Absolute error bounds for learning linear functions online

Authors:
Ethan J. Bernstein
Affiliations:
Computer Science Division, University of California, Berkeley, Berkeley, CA
Venue:
COLT '92 Proceedings of the fifth annual workshop on Computational learning theory
Year:
1992

Citing 2
Cited 2

On-line learning of linear functions

STOC '91 Proceedings of the twenty-third annual ACM symposium on Theory of computing
ON-LINE LEARNING OF LINEAR FUNCTIONS

ON-LINE LEARNING OF LINEAR FUNCTIONS

From noise-free to noise-tolerant and from on-line to batch learning

COLT '95 Proceedings of the eighth annual conference on Computational learning theory
On-line evaluation and prediction using linear functions

COLT '97 Proceedings of the tenth annual conference on Computational learning theory

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this note, we consider the problem of learning a linear function from ** to ** online. Two previous algorithms have been presented which achieve an optimal sum of squared error terms. We show that neither algorithm performs well with respect to absolute error. We also show that in both problem settings, very simple ideas given an algorithm which achieves optimal or close to optimal worst case absolute error.