PRAXIS - BRENT ALGORITHM FOR FUNCTION MINIMIZATION

被引：55

作者：

GEGENFURTNER, KR ^{[1
]}

机构：

[1] NYU,CTR NEURAL SCI,NEW YORK,NY 10003

来源：

BEHAVIOR RESEARCH METHODS INSTRUMENTS & COMPUTERS | 1992年 / 24卷 / 04期

关键词：

D O I：

10.3758/BF03203605

中图分类号：

B841 [心理学研究方法];

学科分类号：

040201 ;

摘要：

Implementations of Brent's (1973) PRincipal AXIS (PRAXIS) algorithm in the widely used C and PASCAL programming languages are presented The algorithm minimizes a multivariate function without using derivatives. An example computer program that calculates a maximum likelihood estimate of the parameters of a psychometric function illustrates the use of the routine. Another algorithm, Localmin, also due to Brent (1973), efficiently finds the minimum of a univariate function. An example program uses this algorithm to estimate the polychoric correlation coefficient from a p x q table of observed frequencies.

引用

页码：560 / 564

页数：5

共 9 条

[1]

Bock R. D., 1968, MEASUREMENT PREDICTI

[2]

Brent R.P., 1973, ALGORITHMS FUNCTION

[3] QUASI-NEWTON METHODS AND THEIR APPLICATION TO FUNCTION MINIMISATION [J].