binofit

Binomial parameter estimates

Syntax

phat = binofit(x,n)
[phat,pci] = binofit(x,n)
[phat,pci] = binofit(x,n,alpha)

Description

phat = binofit(x,n) returns a maximum likelihood estimate of the probability of success in a given binomial trial based on the number of successes, x, observed in n independent trials. If x = (x(1), x(2), ... x(k)) is a vector, binofit returns a vector of the same size as x whose ith entry is the parameter estimate for x(i). All k estimates are independent of each other. If n = (n(1), n(2), ..., n(k)) is a vector of the same size as x, the binomial fit, binofit, returns a vector whose ith entry is the parameter estimate based on the number of successes x(i) in n(i) independent trials. A scalar value for x or n is expanded to the same size as the other input.

[phat,pci] = binofit(x,n) returns the probability estimate, phat, and the 95% confidence intervals, pci. binofit uses the Clopper-Pearson method to calculate confidence intervals.

[phat,pci] = binofit(x,n,alpha) returns the 100(1 - alpha)% confidence intervals. For example, alpha = 0.01 yields 99% confidence intervals.

Note

binofit behaves differently than other Statistics and Machine Learning Toolbox™ functions that compute parameter estimates, in that it returns independent estimates for each entry of x. By comparison, expfit returns a single parameter estimate based on all the entries of x.

Unlike most other distribution fitting functions, the binofit function treats its input x vector as a collection of measurements from separate samples. If you want to treat x as a single sample and compute a single parameter estimate for it, you can use binofit(sum(x),sum(n)) when n is a vector, and binofit(sum(X),N*length(X)) when n is a scalar.

Examples

This example generates a binomial sample of 100 elements, where the probability of success in a given trial is 0.6, and then estimates this probability from the outcomes in the sample.

r = binornd(100,0.6);
[phat,pci] = binofit(r,100)
phat =
0.5800
pci =
0.4771  0.6780

The 95% confidence interval, pci, contains the true value, 0.6.

References

 Johnson, N. L., S. Kotz, and A. W. Kemp. Univariate Discrete Distributions. Hoboken, NJ: Wiley-Interscience, 1993.