Implement logprob derivation for Max and Min #6773
Description
Description
Description
The idea would be to derive the logprob for the following type of graphs
import pytensor.tensor as pt
from pymc import logp
x_rv = pt.max(at.random.uniform(0, 1, size=3))
y_rv = pt.min(at.random.uniform(0, 1, size=3))
x_vv = x_rv.clone()
y_vv = y_rv.clone()
x_max_logprob = logp(x_rv, x_vv)
y_max_logprob = logp(y_rv, y_vv)Reference: https://en.wikipedia.org/wiki/Order_statistic#Probabilistic_analysis