Multilayer Perceptron for non-linear regression within a common framework with Support Vector Machines and Radial Basis Function Regularised Networks for non-linear regression is presented. The aim is taking advantage of Support Vector Machines training paradigm to overcome curse of dimensionality and too strict hypothesis on statistics of errors in traditional Multilayer Perceptron for non-linear regression. In this context, an alternative strategy to Quadratic Programming based on 1-norm minimisation to avoid computational problems of Support Vector Machines is proposed.
|Autori interni:||GIUSTOLISI, Orazio|
|Titolo:||Sparse solution in training artificial neural networks|
|Data di pubblicazione:||2004|
|Digital Object Identifier (DOI):||10.1016/j.neucom.2003.09.005|
|Appare nelle tipologie:||1.1 Articolo in rivista|