes Transforms of Compactly-Supported Probability Distributions 13
[24] Soltani A.R., Roozegar R., On distribution of randomly ordered uniform incremental weighted averages:
divided difference approach, Statist. Probab. Lett. 82 (2012), 1012–1020.
[25] Srivastava H.M., Manocha H.L., A treatise on generating functions, Ellis Horwood Series: Mathematics and
its Applications, Ellis Horwood Ltd., Chichester, Halsted Press, New York, 1984.
[26] Szegő G., Orthogonal polynomials, Colloquium Publications, Vol. 23, 4th ed., Amer. Math. Soc., Providence,
R.I., 1975.
[27] Van Assche W., A random variable uniformly distributed between two independent random variables,
Sankhyā Ser. A 49 (1987), 207–211.
[28] Widder D.V., The Stieltjes transform, Trans. Amer. Math. Soc. 43 (1938), 7–60.
http://dx.doi.org/10.1016/j.spl.2012.02.007
http://dx.doi.org/10.2307/1989901
1 Reminder
2 Special functions
3 Generalized Stieltjes transform of beta distributions
3.1 Symmetric beta distributions
3.2 Nonsymmetric beta distributions
4 Generalized Stieltjes transforms as powers of Stieltjes transforms: further examples
4.1 The square diagram
4.2 The Shrinkage process
4.3 The free Poisson distribution
5 Further developments
References
|