Direct and indirect algorithms for on-line learning of disjunctions

Theoretical Computer Science

Volumn 284, Issue 1, 2002, Pages 109-142

  Helmbold, D P ^a   Panizza, S ^b   Warmuth, M K ^a  

^a University of California Santa Cruz   (United States)

^b CITY UNIVERSITY OF NEW YORK   (United States)

Author keywords

Bayes algorithm; Mistake bounds; Multiplicative updates; On line learning

Indexed keywords

LEARNING ALGORITHMS; LEARNING SYSTEMS; ONLINE SYSTEMS; PROBABILISTIC LOGICS;

BAYES ALGORITHM;

COMPUTER SCIENCE;

EID: 0037031030     PISSN: 03043975     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0304-3975(01)00081-0     Document Type: Conference Paper

References (14)

1. Tracking the best disjunction
2. Beyond independence: Conditions for the optimality of the simple Bayesian classifier
3. Hinge loss and average margin
4. General convergence results for linear discriminant updates
5. Learning when irrelevant attributes abound: A new linear-threshold algorithm
6. Comparing several linear-threshold learning algorithms on tasks involving superfluous attributes
7. An apobayesian relative of Winnow
8. The weighted majority algorithm