Bayesian Statistics Made Simple | Scipy 2019 Tutorial | Allen Downey

Поділитися
Вставка
  • Опубліковано 21 лис 2024

КОМЕНТАРІ • 18

  • @joaopedrorocha5693
    @joaopedrorocha5693 Рік тому

    This was amazing! The grid approach is great to understand the concepts ... i was stuck while learning Bayesian statistics because i was being directly presented to the MCMC approach, which made things look more complex ... now things clicked for me. Thanks!

  • @tizgelmi
    @tizgelmi Місяць тому

    what an amazing lecture, thank you so much!!!

  • @sammathew243
    @sammathew243 5 років тому

    Allen has been lecturing this topic for several years now on Scipy and I am very grateful that he's teaching the common public such important content.

  • @arainboldt
    @arainboldt 5 років тому +10

    great lecture! really appreciate the thorough introduction. Mr. Downey is super enthusiastic and its great!

  • @henrmota
    @henrmota 4 роки тому

    Woww, underrated workshop. For me this workshop and Bayesian Inference is Just Counting, give me the necessary clarity as a beginner to start in this subject. Surprisingly they explain the problem in two different manners that complement each other.

  • @niceday2015
    @niceday2015 3 роки тому +1

    Thank you very much Mr. Allen Downey and Scipy ,Enthought

  • @sousou_no_freiren
    @sousou_no_freiren 4 роки тому +1

    Great tutorial! Would suggest that for future tutorials, you should have the attendees generate their own binomial pfm via spreadsheet. More than anything, doing all the steps manually really helped me gain understanding of the mechanical steps involved.

  • @sammathew243
    @sammathew243 5 років тому

    At 1:42:20, Allen corrects the verbose state of the function, but this error still persists on the GitHub page - Allen, just in case you are hearing!

  • @sau002
    @sau002 2 роки тому

    I like the passion.

  • @sammathew243
    @sammathew243 5 років тому

    I wanted to know about the Euro problem, whether the assessment by Barry Light (p < 0.07) was right or wrong, based on the analysis done through the Bayesian estimation?

    • @sousou_no_freiren
      @sousou_no_freiren 4 роки тому

      The point is that you can determine exact prob distribution via Bayes where as frequentist method only provides a point estimate.

  • @yt-1161
    @yt-1161 2 роки тому

    1:31:18 "the disease"

  • @Garnol
    @Garnol 5 років тому +2

    notebooks github.com/AllenDowney/BayesMadeSimple

  • @maxsangym1656
    @maxsangym1656 5 років тому

    There are typos in the import statements for empyrical _dist in notebook 2,3

    • @sammathew243
      @sammathew243 5 років тому +1

      In case you did not figure it out yet, empyrical_dist has now been replaced with empiricaldist package, that you can install through pip. The latest notebooks on the github page work with this newer package that Allen updated to.

  • @maxsangym1656
    @maxsangym1656 5 років тому

    d6+d6 is not working as shown it is returning 6 elements with the corresponding probabilities

    • @sammathew243
      @sammathew243 5 років тому

      d6+d6 will work only if you are using empyrical_dist package, which Allen used at the workshop. Now, the new notebooks rather use empiricaldist package (available on PyPI), which has d6.add_dist(d6).