File	Date	Author	Commit
DataGenerator	2012-04-13	Sergio Roa	[7ddb0a] cleaning up setRefVectors functions
Graphs	2012-10-10	Sergio Roa	[2ab9c5] adding more pool allocators and cleaning code
GrowingNeuralGas	2012-10-17	Sergio Roa	[37f439] small changes for memory management
VoronoiDiagramm	2012-09-19	Sergio Roa	[eb4d84] Trying to solve eventual memory allocation issu...
doc	2012-03-27	Sergio Roa	[f0d178] Adding revision nr. to documentation
tools	2012-09-25	Sergio Roa	[a730e1] redefining pool allocator
CMakeLists.txt	2012-10-10	Sergio Roa	[2ab9c5] adding more pool allocators and cleaning code
LICENSE	2012-01-10	Sergio Roa	[ea2432] adding license information
NeuralGas.h	2012-09-19	Sergio Roa	[eb4d84] Trying to solve eventual memory allocation issu...
NeuralGasSuite.h	2012-04-13	Sergio Roa	[7ddb0a] cleaning up setRefVectors functions
README	2012-03-05	Sergio Roa	[ae5c3a] better memory management and some leaks fixing
logo.svg	2012-01-27	Sergio Roa	[4757bf] adding logo
mainNGSuite.cpp	2012-03-19	Sergio Roa	[344654] Much better memory handling when looking for gr...

Read Me

These are instructions to install NeuralGas on an Ubuntu machine.
Probably this code works also in MacOS (however it has not been tested).
Please let us know if you want a version for Windows or feel free to
provide the appropriate compilation scripts.

Install following packages (Ubuntu 10.04 and higher):
qt4-qmake
libqt4-dev

Install the boost dev package. In Ubuntu 10.04 you can install libboost1.40-dev 
or libboost-dev and libboost-program-options-dev. In Ubuntu 11.04 just
libboost1.42-dev and libboost-program-options1.42-dev.

Procedures to install NeuralGas:

Go to the NeuralGas root directory and type the following commands:

$ mkdir build
$ cd build
$ ccmake ..

At this point you can provide additional flags to your configuration and type 
the character [c] (configure) twice. If everything worked fine, you can finally 
type [g] for generating the makefile. The final commands are:

$ make
$ sudo make install

Installed programs are:

- NGSuite. Used for testing learning with two different data sets (MackeyGlass and NoisyAutomata) with different learning algorithms.
- LLRGNG. Testing of the Lifelong Robust GNG alg. with a simple distribution of 3 Gaussians.
- GaussianNoise. Testing of the Lifelong Robust GNG alg. with any mixture of Gaussians.
- EBGNG. Testing of a simple Error based GNG alg. with the MackeyGlass data set.
- GNGTesting. Verify the Average Error of some GNG graph given some data.
- VoronoiDiagramm. Generation of a Voronoi diagram as an image, given nodes and data.
- VoronoiWidget. Generation of a Voronoi diagram in a Qt Window, given nodes and data.

Enjoy!

Please send your comments to:

Sergio Roa (sergio.roa@dfki.de)