experchange > ada

Daniel (07-10-19, 09:25 AM)
Does anybody knows pure Ada libraries for AI?

Specially, I'm interested in Decission Trees, but I can't find anything on internet.

I case of a negative answer, does anybody knows a good CPU perforamnce AI C/C++ Library working good binded to Ada code?
J-P. Rosen (07-10-19, 09:39 AM)
Le 10/07/2019 à 09:25, Daniel a écrit :
> Does anybody knows pure Ada libraries for AI?
> Specially, I'm interested in Decission Trees, but I can't find anything on internet.
> I case of a negative answer, does anybody knows a good CPU perforamnce AI C/C++ Library working good binded to Ada code?

There is FannAda (), a binding to
the Fann neural network library. No idea what it's worth.
Daniel (07-10-19, 10:57 AM)
El miércoles, 10 de julio de 2019, 9:39:39 (UTC+2), J-P. Rosen escribió:
> Le 10/07/2019 à 09:25, Daniel a écrit :
> There is FannAda (), a binding to
> the Fann neural network library. No idea what it's worth.
> --
> J-P. Rosen
> Adalog
> 2 rue du Docteur Lombard, 92441 Issy-les-Moulineaux CEDEX
> Tel: +33 1 45 29 21 52, Fax: +33 1 45 29 25 00
>


Thank you for the answer. I was reading FANN but it looks like it only contains neural networks (perceptron, etc..). Im looking for Machine learning, specialy Decision trees and SVM.
Dmitry A. Kazakov (07-10-19, 12:52 PM)
On 2019-07-10 09:25, Daniel wrote:

> Does anybody knows pure Ada libraries for AI?
> Specially, I'm interested in Decission Trees, but I can't find anything on internet.




This includes decision trees both fuzzy and crisp. It is 100% Ada,
except the database persistence back ends.
Daniel (07-10-19, 01:30 PM)
El miércoles, 10 de julio de 2019, 12:52:47 (UTC+2), Dmitry A. Kazakovescribió:
> On 2019-07-10 09:25, Daniel wrote:
>
> This includes decision trees both fuzzy and crisp. It is 100% Ada,
> except the database persistence back ends.
> --
> Regards,
> Dmitry A. Kazakov
>


Thank you Dmitry. It looks really interesting. For now i've only tested ID3, and CART (C4.5) Trees for simple clasification problems. I've never used fuzzy logic and i don't know yet how it really works, but i will investigate all this work.

Generally, to translate every python example of Machine learning to ADA it's being painfull.

I don't know so much about fuzzy logic, but definitly it worths
Philip Munts (07-10-19, 04:44 PM)




When these were originally published, I posted copies (along with a similar column on the Taguchi Method) on the bulletin board at my workplace.

Not entirely Well Received...
Dmitry A. Kazakov (07-10-19, 05:08 PM)
On 2019-07-10 13:30, Daniel wrote:

> It looks really interesting. For now i've only tested ID3, and CART (C4.5) Trees for simple clasification problems. I've never used fuzzy logic and i don't know yet how it really works, but i will investigate all this work.


Do not worry about fuzzy logic.

Firstly, what is traditionally called "fuzzy logic" is not what is used
in the library[*].

Secondly, the fuzzy decision tree is a proper extension/continuation of
the crisp decision tree. If you have crisp data sets the result of
learning will be a crisp tree. No magic, no snake oil.
Jeffrey R. Carter (07-10-19, 06:13 PM)
On 7/10/19 9:25 AM, Daniel wrote:
> Does anybody knows pure Ada libraries for AI?


I guess you're not interested in neural networks, but there's an implementation
of REM NNs in the PragmAda Reusable components.

Björn Lundin (07-11-19, 09:42 AM)
On 2019-07-10 18:13, Jeffrey R. Carter wrote:
> On 7/10/19 9:25 AM, Daniel wrote:
>> Does anybody knows pure Ada libraries for AI?

> I guess you're not interested in neural networks, but there's an
> implementation of REM NNs in the PragmAda Reusable components.
>


REM NN?
That is not the same as recurrent neural network is it?

Recurrent neural networks would interest me.
I'm playing with it using python, but if there is an Ada alternative,
I'd be happier.

--
Jeffrey R. Carter (07-11-19, 06:49 PM)
On 7/11/19 9:42 AM, Björn Lundin wrote:
> On 2019-07-10 18:13, Jeffrey R. Carter wrote:
> REM NN?
> That is not the same as recurrent neural network is it?


No. It's NNs with the REM 2nd-order learning algorithm.

Similar Threads