CORNELL
CS 4780 -
Perceptron
Unformatted text preview:
View Full Document
Outline Perceptron CS4780 Machine Learning Fall 2009 Linear classification rules Perceptron learning algorithm Mistake bound model Perceptron mistake bound Thorsten Joachims Cornell University Reading Mitchell Chapter 4 4 4 4 2 Chapter 7 5 Cristianini Shawe Taylor Chapter 2 2 1 1 Linear Classification Rules Example Spam Filtering Hypotheses of the form unbiased Instance Space X Feature vector of word occurrences binary features N features N typically 50000 Target Concept c Spam 1 Ham 1 biased Parameter vector w scalar b Hypothesis space H Notation Online Perceptron Algorithm Margin of a Linear Classifier 1 Batch Perceptron Algorithm Example Reuters Text Classification optimal hyperplane 2
View Full Document
Unlocking...