11
$\begingroup$

I am wondering if prediction interval and credible interval evaluate the same thing.

For instance with a linear regression, when you estimate the prediction interval of a fitted values, you estimate the $(1-\alpha)\%$ limits of the interval in which you expect your value to fall. Conversely to a confidence interval, you do not focus on a distribution parameter such as the mean value, but on the value that your explained variable could take for a given X value (supposing that $\ Y = a + b.X$).

When you estimate the fitted value for a given $X$ value within a Bayesian framework, from the posterior probability distribution, you can estimate a credible interval. Does this interval give you the same information on the fitted value or not?

$\endgroup$
18
$\begingroup$

They live in different spaces and mean different things.

A credible interval $[a,b]$ is a subset of the parameter space such that $$ P(a\leq\Theta\leq b\mid X_1=x_1,\dots,X_n=x_n) = \alpha \, , $$ and it means that, after seeing the data, you believe that with probability $\alpha$ the parameter value is inside this interval.

A prediction interval $[u,v]$ is a subset of the sampling space such that $$ P(u\leq X_{n+1}\leq v\mid X_1=x_1,\dots,X_n=x_n) = \gamma \, , $$ and it means that, after seeing the data, you believe that with probability $\gamma$ the value of a future observation $X_{n+1}$ will be inside this interval.

$\endgroup$

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.