反向常態累加分布函數之數值方法的性質之分析

Abstract

The inverse normal cumulative distribution function does not have a close form. Several algorithms for evaluating (or approximating) the inverse normal cumulative distribution function have been proposed. The aim of the thesis is to compare a few numerical methods of the inverse normal cumulative distribution function, which are used for calculating approximate values of the inverse normal cumulative distribution function. They include the built-in function in EXCEL, a numerical method by Peter J. Acklam, and a numerical method by Moro. The numerical method of Moro is implemented with MATLAB in this thesis. Related errors of the above-mentioned numerical methods are analyzed in the current thesis as well.