Exclusive Or

Example: The Obligatory Xor Function

Neural Network

This example is the neural network equivalent of Hello World. First load the packages required for the example.

The neural network created here has two input, three hidden and one output neuron. The classic incremental backpropagation algorithm is used.

Data

The data is a truth table for the exclusive or function with False replaced by zero and True replaced by one.

Training

The weights are initialized to assist the neural network in converging on a solution and the network is trained.

Visualization

The neural network and the mean square error can be visualized.

Results

The input values from the truth table are mapped over the neural net Execute function.

Mapping the output values to booleans gives the boolean value for each of the inputs.

Credits

This notebook is part of Fann for Mathematica © 2004 freegoldbar (http://www.geocities.com/freegoldbar/).

Created by freegoldbar  (September 29, 2004)