summaryrefslogtreecommitdiff
path: root/PBS1/Couple Aléatoire.md
blob: c733165a935e2ad62a7b5e577adb6bb568970904 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
# Couple aléatoire
## Définition

Un couple aléatoire discret est un couple $(X,Y)$ de V.A. définies sur le même univers $\Omega$ et à valeurs dans $X(\Omega) \times Y(\Omega)=\{(x,y):x \in X(\Omega), y \in Y(\Omega)\}$

On note $\{ X = x, Y = y \}$ l'évènement $X = x \cap Y = y$
## Loi du couple
On appelle loi de probabilité ou loi jointe du couple $(X,Y)$ l'application
$$
\begin{align}
\mathbb{P}_{XY}: \qquad X(\Omega) \times Y(\Omega) &\longrightarrow [0,1] \\
(x,y) &\longmapsto \mathbb{P}_{XY}(x,y) = \mathbb{P}(X=x,Y=y)
\end{align}
$$
## Fonction de répartition
La connaissance de $\mathbb{P}_{XY}$ équivaut à celle de la fonction de répartition définie pour tout $(x,y) \in \mathbb{R}^2$ par :
$$
F_{XY}(x,y) = \mathbb{P}(X \leq x, Y\leq y)
$$
## Lois marginales
On appelle loi marginale de $X$ l'application $\mathbb{P}_{X}$ de $X(\Omega)$ dans $[0,1]$ définie pour tout $x \in X(\Omega)$ par
$$
\mathbb{P}_{X}(x) = \mathbb{P}(X = x) = \sum_{y \in Y(\Omega)} \mathbb{P}_{XY}(x,y)
$$
Par analogie, on a
$$
\mathbb{P}_{Y}(y) = \mathbb{P}(Y = y) = \sum_{x \in X(\Omega)} \mathbb{P}_{XY}(x,y)
$$
## Lois conditionnelles
Soit $(X,Y)$ un couple aléatoire discret. On appelle loi conditionnelle de $X$ sachant $Y$ l'application $\mathbb{P}_{X|Y}$ de $X(\Omega)$ dans $[0,1]$ définie pour tout couple $(x,y) \in X(\Omega) \times Y(\Omega)$ par
$$
\mathbb{P}_{X|Y}(x|y) = \mathbb{P}(X = x | Y = y) = \frac{\mathbb{P}_{XY}(x,y)}{\mathbb{P}_{Y}(y)}
$$
De façon analogue, on a $\mathbb{P}_{Y|X}$ :
$$
\mathbb{P}_{Y|X}(y|x) = \mathbb{P}(Y = y | X = x) = \frac{\mathbb{P}_{XY}(x,y)}{\mathbb{P}_{X}(x)}
$$
## Indépendance
On dit que $X$ et $Y$ sont **indépendantes** ssi pour tout $x \in X(\Omega)$ et tout $y \in Y(\Omega)$,
$$
\mathbb{P}(X = x | Y = y) = \mathbb{P}(X = x) \quad et \quad \mathbb{P}(Y = y | X = x) = \mathbb{P}(Y = y)
$$
## Covariance
Si $X$ et $Y$ sont définies sur le même $\Omega$, on appelle **covariance** de ces deux variables le réel :
$$
Cov(X,Y) = \mathbb{E}[(X - \mathbb{E}(X))(Y - \mathbb{E}(Y))]
$$
Ou aussi $Cov(X,Y) = \mathbb{E}(X \cdot Y) - \mathbb{E}(X) \cdot \mathbb{E}(Y)$
### Interprétation
- $Cov(X,Y) > 0 \implies$ Plus $X$ grandit, plus $Y$ grandit aussi
- $Cov(X,Y) < 0 \implies$ Quand $X$ augmente, $Y$ diminue (et vice-versa)
- $Cov(X,Y) \approx 0 \implies$ Les 2 variables ne sont pas vraiment liées

## Corrélation
$$
\rho(X,Y) = \frac{Cov(X,Y)}{\sigma_{X}\sigma_{Y}}
$$
Dépendance linéaire entre $X$ et $Y$.
![[Pasted image 20250414173315.png]]
# Couple aléatoire à densité
## Propriétés d'une densité
Une fonction $f : \mathbb{R}^2 \rightarrow \mathbb{R}_+$ est une densité de probabilité si elle vérifie les conditions suivantes :
- **Positivité** : $f(x,y)\geq 0 \quad \forall (x,y) \in \mathbb{R}^2$
- **Normalisation** : $\int_{\mathbb{R}}\int_{\mathbb{R}}f(x,y)dxdy = 1$
## Couple aléatoire à densité
Un **couple aléatoire à densité** est un couple de variables aléatoires $(X,Y)$ défini sur un espace probabilisé et possédant une **densité conjointe $f_{X,Y}(x,y)$**.
Celà signifie que la probabilité que $(X,Y)$ appartienne à une région $A \subset \mathbb{R}^2$ est donnée par :
$$
P((X,Y) \in A) = \int_{\mathbb{R}}\int_{\mathbb{R}}f(x,y)dxdy
$$

