# l5 - Outline Discrete Distributions Oct 4 2011 Discrete...

This preview shows pages 1–11. Sign up to view the full content.

Outline Discrete Distributions Oct 4, 2011 Discrete Distributions

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Outline Outline 1 Distributions pt 1 2 Distributions pt 2 Discrete Distributions
Distributions pt 1 Distributions pt 2 Bernoulli Binomial Outline 1 Distributions pt 1 2 Distributions pt 2 Discrete Distributions

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Distributions pt 1 Distributions pt 2 Bernoulli Binomial Up to now, we’ve deﬁned the pdf’s for our R.V.s by listing out all the possible values the R.V. could take and giving a probability for each outcome Rather cumbersome in real life What we’d love is if a small number of parameters in some function could deﬁne the probabilities for us E.g., Pr ( X = x | a , b ) where a and b are some ﬁxed parameters We’ll spend the next few weeks talking about some common functions Discrete Distributions
Distributions pt 1 Distributions pt 2 Bernoulli Binomial In many ways this is the simplest distribution. The Bernoulli distribution may be appropriate when there are only two possible outcomes (heads/tails; yes/no; dead/alive) and you only observe it once. Bernoulli pdf: X Bern ( p ) pr ( X = 0 ) = 1 - p pr ( X = 1 ) = p For the Bernoulli distribution the parameter is p is the probability of observing X = 1. The Bernoulli distribution serves as the basis of logistic regression Discrete Distributions

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Distributions pt 1 Distributions pt 2 Bernoulli Binomial Bernoulli PDF and CDF Discrete Distributions
Distributions pt 1 Distributions pt 2 Bernoulli Binomial E [ X ] , V [ X ] E [ X ] = X x xpr ( x ) = 1 × p + 0 × ( 1 - p ) = p V [ X ] = X x ( x - p ) 2 pr ( x ) = ( 0 - p ) 2 ( 1 - p ) + ( 1 - p ) 2 p = p 2 ( 1 - p ) + p ( 1 - p ) 2 = p ( 1 - p ) Discrete Distributions

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Distributions pt 1 Distributions pt 2 Bernoulli Binomial Suppose we want to model some disease using the Bernoulli distribution D i Bernoulli ( p ) , where D i is disease status for individual i . What are we assuming about the disease? Outcomes are independent in the sense that the i th person’s disease status doesn’t effect the j th person’s disease status Risk of disease ( p ) is homogenous Discrete Distributions
Distributions pt 1 Distributions pt 2 Bernoulli Binomial Rather than a single trial, we might have n trials: x 1 ,..., x n . Each R.V. x i follows a Bernoulli dist’n. We might be interested in the R.V. Y = i x i . Binomial Distribution Pr ( Y = y ) = ± n y ² p y ( 1 - p ) n - y Discrete Distributions

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Distributions pt 2 Bernoulli Binomial Its easy to justify the binomial dist’n intuitively. Say n = 2 so we have the sample space S = { x 1 = 1 , x 2 = 1 } , { x 1 = 1 , x 2 = 0 } , { x 1 = 0 , x 2 = 1 } , { x 1 = 0 , x 2 = 0 } . Further, let x i Bern ( p ) . x 1 x 2 y Pr ( X 1 = x 1 ) Pr ( X 2 = x 2 ) Pr ( x 1 , x 2 ) 1 1 2 p p p 2 1 0 1 p 1 - p p ( 1 - p ) 0 1 1 1 - p p p ( 1 - p ) 0 0 0 1 - p 1 - p ( 1 - p ) 2 If we only care about Y = x i , then { x 1 = 1 , x 2 = 0 } and { x 1 = 0 , x 2 = 1 } both yield y = 1 and we don’t need to distinguish between them. We count the number of times that Y = y out of n
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### Page1 / 43

l5 - Outline Discrete Distributions Oct 4 2011 Discrete...

This preview shows document pages 1 - 11. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online