Partial Least Squares Regression 1 Introduction (2/4)

แชร์
ฝัง
  • เผยแพร่เมื่อ 11 พ.ย. 2024

ความคิดเห็น • 24

  • @jimpappas2904
    @jimpappas2904 9 ปีที่แล้ว +24

    The way you presented this using the video within the video (loading vector rotating so that correlation between X and Y scores was maximized) was incredibly effective. I did not completely understand the concepts UNTIL I watched this video. You are such a good explainer and I look forward to seeing more of these videos.

    • @RasmusBroJ
      @RasmusBroJ 9 ปีที่แล้ว +1

      Jim Pappas Thanks so much. This is really, really nice to hear. Much appreciated.

  • @alexandermsutherland
    @alexandermsutherland 11 ปีที่แล้ว +6

    Thanks so much for this video! The animation in the beginning of the video clarified the difference between PCA vs. PLSR loadings in a crystal clear way.

    • @RasmusBroJ
      @RasmusBroJ 11 ปีที่แล้ว

      Thanks so much Alexander. This is really nice to hear.

  • @ghazalazadi3376
    @ghazalazadi3376 4 ปีที่แล้ว +1

    The video was so great. I finally realized what means when we say PLS does not actually do PCA on X and Y :)

  • @paulasouto5781
    @paulasouto5781 7 ปีที่แล้ว +1

    I completely agree with you, +Jim Pappas!!! Amazing video! Many thanks, +Rasmus Bro!

    • @RasmusBroJ
      @RasmusBroJ 7 ปีที่แล้ว +1

      I am very happy to hear that it is helpful. Makes it all worthwhile thanks

  • @adityanjsg99
    @adityanjsg99 4 ปีที่แล้ว +1

    Wikipedia and books didn't..this way of presentation did😊

  • @User_27a1
    @User_27a1 4 หลายเดือนก่อน

    So what is the biggest difference between PCA and PLS? PCA ends when loading and score vectors are created and a prediction of T vs U can be done AND PLS further more adjusts the loading vectors to improve the prediction factor r ?

    • @RasmusBroJ
      @RasmusBroJ 4 หลายเดือนก่อน

      PCA doesn't really try to predict. PCA aims to just describe the data in your matrix X. PLS tries to use X to predict some other property y. So PCA is essentially for visualizing your X, PLS is for using X to predict y

  • @RasmusBroJ
    @RasmusBroJ 11 ปีที่แล้ว

    You are most welcome

  • @malikkhan1391
    @malikkhan1391 5 ปีที่แล้ว

    At 5:50, do you know how the explained Variance in X and explained Variance in Y is computed? I am working on python with the MBPLS library, I have the scores, loadings, and weights, but I am not sure how to choose the components (more specifically, for my job, I want to make comparisons of explained Variance in X with PLS vs PCA)

    • @RasmusBroJ
      @RasmusBroJ 5 ปีที่แล้ว

      It is calculated from the model of X. Variance of the residual versus variance of the data.

  • @TheSepideh67
    @TheSepideh67 11 ปีที่แล้ว

    I was wondering if I am allowed to take a picture of one of your slides.

  • @futureramatb2443
    @futureramatb2443 7 ปีที่แล้ว

    This video is still one of the best videos that explain PLS very nicely. By the way, what software and how did you make the X variable loading to rotate? I would like to demonstrate that too! Thank you in advance !

    • @RasmusBroJ
      @RasmusBroJ 7 ปีที่แล้ว

      Thanks so much. It's been terribly many years but I believe this one was simply made by creating many plots in matlab using either the movie or movie2avi function. Maybe by first saving the frames/plots as jpgs.

  • @SkyGreatlove
    @SkyGreatlove 12 ปีที่แล้ว +1

    Thank you sooooooooooooooo much

  • @claudiaazevedo4073
    @claudiaazevedo4073 ปีที่แล้ว

    Hello where are u working? it's not R

    • @RasmusBroJ
      @RasmusBroJ ปีที่แล้ว

      We mainly work in MATLAB. It is perfectly suited for the kind of work we do

  • @TheSepideh67
    @TheSepideh67 11 ปีที่แล้ว

    Thank you. It was great.

  • @TheSepideh67
    @TheSepideh67 11 ปีที่แล้ว

    Thanks.

  • @MatsobaneTabogaMathiba
    @MatsobaneTabogaMathiba 10 ปีที่แล้ว

    how can i use PLS on Mathematica?

    • @VegarOttesen
      @VegarOttesen 9 ปีที่แล้ว

      Matsobane Taboga Mathiba I haven`t the foggiest myself, but I have used R for some time. It is very powerful and open source (so it is also free)

    • @RasmusBroJ
      @RasmusBroJ 9 ปีที่แล้ว

      ***** And it's good in MATLAB :-)