C++ Neural Networks and Fuzzy Logic: Preface

Representing the Weight Vector

Although on and off values are fine for the input vectors mentioned, you need to see grayscale values for the

weight vector. This can be accomplished by quantizing the weight vector into four bins, each represented by a

different ASCII graphic character, as shown in Table 12.1.

Table 12.1 Quantizing the Weight Vector

<= 0White rectangle (space)

0 < weight <= 0.25Light−dotted rectangle

0.25 < weight <= 0.50Medium−dotted rectangle

0.50 < weight <= 0.75Dark−dotted rectangle

weight > 0.75Black rectangle

The ASCII values for the graphics characters to be used are listed in Table 12.2.

Table 12.2 ASCII Values for Rectangle Graphic Characters

White rectangle255

Light−dotted rectangle176

Medium−dotted rectangle177

Dark−dotted rectangle178

Black rectangle219