[svm] About support vector instances

0	If the dataset is not linearly seperatable, i.e. we can't find a perfect hyperplane to seperate the instances. Thus, which instances will become support vectors? Thanks in advance svm asked Mar 25 '13 at 22:32 lovejesus 0●1●1●2

2 Answers:

oldest newest most voted

Thank you for your quick response, Daniel. What about those instances that lie exactly in the "margin" (red dots in the image), are they support vectors?

alt text

answered Mar 26 '13 at 03:02

lovejesus
0●1●1●2

edited Mar 26 '13 at 03:07

points on the margin are unbound support vectors while those properly inside the margin are called bound vectors. btw, you do not want to respond to an answer with another answer. That is what comments are for.

(May 22 '14 at 04:34) Daniel Mahler

0	Support vectors are by definition those that are on the on the wrong side of the optimal hyperplane or or on the correct side, but still inside the margin. answered Mar 26 '13 at 01:10 Daniel Mahler 1226●3●13●22

Subscription:

Once you sign in you will be able to subscribe for any updates here

Tags:

svm ×134

Asked: Mar 25 '13 at 22:32

Seen: 776 times

Last updated: May 22 '14 at 04:34

[svm] About support vector instances

Subscription:

Related questions