Kernel Methods for Pattern Analysis

by John Shawe-Taylor, Nello Cristianini

This book provides professionals with a large selection of algorithms, kernels and solutions ready for implementation and suitable for standard pattern discovery problems in fields such as bioinformatics, text analysis and image analysis. It also serves as an introduction for students and researchers to the growing field of kernel-based pattern analysis, demonstrating with examples how to handcraft an algorithm or a kernel for a new specific application, and covering all the necessary conceptual and mathematical tools to do so.

Genres: Computer Science, Artificial Intelligence, Science, Mathematics

478 pages, Hardcover

First published January 1, 2004

Reviews

[Manny · Rating 3 out of 5]

John Shawe-Taylor apparently lives and breathes for kernel methods. I suppose that if I were a hard-core kernel user I would worship him, but as it is I find him a bit too much.

I heard him give an invited talk at the COLING 2008 conference in Manchester. Usually, people who give invited talks treat it as an opportunity to sell their pet subject to researchers who don't know much about it. They present a non-technical overview, and try to find cool and exciting things that will make some of their listeners want to go and look up the details later on. Shawe-Taylor assumed everyone was already an expert: he just presented an hour of straight-up mathematics to an audience of computational linguists.

I didn't get more than 10% of the talk, and I've read all of his other book and perhaps half of this one. He's a strange guy. I couldn't help wondering if he might be Aspergers - it's fairly common in the research world.

[Michiel · Rating 4 out of 5]

Very clearly written background manual about the different kernel-based algorithms and the different kernels for specific tasks. The derivations of generalisation bound where the most lucid I have ever encountered.