$$
P(a \leq X \leq b ; c \leq Y \leq d) = \int^{d}_{c}\int^{b}_{a}f(x,y)dxdy
$$
## Fonction de répartition
$$
F_{X,Y}(x,y) = P(X \leq x ; Y \leq y) = \int^{y}_{-\infty}\int^{x}_{+\infty}f(x,y)dxdy
$$
## Lois jointes et marginales
Si un couple $(X,Y)$ a pour densité jointe la fonction $f$, on a :
$$
\begin{align}
	\forall x \in X(\Omega), \quad f_{X}(x) = \int^{+\infty}_{-\infty} f(x,y)dy \\
	\forall y \in Y(\Omega), \quad f_{Y}(y) = \int^{+\infty}_{-\infty} f(x,y)dx \\
\end{align}
$$
## Densités conditionnelles
Soit $(X,Y)$ un couple aléatoire de densité jointe $f_{X,Y}$ et $f_X$, $f_Y$ les densités marginales de **X** et **Y** respectivement. On appelle densité conditionnelle de $X$ sachant $Y = y$ par :
$$
\forall y \in \mathbb{R}, \forall x \in \mathbb{R}, f_{X}(x) \neq 0, f_{Y|X}(x,y) = \frac{f(x,y)}{f_{Y}(y)}
$$
De façon analogue, on a $Y$ sachant $X = x$ :
$$
\forall x \in \mathbb{R}, \forall y \in \mathbb{R}, f_{Y}(y) \neq 0, f_{X|Y}(x,y) = \frac{f(x,y)}{f_{X}(x)}
$$
## Indépendance
C.N.S. :
$$
\forall (x,y) \in X(\Omega) \times Y(\Omega), f(x,y) = f_{X}(x)f_{Y}(y)
$$
On a alors $f_{Y|X}(y,x) = f_{Y}(y)$ et $f_{X|Y}(x,y) = f_{X}(x)$

## Moments de couple
Soit $Z = \begin{pmatrix}X \\ Y\end{pmatrix}$ un vecteur aléatoire de densité $f$.
- Espérance $\mathbb{E}(Y | X = x) = \int^{+\infty}_{-\infty}yf_{Y|X=x}(y)dy$
- Espérance $\mathbb{E}(Y) = \int_{\mathbb{R}}\mathbb{E}(Y|X=x)f_{X}(x)dx = \int \int_{\mathbb{R}^2}yf(x,y)dydx$
- Espérance $\mathbb{E}(Z) = \int \int_{\mathbb{R}²}\begin{pmatrix}x \\ y\end{pmatrix} f(x,y)dxdy = \begin{pmatrix}\int \int_{\mathbb{R}²}xf(x,y)dxdy \\ \int \int_{\mathbb{R}²}yf(x,y)dydx\end{pmatrix} = \begin{pmatrix}\mathbb{E}(X) \\ \mathbb{E}(Y)\end{pmatrix}$
- Covariance $Cov(Z) = \begin{pmatrix}\mathbb{V}(X) & Cov(X,Y) \\ Cov(X,Y)& \mathbb{V}(Y)\end{pmatrix}$
	- $Cov(X,Y) = \mathbb{E}((X - \mathbb{E}(X)) \times (Y - \mathbb{E}(Y))) = \int \int_{\mathbb{R}²}(x - \mathbb{E}(X))(y - \mathbb{E}(y))f(x,y)dxdy$
	- $\mathbb{V}(X) = \mathbb{E}((X - \mathbb{E}(X))²), \qquad \mathbb{V}(Y) = \mathbb{E}((Y - \mathbb{E}(Y))²)$

## Propriétés
- $\forall (a,b) \in \mathbb{R}², Var(aX + bY) = a²Var(X) + 2abCov(X,Y) + b²V(Y)$
- $Cov(X,X) = Var(X)$
- Symétrie : $Cov(X,Y) = Cov(Y,X)$
- Bilinéarité : Soient $a,b \in \mathbb{R}$. Alors $Cov(aX+bY,Z) = aCov(X,Y) + b(Cov(Y,Z))$ et $Cov(X,aY+bZ) = aCov(X,Y) + bCov(X,Z)$
- Indépendance : Si $X$ et $Y$ sont indépendantes, alors $Cov(X,Y) = \mathbb{E}(XY) - \mathbb{E}(X)\mathbb{E}(Y) = 0$ (en revanche $Cov(X,Y) = 0)$ n'implique pas forcément l'indépendance de $X$ et $Y$)
- $|Cov(X,Y)| \leq \sigma(X)\sigma(Y)$ (égalité $\Leftrightarrow Y = aX + b$)
## Corrélation
Soient $X$ et $Y$ des v.a. de moyennes finies et de variance non nulles. Le **coefficient de corrélation** de $X$ et $Y$ est alors : $r(X,Y) = \frac{Cov(X,Y)}{\sigma(X)\sigma(Y)}$.
$\rightarrow$ Mesure la relation linéaire entre 2 variables
Si :
- $r(X,Y) = 0$ alors $X$ et $Y$ sont "non corrélées"
- $r(X,Y) > 0$ alors $X$ et $Y$ sont positivement corrélées
- $r(X,Y) = 1$ alors $X$ et $Y$ sont parfaitement positivement corrélées
- $r(X,Y) < 0$ alors $X$ et $Y$ sont négativement corrélées
- $r(X,Y) = 1$ alors $X$ et $Y$ sont parfaitement négativement corrélées
![[Pasted image 20250509112837.png]]

## Loi d'une somme de variables continues indépendantes
![[Pasted image 20250509112950.png]]
![[Pasted image 20250509113011.png]]