This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: STAT5101 Lecture Notes 9 • Discrete distributions – Bernoulli – Binomial – Geometric – Poisson • Continuous distributions – Normal – Bivariate Normal – Gamma – Exponential – Chisquare • Briefly mention – Negative Binomial – Lognormal – Beta 1 Bernoulli Distribution and Bernoulli trials We say X follows a Bernoulli distribution with parameter p if X take only 0 and 1 and the probabilities are P ( X = 1) = p and P ( X = 0) = 1 p . We write X ∼ Ber ( p ). We have studied the basic properties of Ber ( p ). For example, EX = p , V ar ( X ) = p (1 p ) and its MGF is ψ ( t ) = 1 p + pe t . n Bernoulli trials: n independent Ber ( p ) ran dom variables form n Bernoulli trails. For each Bernoulli trial, if X = 1 then we say we have a success, otherwise we have a failure. So p = P ( X = 1) is also called the probability of success. 2 Binomial Distribution Definition: We say X follows a Binomial dis tribution with parameters n and p if P ( X = k ) = n k p k (1 p ) n k , k = 0 , 1 ,...,n We write X ∼ Bin ( n,p ) . We have studied the basic properties of Bin ( n,p ). For example, EX = np , V ar ( X ) = np (1 p ) and the MGF is ψ ( t ) = (1 p + pe t ) n . The model Suppose we have n Bernoulli trails with probability of success p . Let Y be the number of success in the n Bernoulli trails. Then Y ∼ Bin ( n,p ) . Why? let X i be the ith Ber ( p ) rv in the n Bernoulli trails. Then we observe Y = X 1 + X 2 + ··· + X n 3 Geometric Distribution Definition: We say X follows a Geometric dis tribution with parameter of success p if P ( X = k ) = (1 p ) k p, k = 0 , 1 , 2 ,..., We write X ∼ Geo ( p ) . For convenience we also let q = 1 p . P ( X = k ) = q k p . The model Suppose we keep doing the Bernoulli trails until we have a success. Let X be the number of trials before the success . Then X ∼ Geo ( p ) . MGF of Geo ( p ) ψ ( t ) = E [ e tX ] = ∞ X k =0 e tk q k p = p ∞ X k =0 ( e t q ) k ψ ( t ) = p 1 e t q for t < log(1 /q ) 4 Mean and Variance of Geo ( p ) If X ∼ Geo ( p ) then EX = q p V ar ( X ) = q p 2 . Explanation: ψ (1) ( t ) = p (1 e t q ) 2 qe t EX = ψ (1) (0) = p (1 q ) 2 q = q p ψ (2) ( t ) = p (1 e t q ) 2 qe t + 2 p (1 e t q ) 3 qe t · qe t E ( X 2 ) = ψ (2) (0) = p (1 q ) 2 q + 2 p (1 q ) 3 q 2 E ( X 2 ) = q p + 2 q 2 p 2 V ar ( X ) = E ( X 2 ) ( EX ) 2 = q p 2 5 Poisson Distribution Definition: We say X follows a Poisson dis tribution with parameter λ if P ( X = k ) = e λ λ k k ! , k = 0 , 1 , 2 ,..., We write X ∼ Poi ( λ ) . MGF of Poi ( λ ) : Let X ∼ Poi ( λ ) then its MGF is ψ ( t ) = e λ ( e t 1) . ψ ( t ) = E ( e t X ) = ∞ X k =0 e tk e λ λ k k ! = e λ ∞ X k =0 ( e t λ ) k k !...
View
Full Document
 Fall '02
 Staff
 Statistics, Bernoulli, Binomial

Click to edit the document details