NNFpp is a C++ evolution of the C library NNF developed by Alessandro Presta. NNFpp incorporate some algorithms from NNF and other created specifically for NNFpp. For every question about this library send me an e-mail. If you use this library for some project, and if you want, send me an e-mail with: project's name, author, a little description and eventually a link to that project so I can maintain an updated list!!
A neural network can perform many activities: categorization, recognition and many more. NNFpp can help you to develope C++ code that uses neural nets in easy and fast way. This class is compatible with the C++ standard and can be used on all OS. In the package you can find the Borland C++ Builder project file and a tutorial which can be compiled with all the compilers.
NEWS
February 2007 | Released version 3.0.1 Fixed some bug and added an STL version of NNFpp (written by Worik). |
October 2006 | Released version 3.0 New function for shuffle training; a new function to test neural networks. |
May 2006 | Released version 2.0 New activation function; a new class for errors management. |
April 2006 | Released version 1.0.2 Fixed bug in Learning function. |
March 2006 | Released version 1.0.1 Fixed problems with ResetNet function; new structure of file for NNFpp, for compatibility with new version of NNFpp add in top of your .nnf file the string "MLP". |
March 2006 | Released version 1.0 |