This preview shows page 1. Sign up to view the full content.
Unformatted text preview: est predict whether a particular snowman can be thus transformed are the
snowman's height and the number of coal buttons on its chest. Their preliminary data is below: each
“+” represents a magical talking snowman, and each “–” represents an ordinary, non-transforming
snowman. A1 (5 points): You decide to use a linear SVM to classify magical vs. non-magical snowmen. Draw
the resulting decision boundary in the graph above, and circle the support vectors. 24 A2 (10 points): In this SVM, what is the vector w and the constant b? A3 (10 points): Suppose the F.R.O.S.T. scientists discover another magical, talking snowman that is
five feet tall and has six coal buttons on his chest. What would the α value of this new data point be?
Justify your answer. 25 Part B: Kernels (25 Points)
In this section, you will project the data below into a new space with ϕ(u) = < | x1 - x2 | >. That is, you
project the two-dimensional vector u into a one-dimensional vector in a one-dimensional space. B1 (7 points): What is the kernel function K(u,v) for this transformation? 26 B3 (10 points): In the transformed spa...
View Full Document
This document was uploaded on 03/17/2014 for the course EECS 6.034 at MIT.
- Fall '10
- Artificial Intelligence