Efficiency of logistic-normal stochastic supervision

作者:

Highlights:

摘要

We consider the case of a stochastic supervision scheme in which the logarithm of odds ratio of the supervisor classification has a normal distribution and is independent of the feature vector. We consider the two-group multinormal case with a common covariance matrix and compute the Efron efficiency of this supervision scheme, for various values of parameters of the feature distribution and the supervision distribution. We interpret these efficiency values to draw conclusions on the worth of stochastic supervision.

论文关键词:Discriminant analysis,Stochastically classified initial samples,Asymptotic relative efficiency,Logistic-normal distribution

论文评审过程:Received 7 December 1989, Available online 19 May 2003.

论文官网地址:https://doi.org/10.1016/0031-3203(90)90121-Z