New on the blog: Using Bayesian tools to be a better frequentist
Turns out that for negative binomial regression with small samples, standard frequentist tools fail to achieve their stated goals. Bayesian computation ends up providing better frequentist guarantees. Not sure this is a general phenomenon, just a specific example.
https://www.martinmodrak.cz/2025/07/09/using-bayesian-tools-to-be-a-better-frequentist/
#statstab #382 The JASP Guidelines for Conducting and Reporting a Bayesian Analysis
Thoughts: @JASPStats is often people's first attempt at Bayesian statistics. But proper inference and reporting is crucial.
#JASP #Bayesian #BayesFactor #guide #tutorial
https://link.springer.com/article/10.3758/s13423-020-01798-5
Happy birthday Edwin T. Jaynes, champion of sound physics & probability theory! 🎂 🚀 🎓
I find his text still the most illuminating about probability theory: <https://doi.org/10.1017/CBO9780511790423>
And also still a classic in statistical mechanics: <https://doi.org/10.1103/RevModPhys.34.143>
Dear colleagues working with Markov-chain Monte Carlo, especially for Bayesian posterior analysis:
Do you know of any Monte Carlo Standard Error (MCSE) estimators (or alternatively a p% MC error interval) for the *mean*, in the case where the posterior distribution may *not* have a finite variance (but it's known to have a finite mean)?
The paper by Vehtari & al. <https://doi.org/10.1214/20-BA1221> offers "efficiency" estimates for the mean in the case of non-finite variance, as well as a method to find MC error intervals for quantiles. But I can't find there a method for MC standard error or error intervals for the *mean* in the case of non-finite variance.
Cheers!
#bayesian #bayes #mcmc @avehtari@bayes.club @avehtari@mastodon.social
Dear colleagues working with Markov-chain Monte Carlo: could you share any works that explore Markov-chain "convergence", and precision of mean estimates, with methods that use *quantiles* (or interquartile range, or median absolute deviation, or similar), rather than standard deviation and similar quantities?
Just to be clear, I don't mean estimation of quantiles, but estimation *by means of* quantiles.
Thank you!
#statistics #stats #probability #markovchain #mcmc #bayesian #bayes
#statstab #377 To adjust, or not to adjust, for multiple comparisons
Thoughts: Not all-encompassing, but it does cover some relevant notions about multiplicity adjustments.
#FDR #FWER #typeI #error #bonferroni #bayesian #multiplecomparisons
https://www.jclinepi.com/article/S0895-4356%2825%2900021-6/fulltext
A monument to the triumph of vanity over safety?
Gene expression is tuned by #epigenetic changes, which explains why, say, liver and skin cells have the same genome but are otherwise different. Here we introduce a
#Bayesian model for the analysis of epigenetic changes during development. https://epigeneticsandchromatin.biomedcentral.com/articles/10.1186/s13072-025-00594-6
Die Luxusjacht "Bayesian" ist aus dem Meer geborgen! 🚢🌊 Nach langer Zeit wurde das spektakuläre Wrack erfolgreich geborgen. Mehr dazu im Artikel: https://www.n-tv.de/panorama/Luxusjacht-Bayesian-endlich-aus-dem-Meer-geborgen-article25849786.html #Luxusjacht #Bayesian #Bergung #News
#newz
Sunken British superyacht Bayesian is raised from the seabed.
A superyacht that sank off the coast of the Italian island of Sicily last year has been raised from the seabed by a specialist salvage team.
Seven of the 22 people on board died in the sinking, including the vessel's owner, British tech tycoon Mike Lynch and his 18-year-old daughter.
The cause of the sinking is still under investigation.
Vor Sizilien: Luxusjacht "Bayesian" wird geborgen
Die Bergung des 56 Meter langen Segelschiffes hatte sich mehrfach verzögert. Nun ist es an der Oberfläche. Bei ihrem Untergang starben der britische Milliardär Mike Lynch und sechs weitere Insassen.
Das Wrack wurde jetzt vor Sizilien aus 50 Metern Tiefe an die Wasseroberfläche gehoben.
#Bayesian #MikeLynch
https://floatmagazin.de/leute/luxusjacht-bayesian-von-meeresgrund-gehoben/
Interested in trying out *Bayesian nonparametrics* for your statistical research?
I'd be very grateful if people tried out this R package for Bayesian nonparametric population inference, called "inferno" :
<https://pglpm.github.io/inferno/>
It is especially addressed to clinical and medical researchers, and allows for thorough statistical studies of subpopulations or subgroups.
Installation instructions are here: <https://pglpm.github.io/inferno/index.html#installation>.
A step-by-step tutorial, guiding you through an example analysis of a simple dataset, is here: <https://pglpm.github.io/inferno/articles/vignette_start.html>.
The package has already been tested and used in concrete research about Alzheimer's Disease, Parkinson's Disease, drug discovery, and applications to machine learning.
Feedback is very welcome. If you find the package useful, feel free to advertise it a little :)
'How good is your Laplace approximation of the Bayesian posterior? Finite-sample computable error bounds for a variety of useful divergences', by Mikołaj J. Kasprzak, Ryan Giordano, Tamara Broderick.
http://jmlr.org/papers/v26/24-0619.html
#bayesian #posterior #laplace
#Bayesian analysis simplified (#BAYAS): our paper is out! We hope that many biologists & other users will find BAYAS helpful for experimental planning & data analysis. No programming or installation required. Will hopefully lead to reduction of lab animal numbers. #3R #bayes https://doi.org/10.1093/bioinformatics/btaf276
🚀 New to Variational Inference?
Check out A Beginner’s Guide to VI — a hands-on tutorial from @pymc_labs & @PyData Virginia, led by @fonnesbeck
📺 https://www.youtube.com/watch?v=XECLmgnS6Ng
🎥 Learn VI, ADVI, Pathfinder, & real-world modeling tips from Chris.
Briahna Joy Grey and Cory Doctorow
https://www.iheart.com/podcast/867-bad-faith-71674279/
#scifi - Capitalist Store Dedications / #plug / #patronage / #monetization
selections from Little Brother
. . . yet another instruction manual for Big Brother . . .
“But what if I decreed that from now on, every time you went to evacuate some solid waste, you’d have to do it in a glass room perched in the middle of Times Square and you have to be buck naked.” (pg. 116)
“the human condition is dignified by privacy”
“taking a picture was either piracy . . . or terrorism (in a public place)”
Paradox of the False Positive
[‘(just?) one string of many’] #Bayesian
“the Department of Homeland Security had set itself up to fail badly”
the Universal Machine can do anything - the favela Xbox hack, ParanoidXBox, Xnet ( #USkillPirates )
“ #chaff “ – Bayesian Camouflage (pg. 169)
“never underestimate the determination of a kid who is time-rich and cash-poor”
“all I had to do was convince everyone else to use it too” #EFF
[compare twisted logics now in vogue: not a bloke = average so no reproduction.
spam out normal = the 1% - according to SWEspayson swap pryting]
#statstab #360 Bayes Factor Design Analysis {bfda}
Thoughts: Sample size planning is confusing at first with Bayesian. But BFDA is the quick answer.
#statstab #357 Uncertainty Estimation with Conformal Prediction
Thoughts: Haven't parsed this properly but maybe be an interesting discussion point. How best to quantify uncertainty?
#conformalprediction #bayesian #confidenceintervals #uncertainty
