EULEUR

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

July 1, 2022

Probability Distribution derived from the Gamma function


Par

TOURE LAN SAN A1 , DR TOURE M AM ADOU BA1 et Dr CAMARA ALPHA OU M AR1


1.Departement de Mathématiques Universite Gamal Abdel Nasser de Conakry.

Résumé La fonction d’Euler Gamma a contribué à la formulation d’importantes lois de probabilité.


Ces lois sont à la base de la théorie de l’échantillonnage, à savoir les tests d’hypothèses et la théorie de
l’estimation par intervalles.

Mot-clé : Fonction gamma, Fonction bêta, Probabilité, degré de liberté,

Abstract The function of Euler Gamma has contributed to the formulation of important laws of prob-
ability. These laws are at the basis of the theory of sampling, namely the hypothesis tests and the interval
estimation theory.

Keyword :Gamma Function ,Beta Function,Probability, degree of freedom,

0.1 INTRODUCTION
Application of the functions of Euler contributed and facilitated the obtaining of the important results in
statistics and especially in the theories of distribution of sampling.In this paper we study some properties
of the function Gama and Beta, in the second part we also study some laws of probabilities or probability
distributions derived from the function gama and the last part, the application concerns the laws of
probabilities of certain random variables from normal populations.

0.1.1 Gamma Function


Definition Gamma Function Γ is defined by:
Z ∞
Γ(x) = exp(−t)tx−1 dt for all x > 0 (1)
0

Fundamental relationship

:
Γ(x + 1) = xΓ(x)
Indeed : Z ∞
Γ(x + 1) = exp(−t)tx dt
0

Z ∞
= [−exp(−t)tx ]∞
0 +x exp(−t)tx−1 dt
0
2

but [−exp(−t)tx ]∞
0 = 0 if t = 0 or ∞
We have Z ∞
Γ(1) = exp(−t)dt = 1
0
Therefore
Γ(n + 1) = nΓ(n) = n(n − 1)Γ(n − 2) = n!Γ(1) = n!
Γ(n + 1) = n! (2)

0.2 Beta Function


0.2.1 Definition
Γ(p)Γ(q)
B(p, q) = (3)
Γ(p + q)

Integral expression of the Beta function.


Z ∞ Z ∞
Γ(p) = exp(−t)tp−1 dt = 2 exp(−u2 )u2p−1 du
0 0
with t = u2
Therefore : Z ∞Z ∞
Γ(p)Γ(q) = 4 exp(−u2 )u2p−1 du exp(−v 2 )v 2q−1 dv
0 0
Z ∞Z ∞
=4 exp(−u2 + v 2 )u2p−1 v 2q−1 dudv
0 0
In polar coordinates , for u = ρcos(θ) and v = ρsin(θ)
Z ∞ Z π/2
Γ(p)Γ(q) = 4 exp(−ρ2 )ρ2p−1+2q−1 (cos(θ)2p−1 (sin(θ)2q−1 ρdρdθ
ρ=0 θ=0

Z ∞ Z π/2
2 2(p+q)−1
=4 exp(−ρ )ρ (cos(θ)2p−1 (sin(θ)2q−1 dρdθ
ρ=0 θ=0
Z π/2
= 2Γ(p + q) (cos(θ)2p−1 (sin(θ)2q−1 dθ
θ=0
so Z π/2
B(p, q) = 2 (cos(θ)2p−1 (sin(θ)2q−1 dθ (4)
θ=0
In particular
π/2
Γ[(1/2)]2
Z
B(1/2, 1/2) = = [Γ(1/2)]2 = 2 dθ = π
Γ(1) 0

Γ(1/2) = π (5)
2
Passing in Cartesian coordinates,so by posing cos (θ = t, we find:
Z 1
B(p, q) = tp−1 (1 − t)q−1 dt (6)
0
3

0.3 Some Probability Laws Derived from the Gama Functions


0.3.1 Gamma Distribution
This distribution plays an important role in statistics

definition

A random variable X is said to be distributed as the gamma distribution of parameter r if its density is
for θ > 0 and alpha > 0
α
f (x) = e−θx xα−1 for 0 < x < ∞ (7)
Γ(α)

