Online learning from a signal processing perspective
There is increased interest in kernel learning algorithms inneural networks and a growing need for nonlinear adaptivealgorithms in advanced signal processing, communications, andcontrols. Kernel Adaptive Filtering is the first book topresent a comprehensive, unifying introduction to online learningalgorithms in reproducing kernel Hilbert spaces. Based on researchbeing conducted in the Computational Neuro-Engineering Laboratoryat the University of Florida and in the Cognitive SystemsLaboratory at McMaster University, Ontario, Canada, this uniqueresource elevates the adaptive filtering theory to a new level,presenting a new design methodology of nonlinear adaptivefilters.
* Covers the kernel least mean squares algorithm, kernel affineprojection algorithms, the kernel recursive least squaresalgorithm, the theory of Gaussian process regression, and theextended kernel recursive least squares algorithm
* Presents a powerful model-selection method called maximummarginal likelihood
* Addresses the principal bottleneck of kernel adaptivefilters--their growing structure
* Features twelve computer-oriented experiments to reinforce theconcepts, with MATLAB codes downloadable from the authors' Website
* Concludes each chapter with a summary of the state of the artand potential future directions for original research
Kernel Adaptive Filtering is ideal for engineers,computer scientists, and graduate students interested in nonlinearadaptive systems for online applications (applications where thedata stream arrives one sample at a time and incremental optimalsolutions are desirable). It is also a useful guide for those wholook for nonlinear adaptive filtering methodologies to solvepractical problems.