超几何分布高精度模型研究

Research on High-Precision Model of Hyper-Geometric Distribution

摘要: 本文为了推导超几何分布连乘模型，根据它的分式性质，对随机变量进行了两种分类。一种为0的情况，另一种为1, 2, ∙∙∙, n情况。通过对两组元素的组合式超几何分支模型，分别进行了数学变换，推导出计算机不会溢出的超几何分布连乘模型。此模型准确地计算超几何分布全元素0, 1, ∙∙∙, n的概率值，并N大时计算机不会溢出。

Abstract: In order to derive the hyper-geometric distribution and multiplicative model, this paper classifies random variables according to their fractional properties. One is 0, and the other is 1, 2, ∙∙∙, n. Through the combined hyper-geometric branching model of two groups of elements, mathematical transformations were respectively performed, and a hyper-geometric distribution continuous multiplication model with no computer overflow was deduced. This model accurately calculates the probability value of all elements 0, 1, ∙∙∙, n of the hyper-geometric distribution, and the computer will not overflow when N is large.