作者: Sadanori Konishi , Takakazu Sugiyama
DOI: 10.1007/BF02480916
关键词: Mathematics 、 Sample (statistics) 、 Applied mathematics 、 Combinatorics 、 Population 、 Sample mean and sample covariance 、 Wishart distribution 、 Confidence interval 、 Estimation of covariance matrices 、 Bivariate analysis 、 Simple (abstract algebra)
摘要: Normalizing transformations of the largest and smallest latent roots a sample covariance matrix in normal are obtained, when corresponding population simple. Using our results, confidence intervals for may easily be constructed. Some numerical comparisons resulting approximations made bivariate case, based on exact values probability integral roots.