This function represents a density, because by definition of Γ(α) (see eq 1 ),


Z ∞
f (x)dx = 1
0
,
By definition, Z ∞
E(X) = xf (x)dx
0
We have,
∞ ∞
θα
Z Z
1 Γ(α + 1)
E(X) = e−θx xα dx = e−y y α dy/θ = = α/θ
Γ(α) 0 Γ(α) 0 θΓ(α)

V (X) = E(X 2 ) − E 2 (X)


We have ,
∞ ∞
θα
Z Z
1
V (X) = e−θx xα+1 dx − (α/θ)2 = e−y y α+1 dy/θ2 − (α/θ)2
Γ(α) 0 Γ(α) 0

hence ,
Γ(α + 2) Γ(α + 1) 1
− (α/θ)2 = (α + 1) 2 − (α/θ)2 = 2 α(α + 1) − α2 = α/θ2

V (X) = 2
θ Γ(α) θ Γ(α) θ

0.3.2 Beta Distribution


The beta distribution is the distribution of an X; 0 < x < 1, dependent on two parameters n and p,
whose density is:

1
f (x) = x(n−1) (1 − x)(p−1) (8)
B(n, p)
Γ(n)Γ(p)
n, p > o ; where B(n, p) = Γ(+p)
4

As for the previous distribution we find:


n
E(X) =
n+p
and

np
V (X) =
(n + p + 1)(n + p)2

0.3.3 The Normal Distribution


One of the most important continuous probability distribution is the normal distribution, normal curve
or Gassian distribution defined by the equation:
 2 !
1 1 x−µ
f (x) = p exp − , x ∈ R.(9)
σ (2π) 2 σ

. Where µ = mean ,σ =standard deviation, π = 3.14159..., exp=2.71828...

f (x) > 0, ∀x ∈ R
Z +∞
f (x)dx = 1
−∞

Indeed 2 !
+∞ +∞ 
1 x−µ
Z Z
1
f (x)dx = p exp − dx
−∞ −∞ σ (2π) 2 σ
Z +∞  
1 1
= p exp − z 2 dz
−∞ (2) 2
With the change of variable :  
X −µ
Z=
σ
We have : Z +∞ 
1 1 2
= p exp − z dz = 1
−∞ (2) 2
 
THEOREME 1. If X∼ N (µ, σ) ⇐⇒ Z = X−µσ ∼ N (0, 1)
The variable  
X −µ
Z= N (0, 1)
σ
,  
1 1
f (z) = p exp − z 2
(2π) 2
is the probability density of the variable Z We will show that the variance of Z is equal to 1.
Z +∞  
2 1 1 2
V (Z) = z p exp − z dz
−∞ (2π) 2
5

Z +∞  
2 1
=√ z exp − z 2
2
2π 0 2
2
let’s put t = z /2, the zdz = dt :
Z +∞
2
=√ exp (−t)
π 0
2 3 2 1 1
= √ Γ( ) = √ Γ( )
π 2 π2 2
p
Γ(1/2) = (π) (see equation 5) then V (Z) = 1

0.3.4 Chi-square Distribution


Definition

The Chi Square distribution is the distribution of the sum of squared standard normal deviates. The
degrees of freedom of the distribution is equal to the number of standard normal deviates being summed.
We say that X follows a Chi-square distribution with ν degrees of freedom , denote χ2ν , if the probability
density function of X is:

1
f (x) = exp(−x/2)xν/2−1 for ν ∈ ℵ (10)
2ν/2 Γ(ν/2)

The chi-square distribution is another distribution of Gamma ,Indeed for θ = 1/2 and α = ν/2 (in
equation 7). Therefore: Z ∞
E(X) = xf (x)dx = ν
0
and Z ∞
V (X) = x2 f (x)dx − E 2 (X) = 2ν
0

0.3.5 The Fisher ’F distribution


This law is related o the ratio of two independent quadratic forms.Suppose that χ1 and χ2 are inde-
pendently distributed by chi-square distributions with ν1 and ν2 degrees of freedom ,respectively. We
define

