English

应用科学学报 ›› 1994, Vol. 12 ›› Issue (4): 333-339.

• 论文 • 上一篇    下一篇

离散随机观测的最佳逼近

刘永才1, 林刚2   

  1. 1. 上海大学嘉定校区;
    2. 卢湾区环境监测站
  • 收稿日期:1992-09-19 出版日期:1994-12-31 发布日期:1994-12-31
  • 基金资助:
    自然科学基金和高校科技发展基金

THE BEST APPROXIMATION OF DISCRETE RANDOM OBSERVATIONS

LIU YONGCAI1, LIN GANG2   

  1. 1. Shanghai University, Jiading Campus;
    2. Environment Monitoring Station of Lu Wan District
  • Received:1992-09-19 Online:1994-12-31 Published:1994-12-31

PDF

66

摘要： 在大量科学实验,工程测试,专家系统中蕴含一类问题:如何从有限次随机观测中识别出一个布尔函数,它能最佳表示出此观测集中的关系。该文给出一些求解此类问题的途径。

关键词: 公式回归, 最大似然估计, 函数回归, 单调函数

Abstract: A number of scientific experiments, engineering tests and expert systems involve a class of problems:how can the Boolean functions be identified from finite random observations,so that the Boolean functions can best represent the relationships in a set of observations? Some approaches to this problems are given in this paper.

Key words: functional regression, monotone function, formulaic regression, maximum likehood estimation