Got a YouTube account?

New: enable viewer-created translations and captions on your YouTube channel!

English subtitles

← SVMs and Tricky Data Distributions - Intro to Machine Learning

Get Embed Code
5 Languages

Showing Revision 3 created 05/25/2016 by Udacity Robot.

  1. Now I tricked you,
  2. it's actually this line over here and you might have gotten this right.
  3. Now this does maximize the margin, in some sense to all the data points.
  4. But it makes a classification error.
  5. The red x is on the wrong side of the green line.
  6. Whereas in this case over here, all the points are classified correctly.
  7. Support vector machine puts first and
  8. foremost the correct classification of the labels, and then maximize the margin.
  9. If we didn't care about correct classification,
  10. this line back here has an even larger margin to all the training examples.
  11. That's obviously the wrong line.
  12. So for support vector machines, you are trying to classify correctly, and
  13. subject to that constraint, you maximize the margin.