Hyperprior

Hyperprior is a type of prior – a kind of guess – that is used in Bayesian statistics. Priors are used to give us an idea of what the answer to a certain question might be, before we have enough data or information to be sure. Hyperpriors help us make better guesses by going one step further. They are guesses about the guesses (our priors) that we make. In Bayesian statistics we use hyperpriors to explain the relationship between the data we have and our guesses, while also helping us to learn more from the data we have.