Anzeige
Erschienen in:
2010 | OriginalPaper | Buchkapitel
4. Algorithms for Entropy and Correntropy Adaptation with Applications to Linear Systems
verfasst von : Deniz Erdogmus, Seungju Han, Abhishek Singh
Erschienen in: Information Theoretic Learning
Verlag: Springer New York
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
This chapter develops several batch and online learning algorithms for the error entropy criterion (EEC) that are counterparts to the most widely used algorithms for the mean square error criterion (MSE). Because the chapter assumes knowledge of adaptive filter design, readers unfamiliar with this topic should seek a textbook such as [332] or [253] for a review of fundamentals. But the treatment does not require an in-depth knowledge of this field. The case studies in this chapter address only adaptation of linear systems, not because entropic costs are particularly useful for the linear model, but because the solutions for linear systems are well understood and performance comparisons can be easily drawn. This chapter also considers applications of fast evaluations of the IP using the fast Gauss transform and incomplete Cholesky decomposition, and ends with an application of the error correntropy criterion (ECC) to adaptive noise cancellation.