GLM Intro - 4 - Link Function

Поделиться
HTML-код
  • Опубликовано: 22 янв 2025

Комментарии • 31

  • @MeerkatStatistics
    @MeerkatStatistics  2 года назад

    Full course is now available on my private website. Become a member and get full access:
    meerkatstatistics.com/courses...
    * 🎉 Special RUclips 60% Discount on Yearly Plan - valid for the 1st 100 subscribers; Voucher code: First100 🎉 *
    “GLM in R” Course Outline:
    Administration
    * Administration
    Up to Scratch
    * Notebook - Introduction
    * Notebook - Linear Models
    * Notebook - Intro to R
    Intro to GLM’s
    * Linear Models vs. Generalized Linear Models
    * Least Squares vs. Maximum Likelihood
    * Saturated vs. Constrained Model
    * Link Functions
    Exponential Family
    * Definition and Examples
    * More Examples
    * Notebook - Exponential Family
    * Mean and Variance
    * Notebook - Mean-Variance Relationship
    Deviance
    * Deviance
    * Notebook - Deviance
    Likelihood Analysis
    * Likelihood Analysis
    * Numerical Solution
    * Notebook - GLM’s in R
    * Notebook - Fitting the GLM
    * Inference
    Code Examples:
    * Notebook - Binary/Binomial Regression
    * Notebook - Poisson & Negative Binomial Regression
    * Notebook - Gamma & Inverse Gaussian Regression
    Advanced Topics:
    * Quasi-Likelihood
    * Generalized Estimating Equations (GEE)
    * Mixed Models (GLMM)
    Why become a member?
    * All video content
    * Extra material (notebooks)
    * Access to code and notes
    * Community Discussion
    * No Ads
    * Support the Creator ❤

  • @nhatminhnguyenhoang888
    @nhatminhnguyenhoang888 Год назад +4

    Be honest, this series is one of the best series about GLM that i ever learn before. It explain the math concept very clear

  • @anniezou8901
    @anniezou8901 3 года назад

    Your videos are my life savior.

  • @ryandikdan
    @ryandikdan 4 года назад +5

    Great video series! Really explains these tricky concepts well

  • @ankitbioinfo
    @ankitbioinfo 3 года назад

    I saw many videos on these topics but your is the best one.

  • @J12f41
    @J12f41 3 года назад +3

    I could not find an easy explanation on why we say that we apply the link function to the target mean instead of saying to the target. The ending to your video captures that point. Your very basic initial drawing succinctly answered my question finally. If you are able to explain a complicated subject in simple terms you have mastered it. You showed that here. Thank you very much.

    • @MeerkatStatistics
      @MeerkatStatistics  3 года назад +1

      Thanks for the nice words, but I'm far from mastering it. Maybe one day ;-).
      To elaborate on your point - for the case of logistic regression for example, it makes little sense to transform the target which is just binary (0 or 1). While p (or mu) could be modeled as continuous in [0,1] - it does make sense to model it - and transform it for the reasons mentioned in the video.
      For other distributions it depends on your assumptions. In the case of Normal, y~N and log(y)~N are 2 different assumptions, (which can be checked, e.g., using residuals). Using the link function essentially keeps you in y~N while allowing you to model a different relation between the X's and the Y's.

  • @juanfa98
    @juanfa98 10 месяцев назад

    great series

  • @mivuyompateni9075
    @mivuyompateni9075 2 года назад

    Good video , how can i find the hidden/unavailable video parts after this one ?

    • @MeerkatStatistics
      @MeerkatStatistics  2 года назад

      They are available to members on my website. Check the video description.

    • @mivuyompateni9075
      @mivuyompateni9075 2 года назад

      @@MeerkatStatistics please take me through it , i wanna join the Membership and i need the other parts to prepare for my interview .how do i do it ?

    • @MeerkatStatistics
      @MeerkatStatistics  2 года назад

      @@mivuyompateni9075 Go here: meerkatstatistics.com/courses/generalized-linear-models-glms/, click buy membership, choose monthly/yearly/lifetime - if you choose yearly, click "Click here to enter your discount code." and add the voucher code "FIRST100"; then fill up your details (username and password), and click checkout via PayPal.

