Suppr超能文献

一个在MS-DOS下运行的用于分析流行病学分层或匹配数据的程序。

A program running under MS-DOS for the analysis of epidemiologic stratified or matched data.

作者信息

Thalabard J C, Plu-Bureau G, Falissard B

机构信息

INSERM U. 292, Hôpital de Bicêtre, Le Kremlin-Bicêtre, France.

出版信息

Comput Methods Programs Biomed. 1989 Mar;28(3):191-3. doi: 10.1016/0169-2607(89)90148-x.

Abstract

Breslow and Day's method for computing the conditional likelihood of the logistic model in case of stratified samples requires, at each iteration the generation, for each stratum, of all the permutations of nc elements drawn from a total of nc + nt elements, where nc and nt are respectively the numbers of the cases and of the controls. This procedure rapidly proves costly in time as nc and nt increase, and also when nc and nt differ from one stratum to the next. The use of this technique on microcomputers is therefore limited. Presented here is an algorithm written in FORTRAN, based on a binary representation of the possible permutations, which readily permits the implementation of Breslow and Day's model on a microcomputer.

摘要

布雷斯洛(Breslow)和戴(Day)在分层抽样情况下计算逻辑模型条件似然性的方法,在每次迭代时,对于每个分层,都需要生成从总共nc + nt个元素中抽取的nc个元素的所有排列,其中nc和nt分别是病例数和对照数。随着nc和nt的增加,以及当nc和nt在不同分层之间不同时,此过程很快就会在时间上变得代价高昂。因此,在微型计算机上使用这种技术受到限制。这里给出一个用FORTRAN编写的算法,它基于可能排列的二进制表示,这使得在微型计算机上轻松实现布雷斯洛和戴的模型成为可能。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验