Assessing the reliability of statistical software: Part II

被引：66

作者：

McCullough, BD ^{[1
]}

机构：

[1] Fed Commun Commiss, Washington, DC 20554 USA

来源：

AMERICAN STATISTICIAN | 1999年 / 53卷 / 02期

关键词：

accuracy; benchmarks; random number generation; software testing; StRD;

D O I：

10.2307/2685736

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

Part I outlined a methodology for assessing the reliability of three areas: estimation, random number generation, and calculation of statistical distributions. The present article applies this methodology to SAS, SPSS, and S-Plus, with attention to implementation details. Weaknesses are identified in all the random number generators, the S-Plus correlation procedure, and in the one-way ANOVA and nonlinear least squares routines of SAS and SPSS.

引用

页码：149 / 159

页数：11

共 29 条

[1]

[Anonymous], 1980, Statistical Computing

[2]

Bard Y., 1974, Nonlinear Parameter Estimation

[3]

BROWN BW, 1998, DCDFLIB V 1 1

[4] A USERS GUIDE TO NONLINEAR OPTIMIZATION ALGORITHMS [J].