Holonomic gradient method for multivariate distribution theory

Holonomic gradient method for multivariate distribution theory
复制标题

多元分布理论的完整梯度法

DOI:
10.1007/978-3-030-75494-5_1
复制
发表时间:
2021
期刊:
Multivariate, Multilinear and Mixed Linear Models(Contributions to Statistics)
影响因子:
--
通讯作者:
A. Takemura
A. Takemura
中科院分区:
--
文献类型:
--
作者:
山田 道洋 ;菊池 浩明 ;松山 直樹 ;乾 孝治;A. Takemura

文献摘要

相似文献

2011年,我们引入了完整梯度法(HGM)作为多元分布理论的一种新方法。当密度函数在随机变量和参数中都是“完整的”时,它是一种通用的方法。由于多元正态密度是完整的,因此HGM适用于基于多元正态分布的整个经典分布理论。事实上,自2011年以来,我们将HGM应用于经典多元分布理论的许多问题,并证明了一些困难的分布问题,例如涉及矩阵参数的超几何函数的分布问题,可以成功地用HGM处理。本文讨论了HGM的起源、基本理论及其在多元分布理论中的一些应用。
In 2011, we introduced the Holonomic Gradient Method (HGM) as a new approach to multivariate distribution theory. It is a general method applicable when the density functionis “holonomic” in the random variablesxas well as the parameters. Since the multivariate normal density is holonomic, HGM is applicable to the whole classical distribution theory based on the multivariate normal distribution. In fact since 2011, we applied HGM to many problems of classical multivariate distribution theory and showed that some difficult distributional problems, such as those involving hypergeometric functions of a matrix argument, can be successfully treated by HGM. In this article, we discuss the origin, the basic theory and some applications of HGM to multivariate distribution theory.