Dirichlet (gatk 4.0.5.0 API)

```
public class Dirichlet
extends java.lang.Object
```
The Dirichlet distribution is a distribution on multinomial distributions: if pi is a vector of positive multinomial weights such that sum_i pi[i] = 1, the Dirichlet pdf is P(pi) = [prod_i Gamma(alpha[i]) / Gamma(sum_i alpha[i])] * prod_i pi[i]^(alpha[i] - 1) The vector alpha comprises the sufficient statistics for the Dirichlet distribution. Since the Dirichlet is the conjugate prior to the multinomial, if one has a Dirichlet prior with concentration alpha and observes each category i n_i times (assuming categories are drawn from a multinomial distribution pi) the posterior is alpha_i -> alpha_i + n_i

Constructors
Constructor and Description
`Dirichlet(double... alpha)`

All Methods Static Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`double[]`	`effectiveLog10MultinomialWeights()`
`double[]`	`effectiveMultinomialWeights()`
`double[]`	`log10MeanWeights()`
`double[]`	`meanWeights()`
`int`	`size()`
`static Dirichlet`	`symmetricDirichlet(int numStates, double concentration)` Create a symmetric distribution Dir(a/K, a/K, a/K .

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

public double[] effectiveMultinomialWeights()

public double[] effectiveLog10MultinomialWeights()

Class Dirichlet