1
$\begingroup$

Let's say I have a uniform prior $\mu \sim \mathcal{U}(a,b)$, a normal likelihood $y|\mu \sim \mathcal{N}(\mu,\sigma^2)$ with known variance $\sigma^2$, and one observation $y$. Is then the posterior $\mu|y$ a truncated normal distribution with parameters with parameter $\mu=y$, $\sigma=\sigma$, $a=a $, $b=b$?

$\endgroup$
4
$\begingroup$

Yes: $$ p(\mu \mid y) \propto p(y \mid \mu) \, p(\mu) \propto \mathcal N(y; \mu, \sigma^2) \, \mathbb{1}(\mu \in [a, b]) ,$$ where $\mathbb 1$ is the 0-1 indicator function. This is exactly (up to a multiplicative constant) the pdf of the distribution you said.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.