[math] discrete distributions

[Phil Steitz]

I have started work extending the distribution framework to accomodate 
discrete probablity distributions so that we can provide binomial, 
hypergeometric and poisson probabilities and I want to get others' (esp. 
Brent's) opinions on the following plan:

1. Add DiscreteDistribution and AbstractDiscreteDistribution similar to
   the continuous case, but with DiscreteDistribution defining methods
   appropriate for the discrete case -- e.g p(X = x), p(X <= x),
   p(X < x), p(a <= X < b), inf{x: P(X < x) > c} etc.
2. Define discrete distribution interfaces and implementations for
   binomial, hypergeometric and poisson distributions.
3. Extend DistributionFactory, DistributionFactoryImpl to create
   discrete distribution instances by just adding factory methods for
   them.
Step 3. is where I need confirmation.  One could argue that we should 
split off a separate DiscreteDistributionFactory. Thoughts?

Phil



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]