Äh, AINA pitäisi tehdä kirjanmerkki kun jotain joskus näkee. Nyt en löydä mistään sitä, kuka täällä Mastodonissa puhui uudesta avoimen lähdekoodin tilastomatemaattisesta ohjelmasta #WinBUGS / #JAGS / #Stan-perheestä. Ei tule enää nimi mieleen ei ohjelmalle eikä asiasta maininneelle. #tiedejuttuja

Por fin pude zafarme de usar #AnyDesk y #TeamViewer, al emplear VNC para ocupar un software que tengo instalado en una máquina virtual con #Winbugs, lo lamento, es que me había dado flojera investigar antes 😅

My plan for my project in my #bayesian class - recreating (and possibly updating) a master's paper written in 1999 for the Air Force institute of Technology as a basic #bayesian network - I might see if some of the categorical distributions in this paper can be updated to continuous one (is there a way to convert the outcomes of continuous distribution to a categorical one in #winbugs?)

When you choose a million samples in #winbugs and the code takes forever to run and you learn about the "Sunk Cost Fallacy".

Today I learned some weird things about #WinBugs - deviance and Bayesian versions of the AIC. Stochastic variable selection. Plus, on my own: the "sink" command in #rstats #datascience #bayesian

Today in #winbugs: learned about how to handle missing data as well as learning about censored data (that term has a special meaning in #datascience and has not much to do with 'censorship'). I'm amazed at what WinBugs can do.

I now have got a multinomial likelihood with a Dirchlet prior in Winbugs going. It's as elementary as hell, but small triumphs are sometimes the best ones - it's like going from a dark room to someone lighting a match! #winbugs #datascience #dataanalysis

The next step is implenting a multinomial distribution in WinBugs with a Dirichlet prior. The issue: I can't get even the most elementary multinomials and Dirichlets to work. #dataanalysis #winbugs