Causal Inference | Answering causal questions

Поділитися
Вставка
  • Опубліковано 15 лис 2024

КОМЕНТАРІ • 37

  • @FF-sw7rv
    @FF-sw7rv 2 роки тому +5

    This was the best video so far about Casual Inference...so simple and to the point.. Thank you

    • @ShawhinTalebi
      @ShawhinTalebi  2 роки тому

      Thanks for the comment, glad it was helpful!

  • @DeepFindr
    @DeepFindr 2 роки тому +4

    This was an awesome introduction! I've looked into the connection of Graph Neural Networks and Causality and this was a superb introduction! Subbed

  • @ShawhinTalebi
    @ShawhinTalebi  2 роки тому

    More in this series 👇
    Intro to Causality: ua-cam.com/video/WqASiuM4a-A/v-deo.html
    Causal Discovery: ua-cam.com/video/tufdEUSjmNI/v-deo.html
    Intro to Causal Effects: ua-cam.com/video/BOPOX_mTS0g/v-deo.html
    Propensity Scores: ua-cam.com/video/dm-BWjyYQpw/v-deo.html
    Do-operator: ua-cam.com/video/dejZzJIZdow/v-deo.html
    DAGs: ua-cam.com/video/ASU5HG5EqTM/v-deo.html
    Regression techniques: ua-cam.com/video/O72uByJlnMw/v-deo.html
    Towards Data Science article: towardsdatascience.com/causal-inference-962ae97cefda?sk=d68d5191fdb00d3fee47aaa43ed48f3d

  • @aamirkarim9220
    @aamirkarim9220 3 роки тому +1

    One of the best videos on the topic, and the examples are so easy to follow!

  • @SergeySetti
    @SergeySetti Рік тому

    I can't understand the low value of the views counter under this video. Because the way you explain this material is so good!

  • @kaldirYT
    @kaldirYT 3 роки тому +2

    Another great vid!
    Coming for that first spot

  • @rildodemarqui
    @rildodemarqui 2 роки тому +2

    Well explained. Congrats!

  • @ifycadeau
    @ifycadeau 3 роки тому +3

    Super informational! Thanks for sharing Shaw!

  • @zhanbolatmagzumov6409
    @zhanbolatmagzumov6409 3 роки тому +3

    Thank you a lot Shaw! Great video. Will you show some examples of Bayesian Network in python? This example was very helpful
    .

    • @ShawhinTalebi
      @ShawhinTalebi  3 роки тому +1

      Thanks Zhanbolat! That’s a great suggestion. I haven’t done much with Bayesian networks in Python, but it overlaps with causal inference. It could fit in nicely in an appendix to this series.

    • @zhanbolatmagzumov6409
      @zhanbolatmagzumov6409 3 роки тому +1

      @@ShawhinTalebi thank you very much! It will be very helpful!

  • @karannchew2534
    @karannchew2534 6 місяців тому

    9:40
    Average Causal Effect is 0.2 = Having degree increase chance of earning over 50k
    Confounder: Age
    Treatment: HavingDegreeOrNot Response: EarnOver50kOrNot
    Confounder: OverallScore
    Treatment: HaveSolutionOrNot
    Response: LocalScore

  • @hameddadgour
    @hameddadgour 7 місяців тому

    Great video! Thank you for sharing.

  • @TheTessatje123
    @TheTessatje123 Рік тому

    Thank you so much. I read "The book of why", in which the do-operator is mentioned, but it's not explained on how to use it!

    • @ShawhinTalebi
      @ShawhinTalebi  Рік тому

      Great book! Definitely meant for a general audience.
      I talk a bit more about the do-operator in another (technical) video: ua-cam.com/video/dejZzJIZdow/v-deo.html

  • @rezamohajerpoor8092
    @rezamohajerpoor8092 2 місяці тому

    Hi is it possible to use causal inference for feature selection in ML?

    • @ShawhinTalebi
      @ShawhinTalebi  2 місяці тому

      Interesting question. I haven't come across anything like that. Perhaps ideas from causal discovery could be used.

  • @user-wr4yl7tx3w
    @user-wr4yl7tx3w Рік тому

    Great content.

  • @norhanifahcali-ls9xr
    @norhanifahcali-ls9xr Рік тому

    Hahaha the example problem really pokes my anxiety. 😂😂😂

  • @jaspreet0305
    @jaspreet0305 8 місяців тому

    Doubt: Outcome should always be a binary column?

    • @ShawhinTalebi
      @ShawhinTalebi  8 місяців тому

      Not necessarily. This will depend on the method you are using.
      For instance, the T-learner method here can handle non-binary targets. However, using binary targets often simplifies the analysis.

  • @user-wr4yl7tx3w
    @user-wr4yl7tx3w Рік тому

    In what ways do you think causal inference and bayesian inference are related and unrelated? Is the latter only addressing correlation and therefore makes no causal statement.

    • @ShawhinTalebi
      @ShawhinTalebi  Рік тому +1

      Id say these are 2 different animals. Bayesian inference is a particular paradigm for doing inference. On the other hand, causal inference is not so much a particular approach to inference, rather it is any inference where the goal is to answer a question involving cause and effect. From this view, one could even use Bayesian inference to do causal inference
      Hope that helps!

  • @arvoli2596
    @arvoli2596 3 роки тому +1

    Love it!!

  • @ninjanitehawk7338
    @ninjanitehawk7338 3 роки тому +1

    1st

  • @fazlfazl2346
    @fazlfazl2346 6 місяців тому

    bad teaching.

    • @ShawhinTalebi
      @ShawhinTalebi  6 місяців тому

      I'm always looking to improve. What issues do you see here?