
316
Chapter
17
Examples of Neural Network Fuzzy Preprocessing
When a neural network solution is being proposed as a solution to a
problem, one of the many design options for consideration is the
appropriateness of the network's inputs. Certainly, the points
discussed in Chapter 16, Preparing Input Data, (page 253) should be
reviewed. Ideally, though, one should look beyond simple properties
of variables taken individually. What about the possibility of
important properties of individual variables in subranges of their
range? How about interactions of input variables which are easily
seen by the designer, but may be difficult for a network t ...