About
The Neural Network Development Kit (NNDK) is a C++ template-based class library that implements a range of supervised and unsupervised artificial neural network classes, training algorithms; and auxiliary classes that implement algorithms for data splitting and variable selection.
The Neural Network Development Kit was developed as part of research undertaken by the School of Civil, Environmental and Mining Engineering at the University of Adelaide, with a goal to develop best-practices for the application of neural networks to water resources and environmental modelling applications.
The aim was to develop a class library to implement fundamental ANN structures and learning algorithms, which could be used to rapidly develop and evaluate new model development procedures.
License
This software is licensed under the GNU Public License. A copy of this license should be included with the source code repository. Otherwise, please read the license at the following page: http://www.gnu.org/copyleft/gpl.html.
Academic Citations
We are glad to share this software with those who may find it useful for their own research. If you intend to use this software for an academic publication or report, we would appreciate your acknowledgement of the origin of the software by referencing either of the following related articles as appropriate:
May R. J., Maier, H. R., Dandy, G. C. Data splitting for artificial neural networks using SOM-based stratified sampling. Neural Networks, 23(2), pp. 283-94, 201. Available for download at http://dx.doi.org/10.1016/j.neunet.2009.11.009 For input variable selection using PMI
R. J. May, H. R. Maier, G. C. Dandy, nd T. M. K. G. Non-linear variable selection for artificial neural network development using partial mutual information. Environmental Modelling and Software, 23(10-11), pp. 1312-1326, 2008. Available for download at http://www.sciencedirect.com/science/article/pii/S1364815208000467
Related Software An Excel add-in that provides a GUI for a pre-compiled set of tools for the development of ANN models for time-series forecasting and non-linear regression applications, was implemented using on the NNDK.
For further information please visit: http://www.ecms.adelaide.edu.au/civeng/research/water/software/neural-network-excel-add-in/