"New Neural" is an Open Source neural network engine project for the MetaTrader 5 platform. - page 91
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
Open-source? I haven't seen any yet...
And the bazaar black boxes, no sorry, that's for the completely out of their minds, I would not be surprised if 90% of them do not have any neural network at all, but the usual indicators, with GUI visualization of the "learning process", they are bazaars, everything for them is marketing, nothing sacred, after death all hell awaits them.
I agree, the main thing is the process, especially when you already have at least a couple of million dollars in the offshore
I've already posted open source codes of neural network and genetics, this minimal set is enough to write something working ...
here's a thought, maybe I should create a service such as "I'll find and put in your mouth, good quality, fast, inexpensive!", what do you think? - I think it's a good idea.
I've already posted open source codes for neuronics and genetics, this minimal set is enough to make something work...
So I thought, maybe I could create a service like "I'll find and put in my mouth, high quality, fast, inexpensive!", what do you think? - I think it's a good idea.
Before you put it in your mouth, you have to chew it up))
The idea is to chew it up before you put it in your mouth.)
Thank you for the business idea, the chewing can be offered as an additional service at separate rates, for pensioners at an increased rate of 90%.
I've already posted open source codes for neuronics and genetics, this minimal set is enough to make something work...
So I thought, maybe I could create a service like "I'll find and put in my mouth, high quality, fast, inexpensive!", what do you think? - I think it's a good idea.
Not found alas, there is a broken link in your profile to a genalgorithm, about NS nothing, if you do not difficult to give a link to unpack yet do not need to try to figure out myself, and if not, I'll ask you for the money.
When I started to readMachine Learning in Trading: Theory and Practice, I tried to read it but it was confusing, clearly there is a mental decline compared to your topic 8 years ago, apparently a new generation has appeared, that reads and calculates via social networks.
I have not found it alas, there is a broken link in your profile on the genalgorithm, there is nothing about NS, if it is not difficult to give a link, I will try to understand it myself, and if not, I will ask you for money.
When I tried to readMachine Learning in Trading: Theory and Practice, I got confused with the analysis, clearly there is a mental decline compared to your topic 8 years ago, apparently a new generation has appeared, that reads and calculates via social networks.
Thanks, corrected the link, there was a dot at the end of the address.
https://www.mql5.com/ru/forum/4956#comment_108594 - this is a neuron, right on the first page of this thread, it's as simple as a stool, make it into a class and you'll be happy.
And you think wow what geniuses are gathered here, but as a result "the project is stalled" and there is only this and it seems that most of them are just petty liars.
Read the whole thread.
About how the metaquotes jumped off the project curation in any form.
About repositories and projects that didn't get done properly until a few years later, and the MC's vision of projects is, er, peculiar even now.
About how TC pulled the blanket on his moronic genetics.
And so on and so forth.
And you will realise that the project could only succeed if someone took over and did pretty much everything on their own.
And you read the whole thread.
About how the metaquotes jumped off the project curation in any form.
About repositories and projects that were only made normal after a few years, and the MC's vision of projects is, er, peculiar even now.
About how TC pulled the blanket on his moronic genetics.
Etc etc etc.
And you will realise that the project could only succeed if someone took over and did pretty much everything on their own.
Thanks, corrected the link, there was a dot at the end of the address.
https://www.mql5.com/ru/forum/4956#comment_108594 - this is a neuron, right on the first page of this thread, it's as simple as a stool, make it into a class and you'll be happy.
Great! But it lacks training, reverse propagation.
And you should read the whole branch.
About how metaquotes jumped from project curation in any form.
About repositories and projects that were made normal only after a few years, and the MC's vision of projects is, er, peculiar even now.
About how TC pulled the blanket on his moronic genetics.
Etc etc etc.
And you will realise that the project could only work if someone took it all on their own.
On the quiet reading. IMHO of course I should not make immediate "Napoleonic plans" to make a universal toolkit, which includes most popular types of neural networks, in fact it would be more interesting to use something like all known NeuroPro (with one MLP) but open-source and without sophisticated GUI, for C++ training, which produces a dump with weights in a file that is read from mql5 exchange robot, as a matrix of parameters. That is, the combination of C + + MO and mql - execution, of course, all open source on githab, firstly to quickly, so that there was something, then to modify extensibility and then to add new modules and new project participants.
Great! But there's not enough training, backward propagation.
I'm reading up on siliconics. IMHO of course it would be better not to make at once "Napoleonic plans" to make a universal toolkit including most popular types of neural networks, in fact it would be more interesting to see something like all known NeuroPro (with one MLP) but open-source and without sophisticated GUI, for C++ training, which outputs dump with weights into file readable from mql5 exchange robot as parameters matrix. That is, the combination of C + + MO and mql - execution, of course, all open source on githab, firstly to quickly create something that would be at least something, then modify for extensibility and then add as new modules and new project participants will appear.
First it would be worth developing a general approach to the matter. A universal one, allowing for expansion with different types of networks and different learning methods.
Generally, it is not the first attempt to make something large-scale with neural networks, but it's still there. Even if we get involved now and propose options for a common approach, it will not work; people here are incapable of collaborative action.
Great! But there's not enough training, backward propagation.
I'm reading up on siliconics. IMHO of course we should not make "Napoleonic plans" to make a universal toolkit including most popular types of neural networks, in fact it would be more interesting to see something like well-known NeuroPro (with one MLP), but open-source and without sophisticated GUI, for C++ training, which outputs dump with weights into file, which is read from mql5 exchange robot, as a matrix of parameters. That is, the combination of C + + MO and mql - execution, of course, all open source on githab, to start with a rush, so that there was at least something, then modify for extensibility and then add as new modules and new project participants will appear.
A search on the site "unborn"(https://www.mql5.com/ru/search#!keyword=%D0%BD%D0%B5%D0%B9%D1%80%D0%D0BE%D0%BD%D0BD%D0%D0%B0%D1%8F) gives a lot of links with articles and sources in kodobase - MLP, BP, SOM, RBF, PNN and many more - all with sources in MQL. You may link and use all of them in any combination without fancy GUI.
And you don't need any C++. This dependence complicates the product and gives little in return which cannot be done in MQL (with comparable quality) in the context of the above mentioned consumer neural network cofigurations.