MachineLearning version 1.03 DESCRIPTION This distribution includes a base module (MachineLearning.pm) and the following four modules for direct use: NeuralNetwork.pm - Implements a neural network that optimizes its output for quality of results (predictions) without regard to quantity of results (beyond a configurable minimum). NOTE: This module uses simulated annealing (implemented by SimulatedAnnealing.pm--see below) for optimization of the weights and thresholds. DecisionTrees.pm - Implements a "grove" of decision trees built on a single data source (a file containing "training" data). As with NeuralNetwork.pm, the output is optimized for quality at the expense of quantity. SimulatedAnnealing.pm - Implements simulated annealing for optimization of a collection of numbers with predefined ranges. IntegerAnnealing.pm - Implements simulated annealing for optimization of a collection of integers with predefined ranges. INSTALLATION To install the five modules included in this distribution, perform the following simple steps: 1. Copy MachineLearning.pm to a directory in @INC (for example, "/etc/perl" or "D:\Perl\site\lib"). 2. Copy the "MachineLearning" directory (containing the files SimulatedAnnealing.pm, IntegerAnnealing.pm, NeuralNetwork.pm, and DecisionTrees.pm) to a directory in @INC (for example, "/etc/perl" or "D:\Perl\site\lib"). 3. Check that you now have a directory structure like the following: /etc/perl/MachineLearning.pm /etc/perl/MachineLearning/DecisionTrees.pm /etc/perl/MachineLearning/NeuralNetwork.pm /etc/perl/MachineLearning/SimulatedAnnealing.pm /etc/perl/MachineLearning/IntegerAnnealing.pm AUTHOR Benjamin Fitch COPYRIGHT AND LICENSE Copyright 2009 by Benjamin Fitch This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.