This article is cited in 7 papers

Short Communications

Unbiased estimators and classification problems for multivariate normal populations

R. A. Abusev, Ya. P. Lumel'skiĭ

Perm

Abstract: We obtain the unbiased estimates of multivariate normal probability density and the unbiased estimates of probability density of the sample mean and the sample covariance matrix when independent observations have a normal distribution with unknown mean and variance. This unbiased estimators are used for the construction of the point and group classification rules. It is shown that the error of group classification tends to zero when all sample sizes tend to infinity.

Received: 31.01.1978

 English version: Theory of Probability and its Applications, 1981, 25:2, 377–384

Bibliographic databases:

• 
• 
•