Probabilities that contaminated normal observations are extreme
作者:
L. I. Pettit,
期刊:
Journal of Applied Statistics
(Taylor Available online 1992)
卷期:
Volume 19,
issue 1
页码: 111-116
ISSN:0266-4763
年代: 1992
DOI:10.1080/02664769200000009
出版商: Carfax Publishing Company
数据来源: Taylor
摘要:
Results are presented for the probability that a contaminated observation in a normal sample isan outlier. Univariate samples with mean-shift or variance inflation contamination were considered. Multivariate samples with inflation of the covariance matrix in which an outlier is the observation with minimum conditional predictive ordinate were also studied. All the results were obtained by numerical integration or simulation.
点击下载:
PDF (327KB)
返 回