Practical Neural Network Recipies in C++

Fuzzy Data

and

Processing

319

group's triangle. This approach can more equitably distribute the

variable's information across network inputs.

Examples of Neural Network Fuzzy Postprocessing

A neural network can be trained to produce results directly in a form

that is the final goal. But it is frequently the case that it is better to

train the network to produce intermediate, more detailed results that

are then processed to yield final solutions. There Eire two advantages

to this approach. First, it may be easier to train the network to learn

the intermediate variables than to learn the final results. Second, it

provides long-term versatility ...