Exercise 24.6 - Gibbs sampling for probit regression

Answers

With:

p (𝐘, 𝐙, 𝐰 | 𝐗) = p (𝐰) \cdot p (𝐙 | 𝐗, 𝐰) \cdot p (𝐘 | 𝐙),

we can derive all conditional posterior as in exercise 24.5, where the logarithm of $p (𝐘 | 𝐙)$ truncates the Gaussian for $𝐙$ .

solour_lfq

2021-03-24 13:42