Distribuições de Probabilidade Alfredo D. Egídio dos Reis ISEG, Abril de 1999 Conteúdo 1 Distribuições Discretas 1.1 Bernoulli (p) . . . . . . 1.2 Binomial (n, p) . . . . . 1.3 Binomial Negativa (k, p) 1.4 Geométrica(p) . . . . . . 1.5 Hipergeométrica(m, n, k) 1.6 Poisson (λ) . . . . . . . 1.7 Uniforme discreta (n) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 1 1 1 1 1 2 2 Distribuições Contínuas 2.1 Beta (α, β) . . . . . . . . . . . . . . . . 2.2 Cauchy (α, β) . . . . . . . . . . . . . . . 2.3 Exponencial (β) . . . . . . . . . . . . . . 2.3.1 Combinação de Exponenciais . . 2.4 Exponencial Dupla ou de Laplace (α, β) 2.5 F -Snedcor, F (m, n) . . . . . . . . . . . 2.6 Gama (α, β) . . . . . . . . . . . . . . . . 2.7 Logística (α, β) . . . . . . . . . . . . . . 2.8 Gaussiana inversa (α, β) . . . . . . . . . 2.9 Lognormal (µ, σ 2 ) . . . . . . . . . . . . 2.10 Normal (µ, σ) . . . . . . . . . . . . . . . 2.11 Pareto (α, β) . . . . . . . . . . . . . . . 2.12 Pareto Generalizada (α, β, θ) . . . . . . 2.13 Qui-Quadrado χ2 (n) . . . . . . . . . . . 2.14 t-Student, t(n) . . . . . . . . . . . . . . 2.15 Uniforme (a, b) . . . . . . . . . . . . . . 2.16 Weibull (α, β) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 2 2 3 3 4 4 5 5 5 6 7 7 7 8 8 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i 1 Distribuições Discretas 1.1 Bernoulli (p) f (x) = P r(X = x) = px (1 − p)1−x , x = 0, 1 (0 < p < 1) E[X] = p, V [X] = p(1 − p), E[etX ] = (1 − p) + pet Notas: Bernoulli(p) ≡ Binomial(1; p) 1.2 Binomial (n, p) f (x) = µ n x ¶ px (1 − p)n−x , x = 0, 1, 2, · · · , n (0 < p < 1; n = 1, 2, · · · ) £ ¤n E[X] = np, V [X] = np(1 − p), E[etX ] = pet + 1 − p Notas: Binomial(1; p) ≡ Bernoulli(p) 1.3 Binomial Negativa (k, p) f (x) = µ k+x−1 x ¶ pk (1 − p)x , x = 0, 1, 2, · · · (k = 1, 2, · · · , 0 < p < 1 k(1 − p) k(1 − p) E[etX ] = E[X] = V [X] = p p2 µ p 1 − (1 − p)et ¶k , t < − log(1 − p) Notas: BN(1, p)≡Geométrica(p) 1.4 Geométrica(p) f (x) = p(1 − p)x , x = 0, 1, 2, · · · (0 < p < 1) x X F (x) = p (1 − p)i = 1 − (1 − p)x+1 x = 0, 1, 2, · · · i=0 E[X] = (1 − p) p (1 − p) E[etX ] = , t < − log(1 − p) V [X] = p p2 1 − (1 − p)et Notas: Geom(p)≡BN(1, p) 1.5 Hipergeométrica(m, n, k) f (x) = µ m x ¶µ µ n k n−m k−x ¶ E[X] = 1.6 ¶ , x = 0, 1, 2, · · · , k , µ m−n+k ≤x≤m n, m, k ≥ 0 KM KM (N − M )(N − K) , V [X] = N N N (N − 1) Poisson (λ) λx , x = 0, 1, 2, · · · (λ > 0) x! t E[X] = V [X] = λ E[etX ] = eλ(e −1) f (x) = e−λ 1 ¶ 1.7 Uniforme discreta (n) 1 , x = 1, 2, · · · , n (n = 1, 2, · · · ) n x , x = 0, 1, 2, · · · , n n f (x) = F (x) = n X1 et (1 − ent ) (n2 − 1) n+1 V [X] = E[etX ] = ekt = 2 12 n n(1 − et ) E[X] = k=1 2 2.1 Distribuições Contínuas Beta (α, β) f (x) = B (α, β)−1 xα−1 (1 − x)β−1 , 0 < x < 1 (α, β > 0) Z 1 Γ(α)Γ(β) β−1 xα−1 (1 − x) dx = B (α, β) = Γ(α + β) 0 E[X] = 2.2 B(α + k, β) α αβ E[X k ] = V [X] = 2 α+β B(α, β) (α + β) (α + β + 1) Cauchy (α, β) f (x) = α ³ ´ , − ∞ < x < +∞ (α > 0, − ∞ < β < +∞) 2 2 π α + (x − β) E[X] : não existe V [X] : não existe E[etX ] : não existe Notas: Cauchy(1; 0) ≡ t(1) y 0.08 0.07 0.06 0.05 0.04 0.03 -6 -4 -2 0 2 4 Densidade da Cauchy(2; 0) 2.3 Exponencial (β) f (x; β) = βe−β x , x > 0 (β > 0) F (x) = 1 − e−β x E[X] = 1 β 1 V [X] = 2 E[etX ] = β β −t β Notas: Exp(β) ≡ Gama(1, β) 2 6 x y 1.0 0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 Densidade da Exp(1) 2.3.1 Combinação de Exponenciais y 2.0 y 2.0 1.5 1.5 1.0 1.0 0.5 0.5 0.0 0 1 2 3 4 0.0 5 x 0 1 f (x) = e−x y 2 3 f (x) = 2e−2x 2.0 1.5 1.0 0.5 0.0 0 1 2 3 4 5 x f (x) = 2e−x − 2e−2x 2.4 Exponencial Dupla ou de Laplace (α, β) f (x) = β −β|x−α| , − ∞ < x < +∞ e 2 (−∞ < α < +∞, β > 0) E[X] = α V [X] = 2/β 2 E[etX ] = 3 β 2 eαt , |t| < β β 2 − t2 4 5 x y 0.5 0.4 0.3 0.2 0.1 -4 -3 -2 -1 0 1 2 3 4 Densidade da Laplace(0; 1) 2.5 F -Snedcor, F (m, n) f (x) = B (m/2, n/2) E[X] = E[etX ] : −1 ¡ m ¢m/2−1 m nx , x > 0 (m, n > 0) ¢ ¡ (m+n)/2 n 1+ m nx n 2n2 (m + n − 2) , n > 2; V [X] = , n > 4; n−2 m(n − 2)2 (n − 4) não existe Notas: X ∼ F (m, n) ⇒ X −1 ∼ F (n, m); F = χ2 (m)/m χ2 (n)/n ∼ F (m, n) (var. independentes) y 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 1 2 3 4 Densidade da F (10; 4) 2.6 Gama (α, β) β α α−1 −β x e , x > 0 (α, β > 0) x Γ(α) Z ∞ Γ(α) = β α xα−1 e−β x dx f (x) = 0 Para α = n inteiro positivo: Γ(n) = (n − 1)! e F (x|n, β) = 1 − E[X] = α α V [X] = 2 E[etX ] = β β Notas: Gama(1, β) ≡ Exp(β) 4 µ β β−t ¶α E[X k ] = n−1 X k=0 k (βx) e−βx k! Γ(α + k) β α Γ(α) f(x) f(x) x x Densidade da Gama(2; 2) 2.7 Densidade da Gama(3; 5) Logística (α, β) βe−β(x−α) ¡ ¢2 , − ∞ < x < +∞ 1 + e−β(x−α) ³ ´−1 F (x) = 1 + e−β(x−α) f (x) = E[X] = α V [X] = 2.8 (−∞ < α < +∞, β > 0) π2 E[etX ] = eαt Γ(1 − t/β)Γ(1 + t/β), |t| < β 3β 2 Gaussiana inversa (α, β) ½ ¾ 1 2 f (x) = α (2πβ) x exp − (βx − α) , x > 0 2βx r ½ µ ¶¾ 2t E[etX ] = exp α 1 − 1 − β −1/2 −3/2 (α, β) y Densidade da GI(α; β), α = β = 2−2 , 2−1 , 1, 2, 22 , 23 , 24 , 25 2.9 Lognormal (µ, σ 2 ) 2 2 1 √ e−(log x−µ) /(2σ ) , − ∞ < x < +∞ xσ 2π µ ¶ log x − µ F (x) = Φ σ f (x|µ, σ 2 ) = 5 (−∞ < µ < +∞, σ > 0) E[X] = eµ+σ 2 E[X k ] = ekµ+k /2 2 V [X] = e2µ+σ σ 2 /2 2 , k = 1, 2, ... ´ ³ 2 E[etX ] : não existe eσ − 1 Notas: Se X ∼LN(µ, σ 2 ) ⇔ Y = log X ∼N(µ, σ 2 ) y Densidade da LN(0; 1) 2.10 Normal (µ, σ) ¡ ¢−1/2 − 1 (x−µ)2 f (x) = 2πσ 2 e 2σ2 , −∞ < x < +∞ (−∞ < µ < +∞; σ > 0) 1 E[X] = µ; V [X] = σ 2 ; E[etX ] = eµt+ 2 σ 2 2 t 0.4 0.3 0.2 0.1 -3 -2 -1 0 1 2 3 x Densidade da Normal(0; 1) -3 -2 -1 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 1 2 3 4 -4 x Densidades da Normal(0; 1) e Normal(1; 1) -3 -2 -1 0 1 2 3 4 x Densidades da Normal(0; 1) e Normal(0; 1.25) 6 2.11 Pareto (α, β) αβ α , x > 0 (α, β > 0) (β + x)α+1 ¶α µ β ,x>0 F (x) = 1 − β+x f (x) = E[X] = β αβ 2 , α > 1; V [X] = , α > 2; F.g.m. não existe (α − 1) (α − 1)2 (α − 2) y 2.0 1.5 1.0 0.5 0.0 0 1 2 3 4 Densidade da Pareto(2; 1) Notas: • Se X ∼ Pareto(α, β) ⇒ Y = log (1 + X/β) ∼exp(α); • Se X|θ ∼ exp(θ) e θ 2.12 Gama(α, β) ⇒ X Pareto(α, β). Pareto Generalizada (α, β, θ) Γ (α + θ) αβ α xθ−1 , x > 0 (α, β > 0) Γ (α) Γ (θ) (β + x)α+θ µ ¶ x F (x) = β θ, α; ,x>0 β+x f (x) = E[X k ] = 2.13 β k Γ (θ + k) Γ (α − k) , −θ <k <α Γ (α) Γ (θ) E[etX ] : não existe Qui-Quadrado χ2 (n) ³ ´−1 f (x) = Γ(n/2)2n/2 xn/2−1 e−x/2 , x > 0 (n = 1, 2, ...) E[X] = n V [X] = 2n E[etX ] = Notas: χ2 (n) ≡ Gama(n/2, 1/2) 7 µ 1 1 − 2t ¶n/2 , t< 1 2 2.14 t-Student, t(n) µ ¶−(n+1)/2 x2 , −∞ < x < +∞, (n = 1, 2, ...) 1+ n Γ( n+1 ) f (x) = √ 2 n nπΓ( 2 ) n , n>2 n−2 Γ((k + 1) /2)Γ((n − k) /2) 1 · 3 · · · · · (k − 1) E[X k ] = nk/2 = nk/2 , n > k, se k par Γ(1/2)Γ(n/2) (n − k) (n − k + 2) · · · (n − 2) E[X] = 0, n > 1; V [X] = E[X k ] = 0, se kímpar E[etX ] : não existe Notas: X e Y independentes, T = N (0;1) t χ2(n) /n ∼ t(n); t(n) 2 ≡ F(1,n) y 0.35 y 0.30 0.4 0.3 0.25 0.20 0.2 0.15 0.10 0.1 0.05 -3 -2 -1 0 1 2 3 -3 x Densidades da t(1) [...] e t(3) [–] 2.15 -1 0 F (x) = E[X] = Notas: U(0; 1) ≡ Beta(1; 1) 1 2 Densidades da t(5) [–] e N (0; 1) [...] Uniforme (a, b) f (x) = 2.16 -2 1 , a < x < b (−∞ < a < b < ∞) b−a x−a , a≤x≤b b−a a+b ebt − eat (b − a)2 V [X] = E[etX ] = 2 12 (b − a)t Weibull (α, β) α f (x) = αβxα−1 e−βx , x > 0 (α, β > 0) α F (x) = 1 − e−βx , x > 0 Γ (1 + 1/α) E[X] = β 1/α Γ (1 + k/α) E[X k ] = β k/α 8 3 x