You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Can I get the link?
If you like it, please send me the links.
The software is paid.
http://www.neuroshell.com/
You can find cured versions on the web for reference. There is a big branch on the Alpari forum about this software.
http://forum.alpari.ru/showthread.php?t=29694
And what, then, prevents it from being realized.
Or are you one of those "critics". Well, who do not like everything, but do not do anything themselves
It is forbidden: to criticize any, even the most delusional idea, to flub. The level of competence does not matter. If you are interested and have thoughts - speak up!
Welcome: offer new thoughts, develop existing ones.
Cool me down sometimes too :)
Not allowed, it hurts your nerves, immediately grabbing a gun in an attempt to commit harakiri :o)
and we make every person count.
So maybe we should take these three paradigms as a basis too?
No, it hurts your nerves, you immediately grab a gun in an attempt to commit harakiri :o)
Well you can't put the learning algorithm at the top of the list.
And I have nothing against genetics anymore. It's an independent thing, almost like imaging, and requires minimal add-ons from the network to be bolted on.
Until the versioned repository features are enabled in the editor, you will have to use external clients like TortoiseSVN. We will try to add the basic management functionality to MetaEditor as soon as possible.
In addition, it is necessary to lay on effective means of importing/exporting data from known neuro-packages. This will allow to attract those who are already strongly attached to these solutions. In addition, the preliminary development must be done in convenient specialized programs, and then unload the ready/input/preliminary data into the trading system.
Well, you can't put the learning algorithm at the top of the list.
...
I agree, we should be dancing on the grid algorithm,
But I would like the project to evolve towards maximum adaptability to multiple algorithms.
So that we don't have to rewrite the grid algorithm again for every new training algorithm.
1) From the ideas I would like to cancel the possibility to construct a network from the wizard on the basis of the engine.
2) Also, it's necessary to lay on effective means to import/export data from known neuro-packages. This will allow to attract those who are already strongly attached to these solutions. In addition, the preliminary development must be done in a convenient specialized programs, and then unload the ready / raw / preliminary data into a trading system.
1) That's the easiest thing to do, to make a graphical network creation interface that will generate code for an immutable network, like the wizard now generates an EA, although I planned to go further.
2) You have to write a bunch of utilities for recoding nets trained in other packages, there are many packages and even more data storage formats. Different packages don't really care that their networks are easily portable to a competitor's package, hence the mess of formats.