I just would like some reassurance on my interpretation of a few concepts in this paper Xiao, R. & Scott, L. (2011).

The question is about the likelihoods. Am I reading correctly the following expressions of $L$ when evaluated at $$\hat{\theta_0}=(\hat{\mu_0},\hat{\sigma}^2)$$ and $$\hat{\theta_1}=(\tilde{\pi},\tilde{\mu_0},\tilde{\alpha_R},\tilde{\sigma}^2).$$

In other words, are $L(\hat{\theta_0})$ and $L(\hat{\theta_1})$ given by the following expressions?

Likelihood under the null hypothesis: $$L(\hat{\theta_0}) = L(\hat{\theta_0};y_{i\in\{G_i=RR,rr\}}) = \Pi_{i\in\{G_i=RR,rr\}} f(y_i; \hat{\mu_0}, \hat{\sigma}^2)$$

Likelihood under the alternative hypothesis:

$$ L(\hat{\theta_1}) =L(\hat{\theta_1};y_{i\in\{G_i=RR,rr, Rr\}}) = \Pi_{i\in\{G_i=RR,rr\}} f(y_i; \tilde{\mu_0}, \tilde{\sigma}^2) \times \\ \Pi_{i\in\{G_i=Rr\}} \{ \tilde{\pi}\ f(y_i;\tilde{\mu_0}+\tilde{\alpha_R}, \tilde{\sigma}^2) + (1-\tilde{\pi})f(y_i;\tilde{\mu_0}-\tilde{\alpha_R}, \tilde{\sigma}^2)\} $$

enter image description here

  • 1
    $\begingroup$ Both likelihoods must use the same data, I would thus think $L(\hat{\theta}_0)$ is evaluated as a Normal likelihood for all $y_i$'s. $\endgroup$
    – Xi'an
    Mar 28, 2018 at 5:36


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.