Training a neural network, along with testing and prediction, requires
that you specify a data set that contains the data to be used during
training. NeuralTools will either save your trained network directly
in your workbook or optionally, to a file on disk.
If all your data is in a single data set (including both known historical
data and new data where you do not know output values),
NeuralTools allows you to train and test a network, then predict
output values, all in a single step. You select to withhold a certain
percentage of the historical data for testing (20% is shown on the prior
page) and then select to automatically predict output values for cases
with missing dependent values. By doing this you quickly can get the
answers you need in one operation.
NeuralTools supports different neural network configurations to give
the best possible predictions. For classification/category prediction
(where the dependent variable is a category type), two types of
networks are available: Probabilistic Neural Networks (PNN) and
Multi-Layer Feedforward Networks (MLF). Numeric prediction can
be performed using MLF networks, as well as Generalized
Regression Neural Networks (GRNN), which are closely related to
PNN networks.