English subtitles

← PCA for Data Transformation - Intro to Machine Learning

Get Embed Code
4 Languages

Showing Revision 4 created 05/25/2016 by Udacity Robot.

  1. So now through all these misleading quizzes that I gave you,
  2. you now are in the perfect position to understand principal component analysis.
  3. And I will explain to you in 20 seconds, and it's going to stick for
  4. the rest of your life.
  5. If your given data of any shape whatsoever,
  6. like this data point cloud over here, PCA finds a new coordinate system that's
  7. obtained from the old one by translation and rotation only.
  8. And it moves the center of the coordinate system with the center of the data,
  9. right over here.
  10. It moves the x-axis into the principal axis of variation, where you
  11. see the most variation relative to all the data points, and it moves further
  12. axis down the road into a orthogonal less important directions of variation.
  13. So look at the original data here.
  14. I would argue that the data really lies mostly on this line over here.
  15. And then secondly on your orthogonal line over here, and principal component analysis finds
  16. for you these axes and also tells you how important these axes are.
  17. So let's exercise this.