@TAH_Sci @MaartenvSmeden I agree, although I take more of a pragmatic Bayesian viewpoint in that we want to estimate a complex model as accurately as possible (in terms of prediction & variable selection) so we need shrinkage to avoid overfitting, but there are limits to what shrinkage can do.
@ESteyerberg@MaartenvSmeden I would say this is exactly the small n, large p situation where we need to use shrinkage to be able to estimate the model. How well we can do that is the interesting question I try to answer in the paper.
BMS-Aned Fall meeting: Friday October, 7th with @MaartenvSmeden, @SaravanErp and @ErikVanZwet on "Shinkage methods: novel perspectives and applications". Registration is open: https://t.co/ewoYVgZ8Ld
@DarinaPetrov But the basic setup is: I make notes throughout the day in a daily note and review & transfer relevant ones to permanent notes (based on “How to take smart notes). I have a kanban note for all my research projects & tag notes according to projects.
@DarinaPetrov That’s right! I actually recently discovered the usefulness of plugins such as dataview and kanban which made me reorganise my whole system. Thinking of sharing this more extensively once I’ve spend some time working like this and made sure it’s useful. ->
Recording from today's excellent OWYS session on Selection of Priors in Bayesian SEM is online.
Hoping you enjoy watching!
Thanks to our speakers @winterstat, Milica Miočević and Mauricio Garnier-Villarreal and great discussant @SaravanErp!
@YoungStatS2
https://t.co/gc5ukuFt90
Would you like to (learn to) use advanced Bayesian statistical methods to study protective factors in education? Apply for a PhD position on our interdisciplinary Road to Resilience project: https://t.co/WXZnaYTob7
Interested in Bayesian SEM? Join the YSE webinar next week (April 20th) @ 16:00 CEST. Speakers are Milica Miočević, Sonja D. Winter, and Mauricio Garnier-Villareal. Info and registration: https://t.co/H548WdV3vw
Excited for the SEM conference that is starting today with a workshop by Yves Rosseel on SEM with lavaan and one by me on Bayesian SEM! You can still follow online without registration via the Zoom link here: https://t.co/jjfU6YDKvQ
The conference "SEM: New Developments and Applications" starts next week. Registration to attend live is closed, but everyone who is interested can follow online (workshops and/or presentations)! Check out the program here: https://t.co/jjfU6YDKvQ