χ1 /ν1
F = (11)
χ2 /ν2
its density function is:
1 ( νν12 )f ν1 /2−1
g(f ) = (12)
B( ν21 ν22 ) (1 + ν1 ν
ν2 f )1 + ν2 )/2
ν2
E(F ) = (13)
ν2 − 2
and
ν22 ν1 + ν2 − 2
V (F ) = 2 (14)
ν1 (ν2 − 2)2 (n2 − 4)
6

0.3.6 Student’s” t Distribution


Another distribution of considerable practical importance is that of the ratio of a normally distributed
variate to the root of a a variate independently distributed by chi-square distribution. More precisely,
if x is normally distributed with mean µ and variance σ 2 , if u has the chi-square distribution with ν
degrees of freedom, and if x and u are independent distributed, we seek the distribution of.

(x − µ)/σ
t= p
u/ν(15)

Probability density function Student’s t-distribution has the probability density function given by

− n+1
t2

Γ(ν + 1)/2 2

f (t) = √ 1+ for − ∞ < t < +∞ (16)


πνΓ(ν/2) n
E(t = 0 forν > 1 (17)
ν
V ar(t) = for ν > 2 (18)
ν−2
7

0.4 Applications
0.4.1 Distribution of sample variance
The variance s of a sample is also a random variable. The variance of a sample is given by the formula:
n
X (xi − x̄)2
s2 = (19)
i=1
n−1

We also know that for a sample of size n derived from a normal population of mean µ and variance σ 2 ,
the quantity
n n  2
X X xi − µ
Zi 2 =
i=1 i=1
σ
follows the chi-square law with n degrees of freedom. Any sum
Pnof squares of random variables is asso-
ciated with a number of degrees of freedom. Thus, the sum i=1 xi − µ)2 has a n degree of freedom,
2 2
= i=1 (xiσ−x̄)
Pn Pn
but i=1 (xi − x̄)2 has only (n−1) degrees of freedom. Then the ratio (n−1)s
σ2 2 follows
a law of χ2 a (n − 1) degrees of freedom.

(n − 1)s2
χ2 =
σ2

0.4.2 The distribution of the quotient of two variances.


Be two normal populations of σ1 2 and σ2 2 variances respectively.we take two independent samples of
size n1 and n2 , respectively. We know that x1 follows the chi-square law with (n-1) degrees of freedom.
Therefor 2
(n1 −1)s1 /(n1 −1)
σ12 s21 /σ12
F = =
(n2 −1)s22 /(n2 −1) s22 /σ22
σ22

is distributed as F with (n1 − 1) and (n2 − 1) degrees of freedom.

X̄−µ
0.4.3 Distribution of the quantity Z = √
s/ n

2
Let Z be a normal variate with means 0 and variance 1.Let U = (n−1)s
σ2 be a chi-square variable with
(n − 1)degrees of freedom and let U and Z be independent. Then the random variable

Z n−1
t= √
U
is distributed as Student’s with (n − 1) degrees of freedom.
Contents

0.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
0.1.1 Gamma Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
0.2 Beta Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
0.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
0.3 Some Probability Laws Derived from the Gama Functions . . . . . . . . . . . . . . . . . 3
0.3.1 Gamma Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.3.2 Beta Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.3.3 The Normal Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.3.4 Chi-square Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
0.3.5 The Fisher ’F distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
0.3.6 Student’s” t Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
0.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
0.4.1 Distribution of sample variance . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
0.4.2 The distribution of the quotient of two variances. . . . . . . . . . . . . . . . . . 7
X̄−µ
0.4.3 Distribution of the quantity Z = √
s/ n
. . . . . . . . . . . . . . . . . . . . . . . 7

8
Bibliography

[1] Miller Freud ,Probability and statistics seventh edition ,Richard A.Johnson 2005.

[2] Geral Baillargeon ,Probabitity ,SMG


[3] Gilbert Sapota Probabilites Analyse des Donnees et statistiques seconde editions Technip 2006
[4] Bruce E.Hansen,Econometrics, University of Wisconsin, January 2018.
[5] Bernard Haudeville,Econometrie AppliqueeESTEM ,1996

[6] N.Piskunov.Differential and integral calculus


[7] Franklin A.Graybill,Alexandre M.MoodAN Introduction theory of statistics McGraw-Hill 1963

You might also like