Harvard statistics, video 11 note(Poisson Distribution)

7:39 2014-10-11 Saturday
start Harvard statistics, video 11


the Poisson distribution


7:39 2014-10-11
sympathetic magic:


don't confuse r.v. with its distribution


7:46 2014-10-11
"word is not the thing, map is not the territory"


7:48 2014-10-11
the r.v. is a house,


the distributuion is a blue print


7:51 2014-10-11
Poisson distribution 


7:54 2014-10-11
rate parameter


7:57 2014-10-11
1st let's check this is a valid PMF


7:57 2014-10-11
X ~ Pois(λ)


8:00 2014-10-11
often used for the application where


counting the #of successes, where we have a large


number of trials, each one with a small probability 


of success


* #emails receive per hour


8:06 2014-10-11
we have a large number of things that could happen, 


but each one is unlikely // Poisson distribution


8:12 2014-10-11
weakly dependent


8:12 2014-10-11
Poisson Paradigm(Poisson Approximation)


8:15 2014-10-11
Bernoulli(p)


8:16 2014-10-11
we can prove Binomial(n, p) does converge to Poisson


when n gets large & p gets small


8:16 2014-10-11
Binomial => Poisson // Binomial converges to Poisson


8:32 2014-10-11
create a indicator r.v.


8:41 2014-10-11
now we're going to use the "Poisson Approximation"