A Bayesian framework w/o priors or posteriors?

I noticed a paper by Jensen-Clem et al. on the arXiv the other day which I opened into a new tab since its abstract proclaims “In a Bayesian framework, we investigate the correlation between wavelength calibration errors and resulting radial velocity errors.” But then I tried to find the Bayesian section by text-searching “Bayes” (no luck), “prior” (no luck) “posterior” (no luck), and “Bayesian” (no luck) before assuming my text search must be broken (which it wasn’t) and manually trawling through.  Although the study is clearly quite sophisticated in its investigation (through simulations) of the potential of this lock-in amplified spectrometer to achieve precise radial velocity measurement there is certainly no Bayesian statistics going on here.  The only inference procedure I could find (Section 3) concerned a simulation-based study of the distribution of the maximum (pseudo-)likelihood estimator over replicate mock datasets; and I say pseudo-likelihood since the “chi squared test” adopted (in fact, not a chi squared test) is an arbitrary observed-data-to-model discrepancy distance rather than the representation of an assumed sampling distribution.

Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s