Unformatted text preview: as a high dimensionality and so rather than trying to learn an SVM in a high dimension space
you think it would be a smart approach to come up with a series of 1 dimensional stubs that can be used
to construct a boosting classifier. Fill in the classifier table below. Each of the different classifiers are
given a unique ID and a test returns +1 (friend) if true and -1 (enemy) if false.
Classifier Test A Species is a Wizard B Species is an Elf C Species is not a Man D Does not have magic E Is not part of the Fellowship F Has never owned a ring of power G Hair <= 1ft H Hair <= 2 ft I Friend J Misclassified Enemy 1, 5, 6, 9 1, 4, 7, 8 2, 7 2, 3, 8, 10 A2 (4 points)
Looking at the results of your current classifiers, you quickly see two more good weak classifiers
(make fewer than 4 errors). What are they?
Classifier Test Misclassified K
L 7 Part B: Build a Strong Classifier (30 points)
B1 (25 points)
You realize that many of your te...
View Full Document
This document was uploaded on 03/17/2014 for the course EECS 6.034 at MIT.
- Fall '10
- Artificial Intelligence