The multinomial distribution generalizes the binomial distribution to more than two outcome categories: it describes the joint distribution of counts obtained when n independent objects are each classified into one of k disjoint categories according to a probability vector, with a joint PMF derived by a permutation-counting argument analogous to the binomial case. Two of its structural properties — the "lumping property" (merging categories yields a lower-dimensional multinomial) and the marginal/conditional distributions (each component is marginally binomial, and conditioning on one component renormalizes the remaining probability vector) — follow directly from the underlying independent-classification story rather than from algebraic manipulation. Separately, the Cauchy distribution, defined as the ratio of two independent standard normal random variables, is a canonical example in probability theory of a distribution lacking a defined mean or variance, whose density is derived via differentiation of a joint-CDF double integral (or equivalently via the law of total probability) — illustrating techniques for working with joint continuous distributions within the broader domain of probability theory and joint/multivariate distributions.
The multinomial distribution generalizes the binomial distribution to more than two outcome categories: it describes the joint distribution of counts obtained when n independent objects are each clas…