maxent

Compute Maximum Entropy Distributions.

The maximum entropy method, or MAXENT, is variational approach for computing probability distributions given a list of moment, or expected value, constraints.

Here are some links for background info.

A good overview of applications: http://cmm.cit.nih.gov/maxent/letsgo.html

On the idea of maximum entropy in general: http://en.wikipedia.org/wiki/Principle_of_maximum_entropy

Use this package to compute discrete maximum entropy distributions over a list of values and list of constraints.

Here is a the example from Probability the Logic of Science

> maxent 0.00001 [1,2,3] [average 1.5]
 Right [0.61, 0.26, 0.11]

The classic dice example

> maxent 0.00001 [1,2,3,4,5,6] [average 4.5]
 Right [.05, .07, 0.11, 0.16, 0.23, 0.34]

One can use different constraints besides the average value there.