Coding Variation

A blog about molecular biology, biochemistry, genetics and computation.

Tuesday, August 23, 2011

Combinations of weak predictors can be very effective.

From Boosting Algorithms: Regularization, Prediction and Model Fitting:
Kearns and Valiant [52] proved that if individual classifiers perform at least slightly better than guessing at random, their predictions can be combined and averaged, yielding much better predictions.
The reference is to Cryptographic limitations on learning Boolean formulae and finite automata.
Newer Posts Older Posts Home
Subscribe to: Posts (Atom)

Blog Archive

  • ▼  2011 (7)
    • ▼  August 2011 (1)
      • Combinations of weak predictors can be very effect...
    • ►  March 2011 (3)
    • ►  February 2011 (1)
    • ►  January 2011 (2)
  • ►  2010 (23)
    • ►  December 2010 (4)
    • ►  October 2010 (2)
    • ►  August 2010 (1)
    • ►  July 2010 (1)
    • ►  June 2010 (1)
    • ►  May 2010 (3)
    • ►  April 2010 (3)
    • ►  March 2010 (1)
    • ►  February 2010 (7)
  • ►  2009 (1)
    • ►  September 2009 (1)
Copyright 2009-2011 James Thompson. Powered by Blogger.