The joint distribution function of real random variables isIt determines probabilities of rectangles throughIts marginals are and similarly for . independence is equivalent to for every .
If the joint law has a Probability density function , then , and mixed differentiation recovers where justified. A joint density need not exist even when both marginals have densities. For example, take with uniform on . Writing for the uniform Cumulative distribution function gives . The law is concentrated on the diagonal , a set of two-dimensional Lebesgue measure zero. Its two uniform marginals therefore do not make it an absolutely continuous joint law.
A copula is a joint distribution function on a unit cube whose one-dimensional marginal distributions are uniform distributions. Applying each continuous Marginal distribution to its Random variable produces uniform marginals by the probability integral transform; their copula records the remaining dependence. independent uniform coordinates and one common uniform coordinate give different copulas with the same marginals.
Articles by others on the same topic
There are currently no matching articles.