Absolute muppetry …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
So you might have seen me picking fights on twitter with former colleagues at Oxford and probably think that I do this deliberately because I just like to fight.  Actually, I would prefer not to have arguments with people but in this case the analysis is so poor and gives such a deliberately skewed impression of the uncertainties involved in this type of modelling—and was peppered with buzzwords like ‘herd immunity’ to make it tasty for the popular press to whom it was fed directly—that I couldn’t help but lose my shit.  If you want to see the dumpster fire for yourself, it’s over here on dro ..read more
Visit website
Design considerations for a COVID-19 serological survey … part i
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
Since I am, by profession, a statistician employed the field of epidemiology—no longer an astronomer—it makes sense that I should now be devoting my spare research time on projects related to COVID-19 instead of blogging (mostly) about astronomy.  (Spare time since I will be putting aside any of my malaria modelling commitments since this is still very much—in both a relative and an absolute sense—an important disease to control in sub-Saharan Africa & other endemic areas). One topic that I’ve worked on in the past that seems to have some relevance these days is the challenge of map making ..read more
Visit website
“Some of those that work forces, are the same that burn crosses”
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
[Not astro or stats] Boosting the signal on a story that the cowards in the Australian government and top defence brass would rather see buried.  A story that they’d rather lock up journalists and whistleblowers for telling than admit to a failure of military discipline and a failure of justice.  There is strong evidence that Australian SAS members were murdering civilians in Afghanistan, and that their actions were either condoned by senior defence force leaders or else they simply turned a blind eye to the situation.  Credit to those soldiers who have come forward to tell their stories and b ..read more
Visit website
Over-confidence in Bayesian analyses of galaxy rotation curves …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
Yesterday I had a ‘Matters Arising’ in Nature Astronomy published (non-paywalled shareable link here) on the topic given as the title of this blog post; with co-authors, Michael Burgess & Garry Angus.  Aside from identifying some particular technical issues with a published study on the topic of MOND-inspired radial acceleration relationships in galaxies, we highlight the potential of model misspecification to lead to overly tight parameter constraints or overly confident assessments of model choice in Bayesian analyses—and we remind readers of the consequent value in a model testing and d ..read more
Visit website
[Link] A toast to the error detectors …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
This ..read more
Visit website
Non non Gaussian processes …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
I noticed an interesting paper arXived just before Xmas proposing to model stellar spectra as “sparse, data-driven non-Gaussian processes”.  The non-Gaussian part is explained to mean that because a prior is placed on the covariance function and various hyper-parameters, when these are marginalised over the resulting posterior distributions for the latent spectral profiles are highly non-Gaussian.  I don’t personally care for the use of the term “non-Gaussian process” to describe this situation, since it can confuse thinking around how the hierarchical GP model is behaving in the sense of the ..read more
Visit website
Boxing to watch on boxing day …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
I realise that the origin of the holiday name is not the sport, but it’s a great chance to catch up on two fight of the year contenders, both coming out of the World Boxing Super Series program (both available for free on youtube): 1) Taylor vs Prograis 2) Inoue vs Donaire ..read more
Visit website
RDT to Microscopy conversion is neither trivial nor unimportant …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
A common criticism of malaria mapping work that I’ve been involved with runs to the effect that we don’t do the right thing when we correct for the difference between rapid diagnostic test (RDT) estimates of parasite prevalence and microscopic diagnosis estimates of parasite prevalence.  For instance, this paper points out that in Mappin et al (2015) (where we examine the relationship between RDT and microscopy prevalence) the fitted model has “a major limitation” being that we don’t allow for over-dispersion in any unmeasured risk factors, and proposes that the solution is a model in which ea ..read more
Visit website
Astronomers invented Bayesian optimisation in 2019 …
Another Astrostatistics Blog | The random musings of a reformed astronomer …
by drewancameron
4y ago
Long time readers of this blog will know that I’ve been bigging up the potential of Bayesian optimisation approaches for fitting expensive astronomical simulations for many years now.  In particular, when likelihood function computations are computationally very expensive and possibly stochastic due to the need to run some kind of crazy physics simulator at the back end, then it makes sense to approximate the posterior in an iterative manner using Bayesian optimisation.  And if you’re going to do that you should learn from the wealth of experience in the statistics community (and elsewhere in ..read more
Visit website

Follow Another Astrostatistics Blog | The random musings of a reformed astronomer … on FeedSpot

Continue with Google
Continue with Apple
OR