1
$\begingroup$

I am trying to check if the noncentral chi-square distribution is log-concave in its noncentrality parameter. Specifically, given

$p(y ; \lambda, \sigma^2) = \frac{1}{2\sigma^2}\exp\left(-\frac{y+\lambda}{2\sigma^2}\right) I_0\left(\frac{1}{\sigma^2}\sqrt{\lambda y_i}\right)$

is $\lambda \rightarrow \log p(y;\lambda,\sigma^2)$ concave? In addition, is $\lambda^2 \rightarrow \log p(y;\lambda,\sigma^2)$ concave?

I think the first is affirmative, but could not find the second in the literature.

$\endgroup$
1
$\begingroup$

See

http://www3.stat.sinica.edu.tw/statistica/oldpdf/A7n313.pdf

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.