摘要
对样本相关系数矩阵等行和分解算法作了简化和推广,使算法不仅可以应用在基于正态总体非独立样本的假设检验问题,也可以有效地运用在最优化算法中牛顿法等与二次函数极小化有关的问题上.
The algorithm of equal line sum decomposition for the sample correlation coeffi- cient matrix has been simplified and extended. Therefore, the algorithm can be applied not only in hypothesis testing based on dependent sample on normal totality , but efficiently in the problems related to minimization of quadratic function such as the Newton Method in optimization as well.
出处
《数学的实践与认识》
CSCD
北大核心
2011年第21期239-243,共5页
Mathematics in Practice and Theory
关键词
正定矩阵
分解算法
推广
二次函数
最优化
positive definite matrix
decomposition algorithm
extension
quadratic function
optimization
