Andrew R. Webb A loss function approach to model selection in nonlinear principal components. [Citation Graph (0, 0)][DBLP] Neural Networks, 1999, v:12, n:2, pp:339-345 [Journal]
Andrew R. Webb, David Lowe The optimised internal representation of multilayer classifier networks performs nonlinear discriminant analysis. [Citation Graph (0, 0)][DBLP] Neural Networks, 1990, v:3, n:4, pp:367-375 [Journal]