Analysis of two gradient-based algorithms for on-line regression

DOI10.1006/JCSS.1999.1635zbMath0961.68148OpenAlexW2033547792WikidataQ59538608 ScholiaQ59538608MaRDI QIDQ1970199

Nicolò Cesa-Bianchi

Publication date: 19 March 2000

Published in: Journal of Computer and System Sciences (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1006/jcss.1999.1635

zbMATH Keywords

gradient descentexponentiated gradient

Mathematics Subject Classification ID

Nonnumerical algorithms (68W05)

Related Items (9)

Adaptive and optimal online linear regression on \(\ell^1\)-balls ⋮ Optimal anytime regret with two experts ⋮ Randomized prediction of individual sequences ⋮ RECURSIVE FORECAST COMBINATION FOR DEPENDENT HETEROGENEOUS DATA ⋮ A continuous-time approach to online optimization ⋮ Sequential model aggregation for production forecasting ⋮ A quasi-Bayesian perspective to online clustering ⋮ Learning to assign degrees of belief in relational domains ⋮ Adaptive and self-confident on-line learning algorithms

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Exponentiated gradient versus gradient descent for linear predictors
• Decision theoretic generalizations of the PAC model for neural net and other learning applications
• A game of prediction with expert advice
• The weighted majority algorithm
• On-line learning of linear functions
• Universal prediction of individual sequences
• How to use expert advice
• A decision-theoretic extension of stochastic complexity and its applications to learning
• Probability Inequalities for Sums of Bounded Random Variables
• Convergence of stochastic processes
• Relative loss bounds for multidimensional regression problems

This page was built for publication: Analysis of two gradient-based algorithms for on-line regression