Binomial/Poisson Distributions Summary Binomial Poisson: Independent Trials, Whereby Each Trial

Binomial/Poisson Distributions Summary Binomial
Characteristics: (i) n independent trials, whereby each trial yields two possible outcomes- a fixed constant probability of success p or failure q = 1 p (ii) Purely discrete
Poisson
Characteristics: (i) Events occur randomly and singly (ii) A single parameter (average number of occurences) defines the distribution and is proportional to the frame of measurement ( eg = 1 for 1 week and therefore = 2 for 2 weeks) (iii) Purely discrete Definition: X ~ P0 ( )
r nr
Definition: X ~ B ( n, p) Formula:
P ( X = r ) = n C r ( p ) (q )
Formula:
P( X = r ) =
Mean: = np
Variance: 2 = npq
Mean: =
e ( ) r! Variance: 2 =
r
Separate binomial distributions typically cannot be pooled together
For two independent poisson distributions defined under the same frame of measurement, eg X ~ P0 ( ) and Y ~ P0 ( ) , they can be pooled together to form a consolidated poisson model whereby X + Y ~ P0 ( + ) Example Scenario: number of defects along a 10m long piece of cloth manufactured in a factory. Graphic calculator commands : P ( X = r ) poissonpdf P ( X r ) poissoncdf
Example Scenario: number of sixes obtained during ten throws of an unbiased dice. Graphic calculator commands : P ( X = r ) binompdf P ( X r ) binomcdf
(Note that pdf stands for probability density function, while cdf stands for cumulative density function.) Binomial to Poisson Approximation: For a binomial distribution whereby X ~ B ( n, p) , IF np < 5 and p < 0.1 , then X ~ P0 (np, npq ) approximately. (Note that there is NO poisson to binomial approximation) NO continuity correction is required since this is a discrete to discrete approximation.
Some important miscellaneous concepts: (a) It is advisable, amongst all things, to appreciate and convert to memory that P( x = 0) = q n for a binomial distribution and P ( X = 0) = e for a poisson distribution. Example of its relevance: For X ~ B ( n, 0.05), find the least value of n such that P ( X 1) > 0.99 without employing any explicit GC statistical commands. Workings: P ( X 1) > 0.99 1 P ( X = 0) > 0.99 1 (1 0.05) n > 0.99 (continue the solving process)
(b) It is very usual for questions to embed a binomial/poisson distribution within another binomial/poisson distribution-students must be sufficiently discerning to relate one part of the context to another. Put it simply, such questions are multi-layered. Example: Eggs are packed in boxes of 500. On average, 0.8% of the eggs are found to be broken when the eggs are unpacked. (i) Find the probability that in a box of 500 eggs, exactly 3 will be broken. Let the random variable X denote the number of broken eggs within a box of 500. Then X ~ B (500, 0.008) (ii) A supermart unpacks 100 boxes of eggs. What is the probability that there will be exactly 4 boxes containing no broken eggs?
Now we are zooming out and focusing on a more general picture of boxes of eggs. Let the random variable Y denote the number of boxes containing no broken eggs. Then Y ~ B (100, p) where
p = P ( X = 0) based on the random variable X defined in (i).
(c) While less common, finding the mode of a distribution via non GC methods can be examined-a detailed worked example will illustrate this better:
The random variable is the number of successes in 200 independent trials of an experiment in which the probability of success at any one trial is . Given that ( ) = 10.6008, find the exact value of and show that
for
= 0,1,2, . . . . . . . . . ,199
(i) Hence find the value of such that ( = ) is the maximum. (ii) Using a Poisson approximation, find the probability that more than 198 of the 200 trials were not successful.
SOLUTIONS : ~ (200, ) Since ( Then ) = 10.60008, ( )= ( ) [ ( )] = 10.6008 (200p)
200( )(1 ) = 10.6008 (200p) 200p 200 = 10.6008 40000p
39800p + 200p 10.6008 = 0 Solving gives = 0.014 (shown) 200 k + 1 200 k (1 ) (1 )
( = + 1) = ( = )
200 k + 1 . = 200 k .
( . ( . )
)!( ( )!(
! !
)! )!
. .
( . ( . )
) .
(k+1) 986
200-k 14
(shown)
(i)
If
( =
( (
+ 1) > ( = ),
) )
then
>1
>1
< 1.814
where solving the inequality gives
Hence, we have ( = 2) > ( = 1) > ( = 0)------------------(1) On the other hand, if then

( ( ) )
( =
( (
+ 1) < ( = ),
) )
<1
<1
> 1.814
where solving the inequality gives Hence, we have
( = 200) < ( = 199) < ( = 198). . . . . . . . . . . . . < ( = 4) < ( = 3) < ( = 2)----------(2) Reconciling (1) and (2) therefore gives ( = 2)as the maximum, ie (ii) Since ~ = 200(0.014) = 2.8 < 5 and (2.8) approximately = 0.014 < 0.1, = 2 (shown)
( more than 198 of the 200 trials were not successful) = ( 1) = 0.231(shown) (Note: Care must be exercised in interpreting the random variable correctly; was defined from the start as the number of successful trials, NOT unsuccessful trials .)

Binomial/Poisson Distributions Summary Binomial Poisson: Independent Trials, Whereby Each Trial

Uploaded by

Copyright:

Available Formats

Binomial/Poisson Distributions Summary Binomial Poisson: Independent Trials, Whereby Each Trial

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Binomial/Poisson Distributions Summary Binomial Poisson: Independent Trials, Whereby Each Trial

Uploaded by

Copyright:

Available Formats

Binomial/Poisson Distributions Summary Binomial

Separate binomial distributions typically cannot be pooled together

SOLUTIONS : ~ (200, ) Since ( Then ) = 10.60008, ( )= ( ) [ ( )] = 10.6008 (200p)

200( )(1 ) = 10.6008 (200p) 200p 200 = 10.6008 40000p

39800p + 200p 10.6008 = 0 Solving gives = 0.014 (shown) 200 k + 1 200 k (1 ) (1 )

where solving the inequality gives

Hence, we have ( = 2) > ( = 1) > ( = 0)------------------(1) On the other hand, if then

where solving the inequality gives Hence, we have

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.