Multilayer Perceptron Class
Artificial Neural Network Class, programmed in C++.
I programed it years ago for a better understanding of the Neural Networks. Might have some bugs.
MLP has the following characteristics:
- Fully customizable layers, including multiple hidden layers.
- Two activation functions: Sigmoid and Hyperbolic tangent.
- Backward propagation of errors using gradient descent (delta rule).
- NN weights backup once it learns.
- Learning algorithm with learning rate (alpha) and momentum(eta)
Data is displayed in console with several options (weights, bias, activations).
Artificial Neural Network example
Building
Multilayer Perceptron Class can be compiled with any compiler that supports C++.