Arman Oganisian

@StableMarkets

Biostatistics PhD candidate | Nonparametric & at | Associate Fellow at . Bayesed & confused.

Philadelphia, PA
Vrijeme pridruživanja: ožujak 2011.

Tweetovi

Blokirali ste korisnika/cu @StableMarkets

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @StableMarkets

  1. 25. sij
    Prikaži ovu nit
    Poništi
  2. 25. sij

    Schedule for Summer Institute below. and I will be teaching sessions on Causal. Cutting-edge Bayes models w/ hands-on R coding!

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    16. sij

    Just posted this to arxiv: . The goal of the paper is to test the null hypothesis that a causal dose-response curve is flat on the support of the exposure in the no unobserved confounding setting. 1/

    Prikaži ovu nit
    Poništi
  4. 16. sij

    FYI. There will *probably* be some sessions on Bayesian causal inference.

    Poništi
  5. 8. sij

    Chairing a session on Causal Inference Methods for Health Policy Research tomorrow at . Great set of topics: DiD, multiple outcomes, synthetic controls. Worth dropping by!

    Poništi
  6. 3. sij

    Docs have this whole field of "translational medicine" that specifically aims to tie clinical practice and medical research. I wonder if there should be a field called "Translational Statistics" that links statistical methodologists/theorists and practitioners.

    Poništi
  7. 31. pro 2019.

    Mom: this brandy was aged 20 years! Normal person response: wow that’s a long time! Statistician response: huh. I wonder what the distribution is...

    Poništi
  8. proslijedio/la je Tweet
    30. pro 2019.

    I'm excited to share some new work out in ! In it, we propose a flexible multivariate extension of histogram analysis for MRI data | 📄 📦

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    28. pro 2019.

    Really nice blog post discussing the historical context of targeted maximum likelihood estimation (TMLE) and double machine learning (DML). Those not yet closely familiar with TMLE might be best served skimming the technical discussion, but can find helpful references within.

    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet

    Some people say that gifting cash isn't very "thoughtful." The truth is the opposite-- the most thoughtful thing you can do this Christmas is to minimize deadweight loss.

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet

    What can happen when using AI in decision making 🤣🎁🎅

    Poništi
  12. 11. pro 2019.

    Wasn’t expecting this but excited to head to Nashville! Thanks to and Nandita Mitra (still twitter-less!) for being such great advisors.

    Poništi
  13. 28. stu 2019.

    At my mom’s place for thanksgiving and found this copy of “Causal Diagrams for Epidemiological Research”. Printed this out around 2015 (pre phd) to get a feel for causal inference. Found it pretty cool so here I am.

    Poništi
  14. 24. stu 2019.

    Apparently a lot of the hype about MCT oil consumption started with this 2008 study . "There was a trend toward greater loss of fat mass (P = 0.071) " for MCT oil vs. olive oil.

    Poništi
  15. 15. stu 2019.

    This is what distributions drawn from a Dirichlet Process look like: very discrete. And it's that discreteness that induces posterior clustering

    Poništi
  16. 11. stu 2019.

    Good quote: "Nothing is wrong with making assumptions; causal inference is impossible without making assumptions, and they are the strands that link statistics to science. It is the scientific quality of those assumptions, not their existence, that is critical." - Rubin, 1986.

    Poništi
  17. proslijedio/la je Tweet
    9. stu 2019.

    This is really excellent advice on good presentation skills. The point of a good talk is to communicate not to create performance art demonstrating how smart you are.

    Poništi
  18. 27. lis 2019.

    And there's a performance difference. rbinom takes about 40% longer.

    Prikaži ovu nit
    Poništi
  19. 27. lis 2019.

    There are two types of people in the world. Those who simulate random binary events with probability p via rbinom(1,1, p) and those who do it via I(runif(1) < p).

    Prikaži ovu nit
    Poništi
  20. 18. lis 2019.

    “Statistics are one of many pillars of liberalism, indeed of the Enlightenment...yet in recent years, divergent levels of trust in statistics has become one of the key schisms that have opened up in western liberal democracies”.

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·