The distribution associated with the random variable, X, defined as the number of ‘successes’ in n independent trials each having the same probability, p, of success. The random variable X is said to be a binomial variable and to have a binomial distribution with parameters n and p. This is written as X ~ B(n, p). The mean of this distribution is np and the variance is np (1−p). The probability function is given byThe distribution takes its name from the fact that successive probabilities are the terms in the expansion in ascending powers of p, by the binomial theorem, of (q+p)n, where q=1−p. The first published derivation of the distribution was by Jacob Bernoulli in 1713.
As an example, suppose that a computer generates fifteen random integers between 0 and 9 inclusive. The number of these integers that are odd has a B(15, 0.5) distribution. The number that are non-zero has a B(15, 0.9) distribution, and the number that are greater than 7 has a B(15, 0.2) distribution. The diagram shows the graphs of the probability functions for these distributions.
If we note that P(X=0)=qn, successive probabilities can be calculated using the recurrence relationIf (n+1)p is not an integer the graph is unimodal, with mode at the (integer) value of r such that and (n+1)p−1 and (n+1)p are both modal values (as in the B(15, 0.5) case illustrated).
A binomial random variable with parameters n and p may be regarded as the sum of n independent observations of a Bernoulli variable (see Bernoulli distribution) with parameter p. The sum of two independent binomial variables with parameters n1, p and n2, p, respectively, is also a binomial variable, with parameters (n1+n2), p.
For large values of np and nq the normal approximation to the binomial distribution may be used:and Φ is the cumulative distribution function for a standard normal variable (see normal distribution). The ‘½’ is a continuity correction. The result, that a binomial distribution with p = ½ may be approximated by a normal distribution, underlies the derivation of the normal distribution by de Moivre in 1733 and is sometimes referred to as the de Moivre–Laplace theorem. For large values of n and small values of p the Poisson approximation to the binomial distribution may be used:The word ‘binomial’ was used in its mathematical sense in a 1557 text entitled The Whetstone of Witte by Robert Recorde. The ‘binomial distribution’ was so named by Yule in 1911.
http://www.math.uah.edu/stat/applets/BinomialCoinExperiment.html Applet.