Jethro's Braindump

Exchangeability

A set of parameters \(\theta = (\theta_{1}, \theta_{2}, \dots, \theta_{k})\) is exchangeable if the distribution of \(\theta\) is unchanged when the parameter components are permuted. One’s prior belief about \(\theta_{j}\) will be the same as one’s prior beliefs of \(\theta_{k}\).

Links to this note