A probability distribution used to smooth categorical data by adding prior counts to observed frequencies.