Mark van der Laan

@mark_vdlaan

Jiann-Ping Hsu/Karl E Peace Professor in Biostatistics & Statistics . Targeted (machine) Learning for in with .

Vrijeme pridruživanja: studeni 2019.

Tweetovi

Blokirali ste korisnika/cu @mark_vdlaan

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

  1. proslijedio/la je Tweet

    Register for the Turing Institute lecture by Mark van der Laan on 4th March. All welcome and

    Poništi
  2. proslijedio/la je Tweet
    28. sij

    On 3rd of March, and I are giving a causal inference masterclass For details and to apply to attend, visit this link

    Poništi
  3. 27. pro 2019.

    Asked for a comparison of TMLE and double machine learning, this blog post comparing the two approaches in historical context came together quickly

    Poništi
  4. proslijedio/la je Tweet
    23. pro 2019.

    Curious about mediation analysis under intermediate confounding? Check out new effects and robust estimation strategies in exciting new work from , , ! (+ I got to help too.) 📦 coming soon.

    Poništi
  5. proslijedio/la je Tweet
    21. pro 2019.

    Exciting news, and there’s more: together with and we’re running a masterclass the previous day (March 3rd). Applications will open early in the new year. Watch this space!

    Poništi
  6. 16. pro 2019.

    During a panel discussion on bridging theory and application, I commented on the role of the statistician in scientific practice; later received this wonderful note from — always nice to see others agreeing on the importance of transparency in our field. 2/2

    Prikaži ovu nit
    Poništi
  7. 16. pro 2019.

    Greatly enjoyed the workshop last week, plenty of stimulating discussion with many impressive colleagues. 1/2

    Prikaži ovu nit
    Poništi
  8. 4. pro 2019.

    How about a statistically unbiased estimate of something interesting, and don't even bother estimating something that isn't interesting?

    Poništi
  9. 4. pro 2019.

    "The majority that one is taught in statistics is based on such unrealistic [parametric] assumptions and is actually wrong when applied to the real world."

    Prikaži ovu nit
    Poništi
  10. 4. pro 2019.

    "If we replace the real problem by a toy problem, we avoid the real challenges."

    Prikaži ovu nit
    Poništi
  11. 4. pro 2019.

    "It is my philosophy that one should make a sincere effort to define the real statistical estimation problem as accurately as possible, making assumptions that are reasonable and can be defended."

    Prikaži ovu nit
    Poništi
  12. 26. stu 2019.

    Just responded to question regarding article I wrote, the same one that I shared in my last tweet. Answer: Are all models useless? Is any exact model possible — or useful?

    Prikaži ovu nit
    Poništi
  13. 20. stu 2019.

    The problem is that the world is not pre-packaged for this kind of statistical convenience — and the assumptions driving these models are too restrictive to capture the true distribution of data. 3/3

    Prikaži ovu nit
    Poništi
  14. 20. stu 2019.

    By using these kinds of models, we assume that our data is generated from a specific type of probability distribution; but if the model is incorrect, it may lead us to incorrect conclusions. 2/3

    Prikaži ovu nit
    Poništi
  15. 20. stu 2019.

    The world is full of bent coins, each bent in slightly different ways - eg, we often assume that data are well modeled by a normal curve; yet, in many real-life situations, this may be a poor model or only useful for broad, sweeping generalizations. 1/3

    Prikaži ovu nit
    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:

    ·