Note: This library is still under initial development and shouldn't be used as of yet.
The perceptron algorithm is a binary classifier which served as a stepping stone to some of the first artificial neural networks ever produced. The algorithm itself is a single neuron with the unit step function as its activation function. The alorithm returns a true or false, signifying if the given input vector is a part of the perceptron's learned classification space.
Install SCons on your system and simply run the scons
command in the root directory of this repository.
include "includes/perceptron.hpp"
// Construct a perceptron and a trainer...
perceptron::Perceptron perceptron;
perceptron::Trainer trainer;
// Train your perceptron...
trainer.AssignTrainee(perceptron);
trainer.SetRate(RateValue);
while(IsMoreData())
{
trainer.Train(GetNextInput(), GetNextExpectedResult());
}
// Categorize with your perceptron...
if(perceptron.Evaluate(GetInput()))
{
cout << "It's a match!";
}
else
{
cout << "No match here.";
}
working on it...
working on it...