TRANSFORMATION OF NONPOSITIVE SEMIDEFINITE CORRELATION-MATRICES
TRANSFORMATION OF NONPOSITIVE SEMIDEFINITE CORRELATION-MATRICES
复制标题
DOI：
10.1080/03610928308831068
复制
发表时间：
1993-01-01
影响因子：
0.8
通讯作者：
MOLENBERGHS, G
中科院分区：
文献类型：
--
作者：
ROUSSEEUW, PJ;MOLENBERGHS, G
在多元统计中，协方差或相关矩阵的估计是至关重要的。计算和其他论证经常导致使用依赖于坐标的估计器，所产生的矩阵是对称的但不是半正定的。我们简要地讨论了基于收缩的将这类矩阵变换为半正定矩阵的现有方法。文中还考虑了一种基于特征值的简单方法。考虑到相关矩阵的几何结构，提出了一种新的方法，该方法采用了类似于多维尺度的技术。
In multivariate statistics, estimation of the covariance or correlation matrix is of crucial importance. Computational and other arguments often lead to the use of coordinate-dependent estimators, yielding matrices that axe symmetric but not positive semidefinite. We briefly discuss existing methods, based on shrinking, for transforming such matrices into positive semidefinite matrices. A simple method based on eigenvalues is also considered. Taking into account the geometric structure of correlation matrices, a new method is proposed which uses techniques similar to those of multidimensional scaling.