Support Vector Machines
A Support Vector Machine, or SVM, splits two groups with a straight line. Many lines can do this job, so it picks the one with the widest empty road between the groups. In this lesson you see why a wide road is safer, and you move a line yourself.
New words
- Margin
- The width of the empty road between the line and the nearest points.
- Support vector
- A point that touches the edge of the road. Only these points matter.
- Hyperplane
- The dividing line in 2D, or a flat sheet in more dimensions.
- Soft margin
- A road that allows a few points to be on the wrong side.
- C
- A setting for how much each mistake costs. Big C means few mistakes allowed.
- Kernel trick
- Lifting points into a new space where a straight cut works.
What you will learn
- Which line should separate the groups?
- What is the margin?
- How far is a point from the line?
- How wide is the road for a line?
- How does the line classify new points?
- What if the groups overlap? Soft margin and C
- What if a straight line cannot work? The kernel trick
- When should you use an SVM?
- Try it: find the widest road
- Common mistakes
- Explain it to a friend
- Key takeaways
10 quiz questions at the end.
The full lesson, with interactive demos and the quiz, opens after you sign in with Google. It is free.