Demystifying Support Vector Machines for Beginners: A Reading List

Abstraction: Curated SVM learning resources for beginners across levels

Key points:

• SVMs were developed by Vladimir Vapnik from foundations laid in the 1970s; shot to prominence on handwriting recognition tasks matching neural network accuracy.
• SVMs use structural risk minimization (superior to neural network empirical risk minimization), giving them better generalization.
• Recommended starter book: "Support Vector Machines and other Kernel Based Learning Methods" by Nello Cristianini and John Shawe-Taylor — conceptual intro before heavy math.
• "Learning with Kernels" by Bernhard Scholkopf and Alexander Smola also recommended for beginners; Vapnik's own books suggested only after foundational readings.
• Curates 16+ video lecture courses (UC Berkeley, videolectures.net), 13+ PDF tutorials (Andrew Ng CS229 notes, Marti Hearst), and lecture slide sets from CMU, MIT, Stanford, ASU.
• Star-marked resources are personally vetted by the author; beginners advised to skip Bayesian and manifold learning sections initially.

Connections: Vladimir Vapnik · Bernhard Scholkopf · Support Vector Machines · Kernel Methods · Statistical Learning Theory

Source: http://onionesquereality.wordpress.com/2008/08/31/demystifying-support-vector-machines-for-beginners/