← Back to papers
2003 JMLR JMLR 2003

On the Performance of Kernel Classes

Abstract

We present sharp bounds on the localized Rademacher averages of the unit ball in a reproducing kernel Hilbert space in terms of the eigenvalues of the integral operator associated with the kernel. We use this result to estimate the performance of the empirical minimization algorithm when the base class is the unit ball of the reproducing kernel Hilbert space. [abs] [ pdf ][ ps.gz ][ ps ]

🌉 Interdisciplinary Bridge — Machine Learning and Mathematics & Optimization
📈 Trend Setter — Geometry
🧭 Keyword Pioneer — eigenvalue decay
🐝 Cross-Pollinator — Artificial Intelligence, Computer Vision, Data Science & Analytics, Deep Learning, Healthcare & Medicine, Knowledge & Reasoning, Machine Learning, Mathematics & Optimization, Reinforcement Learning, Robotics, Security & Privacy
🐣 Hot Topic Early Bird — reproducing kernel hilbert space

Authors

Shahar Mendelson

Topics

Machine Learning > Core Methods > Classification Machine Learning > Optimization & Theory > Learning Theory Mathematics & Optimization > Mathematics > Geometry

Keywords

reproducing kernel hilbert space eigenvalue decay localized rademacher average empirical minimization learning guarantee
Download PDF

Related papers

Bottom-Up Relational Learning of Pattern Matching Rules for Information Extraction 2003
An Efficient Boosting Algorithm for Combining Preferences 2003
A Multiscale Framework For Blind Separation of Linearly Mixed Signals 2003
Word-Sequence Kernels 2003
An Extensive Empirical Study of Feature Selection Metrics for Text Classification 2003