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 Document
Unformatted text preview: MA2216/ST2131 Probability Notes 4 § 1. Discrete Uniform Distribution. 1. If X assumes the values x 1 < x 2 < ... < x n , with equal probabilities, then X is called a discrete uniform r.v. The probability distribution of X is called the discrete uniform distribution . The probability density function of X is given by f X ( x ) = 1 n , x = x 1 ,x 2 ,...,x n . 2. The mean and variance of X are, respectively, EE ( X ) = 1 n n X i =1 x i , Var( X ) = 1 n n X i =1 x 2 i ˆ 1 n n X i =1 x i ! 2 . 3. Distribution Function. The c.d.f. F X ( x ) is given by F X ( x ) = IP { X ≤ x } = X i : x i ≤ x f X ( x i ) = # of x i ’s with x i ≤ x n (to be more explicit) = , for x < x 1 , j n , for x j ≤ x < x j +1 for 1 ≤ j < n , 1 , for x n ≤ x . (It is obvious to see that F X ( · ) is a step function which is rightcontinuous with left limits.) 1 § 2. Geometric Distribution. 1. The model which generates a geometric distribution can be described as follows: Conduct independent Bernoulli trials, IP(success) = p, IP(failure) = q = 1 p, < p < 1 , until a success is obtained. Let X be the number of trials required. Then, f X ( n ) = IP( X = n ) = q n 1 p, n = 1 , 2 ,..., and X is called a geometric r.v. with parameter p , denoted by X ∼ Geom( p ) . X is also said to have a geometric distribution . Note that f X ( n ) is indeed a genuine p.d.f., for f X ( n ) > 0 and ∞ X n =1 q n 1 p = p ∞ X n =0 q n = p 1 q = 1 . Obviously, the above infinite sum has to do with a geometric series, which explains why it is called a geometric distribution. Remark. Recall Ex.1 in Tut. 1, which can be rephrased as a geometric distribution model. 2. Example: During busy hours, a telephone exchange is very near capac ity, so people cannot find a line to use. If the probability of a connection during busy time is p = 0 . 1, what is the probability that at least 5 attempts are needed for a successful call? Sol. Let X denote the number of attempts so as to get the first success. It is obvious to see that X has a geometric distribution with parameter p = 0 . 1. The desired probability is given by IP { X ≥ 5 } = ∞ X x =5 p (1 p ) x 1 = (1 p ) 5 1 = (0 . 9) 4 . 2 Note. Here, we actually derive a useful formula for geometric r.v.’s. Namely, IP { X ≥ k } = ∞ X x = k p (1 p ) x 1 = (1 p ) k 1 = q k 1 , (2 . 1) for k = 1 , 2 ,.... It is quite straightforward to figure out a “ probabilistic ” derivation. The event { X ≥ k } simply means that each of the first k 1 trials results in a failure, etc. 3. For a geometric r.v. X with parameter p , EEX = 1 p , Var( X ) = EE [ X 2 ] [ EEX ] 2 = 1 p p 2 . (2 . 2) It is wellknown that ∞ X n =1 q n 1 = ∞ X n =0 q n = 1 1 q ....
View
Full Document
 Fall '08
 Forgot
 Normal Distribution, Probability theory, RK, Qn

Click to edit the document details