Difference between revisions of "Winnow Algorithm"
From Cohen Courses
Jump to navigationJump to search| Line 3: | Line 3: | ||
== Relevant Papers == | == Relevant Papers == | ||
# [http://www.springerlink.com/content/j0k7t38567325716/?MUD=MP Nick Littlestone (1988). "Learning Quickly When Irrelevant Attributes Abound: A New Linear-threshold Algorithm", Machine Learning] | # [http://www.springerlink.com/content/j0k7t38567325716/?MUD=MP Nick Littlestone (1988). "Learning Quickly When Irrelevant Attributes Abound: A New Linear-threshold Algorithm", Machine Learning] | ||
| − | # | + | # Littlestone, Nicholas. "Mistake bounds and logarithmic linear-threshold learning algorithms." (1990). |
Latest revision as of 06:09, 6 November 2012
Winnow Algorithm is a for learning a linear classifier/decision hyper-plane from labeled examples. It scales well to high dimensions especially when many of the dimensions are irrelevant. Thus it finds good use for text classification problems using a bag-of-words feature representation.
Relevant Papers
- Nick Littlestone (1988). "Learning Quickly When Irrelevant Attributes Abound: A New Linear-threshold Algorithm", Machine Learning
- Littlestone, Nicholas. "Mistake bounds and logarithmic linear-threshold learning algorithms." (1990).
