Winnow (algorithm): Difference between revisions

Revision as of 02:41, 14 March 2008

Revision as of 20:55, 14 September 2008

Line 1:

{{context}}

The '''winnow algorithm''' <ref>Littlestone, N. (1988) '''Learning Quickly When Irrelevant Attributes About: A New Linear-threshold Algorithm''' Machine Learning 285-318(2)</ref> is a technique from machine learning. It is closely related to the [[Perceptron]], but it uses a differentmultiplicative weight-update rule.scheme Winnowthat isallows effectiveit atperform removingmuch unhelpfulbetter than the perceptron when many dimensions are irrelevant (hence its name). It is not a sophisticated algorithm but it scales well to high-dimensional spaces. During training, winnow is shown a sequence of positive and negative examples. From these it learns a decision hyperplane.

The update rule is (loosely):