Abstract: An algorithm for the computation of the regularized incomplete Beta function is described. This function has important applications in areas such as Statistics, Physics and Information Theory. The computation of the function can be carried out through a continued function evaluation supplemented with series and asymptotic expansions when both parameters are large. Numerical tests demonstrate the accuracy of the algorithm and show that our algorithm is more accurate than Matlab?s built-in function betainc for a wide range of parameters.

Authorship

Download reference