A dyadic distribution is a specific type of discrete or categorical probability distribution that is of some theoretical importance in data compression.
A dyadic distribution on the nonnegative integers 0, 1, 2, ... is a probability distribution whose probability mass function is
where nu is some (positive) integer. More generally it is a categorical distribution in which the probability assigned to any label is of the above form
It is possible to find a code defined on this distribution, which has an average code length that is equal to the entropy.