Unified Conditional Frequentist and Bayesian Testing of Composite Hypotheses (Q4455945)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 2059381
Language Label Description Also known as
default for all languages
No label defined
    English
    Unified Conditional Frequentist and Bayesian Testing of Composite Hypotheses
    scientific article; zbMATH DE number 2059381

      Statements

      Unified Conditional Frequentist and Bayesian Testing of Composite Hypotheses (English)
      0 references
      0 references
      0 references
      16 March 2004
      0 references
      Bayes factor
      0 references
      conditional error probability
      0 references
      default prior distribution
      0 references
      group invariance
      0 references
      nested hypotheses
      0 references
      posterior probability
      0 references
      The authors consider testing \(H_0:X\sim f_0(\cdot\,|\, \vartheta_0)\), \(\vartheta_0\in\Theta_0\), against \(H_1:X\sim f_1(\cdot\, |\;\vartheta_1)\), \(\vartheta_1\in\Theta_1\), where \(f_i\) are PDF families invariant w.r.t. some group \(G\) acting on the observation space. As a test statistic the Bayes factor \[ B(x)=\int f_0(x\mid\vartheta) \,d\mu(\vartheta)/ \int f_1(x\mid \vartheta) \,d\mu(\vartheta) \] is proposed, where \(\mu\) is a prior invariant w.r.t. groups induced on \(\Theta_i\) by \(G\). It is shown that the obtained test can be interpreted in either a Bayesian or conditional frequentist sense. Examples of testing Weibull vs lognormal, exponential vs lognormal and correlated vs uncorrelated normal by i.i.d. samples are considered.
      0 references

      Identifiers