Re: Probability question
- From: "Reef Fish" <Large_Nassau_Grouper@xxxxxxxxx>
- Date: 6 Apr 2006 17:10:59 -0700
Bill Howells wrote:
Eh, not that long ago, just a beginner. But if I can't do a simple
binomial probability, I'd better hang it up right now!
Bigw, you made me curious with the Poisson,
which is me, in French.
which I had to check. But
it looks like the Poisson distribution is derived from the Binomial by
letting n -> infinity
Much more than that requirement. :-)
so with n=4, it is probably not a good
approximation. I calculated with a Poisson mean of lambda = n*p =
4*(2/3) and a count of k=3 events, I get a Poisson probability of about
0.18. Contrast this with the p = 32/81 = 0.395 from the Binomial, and
you can see the Poisson is not working very well.
That's another pain in the neck and time-wastin' rules of thumbs
in textbooks written by author who had nothing better to do.
Discrete Poisson probabilities no matter how many terms or how
large or small n is, can always be equated EXACTLY to an
equivalent tail probability of a well-known continuous random
variable.
I am purposely leaving out all the details of the correspondence.
-- Bob. Le poisson; ou poissons de récif
.
- Follow-Ups:
- Re: Probability question
- From: Bill Howells
- Re: Probability question
- References:
- Probability question
- From: bigwags
- Re: Probability question
- From: Bill Howells
- Re: Probability question
- From: Reef Fish
- Re: Probability question
- From: Bill Howells
- Probability question
- Prev by Date: Re: multivariate linear regression
- Next by Date: Re: Regressio line problem
- Previous by thread: Re: Probability question
- Next by thread: Re: Probability question
- Index(es):
Relevant Pages
|