BP-ANNA (Artificial Neural Network Architecture) is an efficient object oriented Back-Propagation artificial neural network library developed to nicely fit the FLTK library. The training algorithms include the Step Descent, Conjugate Gradient and Levenberg-Marquardt. The distribution include the source code and a two demo applications for Linux systems. BP-ANNA is very flexible to set multiple inputs, layers, neurons per layer and outputs in a simple way. There is a graphical neuronal network structure visualizer.

The input data file are columns in plain text format. There are few examples in the bin directory. The outputs are taken from the last column backward to the first. So if the matrix has four columns you can take maximum the last three ones as outputs.

Get the latest version with Git:

$ git clone https://github.com/eetorres/bpanna.git

or Zip file

https://github.com/eetorres/bpanna/archive/master.zip