Examples of Use

Feb 25, 2011 at 10:51 AM

I like this project, the idea to have something of a baseline ml library in c# is great. On the initial page you provide some examples of quickly setting up a classification and clustering solution using NNs, Trees and K-means. I wonder, once you trained the model what way do you recommend to store it, obviously re-training the models is costly and keeping them in memory might also not be the best solution. btw. thumbs up, the GA also looks cool, will play around with the code over the comming days :-D

Martin

Coordinator
Feb 25, 2011 at 4:47 PM
Edited Feb 25, 2011 at 4:54 PM

Martin:

    Great question! It turns out that these models can be saved to a file (xml). Whenever you wish to reuse a model from before, instead of generating the model you simply load it up again. A good example can be found in the Test_Decision_Tree_Save test method as well as Test_Decision_Tree_Load

Hope this helps!

-Seth

--EDIT

Incidentally, thank Azret Botash (of DevExpress) for the GA implementation. I've been dying to incorporate it into the ML typing system).

Feb 25, 2011 at 5:17 PM

ah, thank you Seth, that's exactly the sort of thing I was after. I am looking forward playing around with the library a bit.

Thanks,

Martin