Ben Ogorek

@OgorekDataSci

Freelance data scientist just trying to figure things out

Vrijeme pridruživanja: lipanj 2019.

Tweetovi

Blokirali ste korisnika/cu @OgorekDataSci

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

  1. 2. velj

    Python test dependencies and : "tests_require should be packages that are used in the tests such as numpy and not packages that are used to conduct testing like pytest or nose." For pytest, use extras_require.

    Poništi
  2. 2. velj

    Finally learning what a Github release is exactly: "A release is a container of one or more assets, associated to a git annotated tag" (second answer at ). The tags often look like "v1.0.0" but I guess you could name them however you like.

    Poništi
  3. 29. sij

    Step by step instructions for releasing to PyPi. Very manual but they work. Going to dig into 's deployment process next.

    Poništi
  4. 26. sij
    Poništi
  5. proslijedio/la je Tweet
    23. sij

    Google Dataset Search is now officially out of beta. "Dataset Search has indexed almost 25 million of these datasets, giving you a single place to search for datasets & find links to where the data is." Nice work, Natasha Noy and everyone else involved!

    Prikaži ovu nit
    Poništi
  6. 20. sij

    While it gives me no joy to answer my own Q&A forum questions, here's what the "paths" are in Combinatorial Purged Cross Validation:

    Poništi
  7. 7. sij

    "Some reflection about priors beats cross-validation hacks anyday. But, in the absence of reflection, cross-validation on a _corpus_ helps pick and evaluate priors, and that’s what our paper hopefully demonstrates." - Aleks Jakulin on paper ().

    Poništi
  8. 7. sij

    Story from Bayesian (Dan Simpson) who used CV (at least) once: - Posterior is very flat - Different values of parameter led to very different out-of-sample predictions - Prior parameters very different to elicit So he used CV to choose this parameter

    Poništi
  9. 7. sij

    Are CV-based tuning methods and Bayesian estimation fundamentally at odds? Good discussion in the comments here:

    Poništi
  10. proslijedio/la je Tweet
    4. sij

    There’s a discovery method called “positive deviance”... basically finding pockets of positive outcome & studying what they are doing differently from e’one else. Often simple and disproportionate interventions are found. Thie article below reminded me of this though tangentially

    Poništi
  11. 4. sij

    Once we escape screens for display, the sky's the limit.

    Poništi
  12. 3. sij
    Poništi
  13. 3. sij

    The grand comedy of my life's work is that, while I was a terrible carpenter during my 2015 "sabbatical," this one YouTube video I made on the framing square has more views than any content I've ever created in Statistics or Data Science: What a joke!

    Poništi
  14. 3. sij
    Poništi
  15. 3. sij
    Poništi
  16. proslijedio/la je Tweet
    2. sij

    Ah yes, Jan 1st 2020... Python 2 is now formally laid to rest, meeting its end of life. No further updates, including security updates, to come. Don't worry though, Python 2.X will live on at large for at least another decade since companies still wont update.

    Poništi
  17. 1. sij

    More of the Friedman quote: "The process of obtaining serially uncorrelated residuals may in effect simply eliminate the permanent components, leaving the analyst to study...pure noise... in the process of seeking to satisfy mechanical statistical tests."

    Poništi
  18. 1. sij

    Nice quote from Milton Friedman: "I have mixed reactions to the current widespread tendency to regard serial correlation of residuals as a pure nuisance, if not the original sin, in analyzing time series."

    Poništi
  19. 1. sij

    The whole is not as coherent as some of the individual parts, but one takeaway is that differencing in the case of a "levels" model (level of y_t on level of x_t) magnifies the impacts of measurement error and model misspecification

    Poništi
  20. 1. sij

    Finally reading "A Brief Parable of Over-Differencing" by John Cochrane. It's starting off pretty good!

    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:

    ·