    • @mivuyompateni9075
      @mivuyompateni9075 2 года назад

      @@MeerkatStatistics hey i did pay for the membership, Invoice #2F08182D94. videos are still not showing

    • @MeerkatStatistics
      @MeerkatStatistics  2 года назад +1

      @@mivuyompateni9075 Thanks for the purchase! The videos are not on RUclips but are on my website (which is better - no ads). Click start course here: meerkatstatistics.com/courses/generalized-linear-models-glms/ you should have full access now for a month (I think I saw you took the monthly plan)

  • @RCNNNNCR1357
    @RCNNNNCR1357 Год назад

    At the end shouldn't mu be equal to 1/(1+e^-xb)? I think what you have is 1- mu.

    • @MeerkatStatistics
      @MeerkatStatistics  Год назад +1

      It's the same. Just multiply the numerator and denominator by e^xb

    • @RCNNNNCR1357
      @RCNNNNCR1357 Год назад

      That's right :D@@MeerkatStatistics

  • @ianstats97
    @ianstats97 Год назад

    What are topics 5-6 and 7? I sent a message to your website, trying to understand what is on the full course prior of purchase cheers

    • @MeerkatStatistics
      @MeerkatStatistics  Год назад

      They are available on my website course, for a pay.
      If you press on the curriculum tab here you should see the table of content:
      meerkatstatistics.com/courses/generalized-linear-models-glms/
      Here it is anyway:
      “GLM in R” Course Outline:
      Administration
      * Administration
      Up to Scratch
      * Notebook - Introduction
      * Notebook - Linear Models
      * Notebook - Intro to R
      Intro to GLM’s
      * Linear Models vs. Generalized Linear Models
      * Least Squares vs. Maximum Likelihood
      * Saturated vs. Constrained Model
      * Link Functions
      Exponential Family
      * Definition and Examples
      * More Examples
      * Notebook - Exponential Family
      * Mean and Variance
      * Notebook - Mean-Variance Relationship
      Deviance
      * Deviance
      * Notebook - Deviance
      Likelihood Analysis
      * Likelihood Analysis
      * Numerical Solution
      * Notebook - GLM’s in R
      * Notebook - Fitting the GLM
      * Inference
      Code Examples:
      * Notebook - Binary/Binomial Regression
      * Notebook - Poisson & Negative Binomial Regression
      * Notebook - Gamma & Inverse Gaussian Regression
      Advanced Topics:
      * Quasi-Likelihood
      * Generalized Estimating Equations (GEE)
      * Mixed Models (GLMM)
      Best,
      David

  • @emiliovillegas2000
    @emiliovillegas2000 Год назад

    Do you have a discount code this black friday for your yearly plan?

    • @MeerkatStatistics
      @MeerkatStatistics  Год назад

      That's a good idea. Here you go, use the coupon code: BlackFriday and get 50% discount on all plans.

  • @JinaneJouni
    @JinaneJouni 2 года назад

    great job. thank you a lot

  • @kwstasvas36
    @kwstasvas36 4 года назад

    Really good video. Thank you a lot

  • @prashant0104
    @prashant0104 3 года назад

    I understand that the GLS transforms the predictors and response variables such that the error term becomes an identity matrix. So does this mean that the GLS is not a type of GLM since you say that the link function does not transform Ys and Xs? Thank you.

    • @MeerkatStatistics
      @MeerkatStatistics  2 года назад

      GLM and GLS are two different things. GLMs are used to model binary/binomial data, strictly positive data (Gamma), count data (Poisson) etc.; GLS is a way to find the best coefficients using least squares when you assume the variance matrix of the residuals is not an identity matrix.

  • @imrul66
    @imrul66 3 года назад

    Hi! how does this translates into practice? For example, when we are calling logit regression of y on x in Stata/R, what is happening behind the scene?

    • @MeerkatStatistics
      @MeerkatStatistics  3 года назад

      Watch the following videos... Basically it does Maximum Likelihood like I show in the Likelihood Analysis video, and solves it numerically, like I show in the Numerical Solution video. I will probably add a specific video for Logistic Regression sometimes soon.

  • @TiagoPereira-hm1nq
    @TiagoPereira-hm1nq 3 года назад

    Great! Bravo!

  • @prashant0104
    @prashant0104 3 года назад

    thnx you!

  • @vncsna
    @vncsna 4 года назад

    Thanks²

  • @Chale1288
    @Chale1288 3 года назад +1

    Didn't help :(