 |
||||
| Ralf Herbrich | ||||
| MIT Press | ||||
| 2002 | ||||
| English | ||||
| 364 pages | ||||
| 2.6 Mb | ||||
[tab] [content title="Summary"] Linear classifiers in kernel spaces have become a significant area of focus in machine learning. The kernel technique enhances the traditional linear classifier—while limited, it is well-established and thoroughly researched—by expanding its use to various nonlinear pattern-recognition tasks, including natural language processing, machine vision, and biological sequence analysis. This book offers the first comprehensive overview of both the theory and algorithms related to kernel classifiers, featuring the latest advancements in the field. It starts by detailing major algorithmic developments such as kernel perceptron learning, kernel Fisher discriminants, support vector machines, relevance vector machines, Gaussian processes, and Bayes point machines. Following this, the book provides an in-depth introduction to learning theory, covering concepts like VC and PAC-Bayesian theory, data-dependent structural risk minimization, and compression bounds. Throughout the text, there is a strong emphasis on the interplay between theory and algorithms, explaining how learning algorithms operate and their underlying rationale. The book is rich with examples, includes complete pseudocode for the presented algorithms, and offers an extensive source code library, making it an invaluable resource for researchers and practitioners in the field. [/content] [content title="Content"] [/content] [content title="Author(s)"] [/content] [/tab]
[facebook src="bibliosciencesorg"/]
Key-Words: Télécharger EBOOK PDF EPUB DJVU Learning kernel classifiers: theory and algorithms. Download EBOOK PDF EPUB DJVU Learning kernel classifiers: theory and algorithms.
