... (3.13) The linear scaling function for [−1, 1], transforming a variable xk into x∗∗ , has the following form: k x∗∗ = · k,t xk,t − min(xk ) −1 max(xk ) − min(xk ) (3.14) A nonlinear scaling method ... network link the input variables x to the encoding neurons C11 and C12, and to the nonlinear principal components The parameters also link the nonlinear principal components to the decoding neurons ... default in credit cards and in banking- sector fragility (Chapter 8) For dimensionality reduction, the race is between linear principal components and the neural net auto-associate mapping We show, in...