Bayesian Testing of a Point Null Hypothesis Based on the Latent Information Prior

Entropy

关键词: conditional mutual information;    discrete prior;    Kullback-Leibler divergence;    prediction;    reference prior;    Jeffreys-Lindley paradox;   

DOI  :  10.3390/e15104416

来源: mdpi

【 摘 要 】

Bayesian testing of a point null hypothesis is considered. The null hypothesis is that an observation, x, is distributed according to the normal distribution with a mean of zero and known variance . The testing problem is formulated as a prediction problem. Bayesian testing based on priors constructed by using conditional mutual information is investigated.

【 授权许可】

CC BY   
© 2013 by the authors; licensee MDPI, Basel, Switzerland.