2011 | OriginalPaper | Buchkapitel
Nonlinear ICA
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
In the former chapters, it is assumed that the mixing model is linear; however, it is very different in practice since the mixture is always nonlinear mixture. Therefore, some error will be introduced into the solution. The difficulty in solving the nonlinear problem without any constraints is that it is exactly non-unique. The nonlinear ICA problem is much more complex than the linear model, so that a significant solution cannot be obtained by the assumption that the resources are statistically cannot mutually independent from each other. Thus, it is reasonable to make some restriction on the nonlinear function or mixing process. In recent years, many researchers have begun to study the nonlinear mixing model of ICA, such as Burel, 1992; Hermann and Yang, 1997; Lee
et al
, 1997; Lin and Cowan, 1997; Lin and Cowan, 1997; Table and Jutten, 1997; Hyvarinen and Oja, 1998; Hochreiter and Schemidhuber, 1998; Hyvarinen and Pajunen, 1999; Jutten, 2000; Jutten and Karhunen, 2003; Theis, 2003; Almeida, 2003, etc
[1–5]
.