Re: [Gneuralnetwork] New member

[John Salatas]

Hi again!

I had a look at the code and it seems pretty straightforward, although I didn't look at the implementation of the various training methods in big detail. I guess I'll look in more detail the implementation of gradient descent and genetic algorithm as I'm already familiar with these. However, I have to admit that I tried a couple of examples (ie the square root that is already in main) with more training data and various training parameters but I couldn't get it to produce any good results :\

Anyway, I guess I'll look at the code again and try to debug it so in order to get a better understanding. In the mean time some general thoughts/questions:

1) Is it planned to be a library that one can use in his own code or a standalone application? If it is going to be a standalone application then I guess it needs to somehow be able to read the input data and the neural network's structure and parameters from a file. Right?

2) Currently it only accepts as input numbers as inputs, but I believe it would be good to accept other types of inputs like for example strings. Imagine for example the case a training dataset containing people's data like sex which can only have distinct values ("male"/"female") or the case in which we have a dataset with containing people's data and we want to train a neural network to output a person's sex (ie to output distinct values instead of numbers).

Please, let me know what you think.

Regards,

John Salatas
Web Page: http://jsalatas.ictpro.gr

On Thu, Mar 17, 2016 at 2:08 PM, Jean Michel Sellier <address@hidden> wrote:

Hi John,

Welcome to this community! I am currently completing a new version of Gneural Network that, hopefully, should be soon released. It will be much more advanced than the one that is online right now. In the meanwhile, it would be great if you could have a look to the code that is online (version 0.0.1). Your experience could be VERY useful for this project so it would be great if, after having a look at the source code, you could share your comments and/or ideas! May be you can point us to some method/technique that nobody of us is aware of. In any case, please feel free to share your thoughts!

Best,

JM

2016-03-17 21:41 GMT+01:00 John Salatas <address@hidden>:

Hi!

I just saw this project and first of all I want to congratulate all of you for this.

In the past I have implemented some neural networks algorithms (Elman, Self Organizing Maps and Learning Vector Quantization) for the WEKA environment (http://www.cs.waikato.ac.nz/ml/weka/) so I could contribute to gneuralnetwork as well. The only drawback is my poor C programming skills/experience but I guess this wouldn't be a major problem.

Regards,

John Salatas
Web Page: http://jsalatas.ictpro.gr

--

Jean Michel Sellier, Ph.D., M.Sc.

www.nano-archimedes.com

www.gnu.org/software/archimedes

www.gnu.org/software/nano-archimedes

www.gnu.org/software/gneuralnetwork

http://en.wikipedia.org/wiki/GNU_Archimedes