The Perceptron Algorithm vs. Winnow: Linear vs. Logarithmic Mistake Bounds when few Input Variables are Relevant.

Jyrki Kivinen Manfred K. Warmuth

Published in: COLT (1995)

Keyphrases

perceptron algorithm
input variables
mistake bound
multi class
variable selection
weight vector

winnow algorithm
artificial neural networks
learning algorithm
neural network model
hyperplane
linear threshold

fuzzy rules
membership functions
worst case
kernel classifiers
concept classes
support vector machine

data sets
pac learning
pairwise
active learning
feature space
neural network