Homepage › Solution manuals › Kevin P. Murphy › Machine Learning: a Probabilistic Perspective › Exercise 3.14 - Posterior predictive for Dirichlet-multinomial
Exercise 3.14 - Posterior predictive for Dirichlet-multinomial
Answers
For question (a). In this concrete case we have , , any component of be , and . To derive , we resort to the deduction in Exercise 3.13:
For question (b), the independence between characters is still ignored, bringing no significant change to the computation: