... useful information for making decisions basedon information from markets with large numbers of prices.As Fotheringhame and Baddeley (19 97) point out, if the underlying truestructure interrelating ... −min(xk)(3.13)The linear scaling function for [−1, 1], transforming a variable xkintox∗∗k, has the following form:x∗∗k,t=2·xk,t− min(xk)max(xk) −min(xk)− 1 (3.14)A nonlinear scaling method ... network link the input variables x to the encodingneurons C11 and C12, and to the nonlinear principal components. Theparameters also link the nonlinear principal components to the decodingneurons...