Machine learning in trading: theory, models, practice and algo-trading - page 2681

[mytarmailS]

Renat Fatkhullin #:

Make an api for the tester so that everyone can do what they want with any language they are used to, so that they can use the libraries they like....

That's what people want, not an outdated copy of some library on MO, which you have to learn again, which will always be 10 steps behind the best libraries....

[Serqey Nikitin]

Aleksey Nikolayev grail (or proof of its impossibility)

Any mathematical formulas are always derived from some assumptions, which are not necessarily always and everywhere fulfilled.

Wrong conclusions...

Mathematical formulae are ALWAYS fulfilled!

And what follows from this...?

Only that a profitable strategy does not rely on maths alone....

[Renat Fatkhullin]

mytarmailS #:
Make an api for the tester so that everyone can do what they want from any language they are used to, so they can use the libraries they like...

That's what people want, not an outdated copy of some library on MO, which you have to learn again, which will always be 10 steps behind the best libraries....

In short: it is not even theoretically possible. That's why there are so many retarded crutches around (not us) trying to combine diverse things.

We make complete, resource-saving and high-speed solutions. To have one(!) protected *.ex5 file that can be distributed and sold.

What is important: we do a couple of orders of magnitude more complex work implementing ML/matrices/mathematics in MQL5. Because the testing/implementation processes are based on speed, speed and speed again. Without it, there is no possibility to make trillion calculations, delving into the details of deep modelling.

Native maths gets rid of the burden of external libraries and allows you to perform calculations as fast as possible. And if possible, using GPU in a completely transparent way.

[Maxim Dmitrievsky]

Renat Fatkhullin #:

Posted documentation on the new features of the built-in matrices and vectors:

• Matrixand Vector Methods
  
• Matrix and vector types
• Initialisation
• Manipulations
• Operations
• Derivations
• Transformations
• Statistics
• Characteristics
• Solutions
• Machine learning

Please go through the descriptions of the functionality.

If you can give us your comments and ideas, we will be very grateful.

It turns out that layers can already be built through matrices and vectors and calculate the gradient on them, we just need to add a couple of optimisers like Adam and we can try to write a neural network.

And I haven't understood yet how the weights of the NS will be stored

[mytarmailS]

Renat Fatkhullin #:

We make one-piece, resource-saving and high-speed solutions. To end up with a single(!) protected *.ex5 file that can be distributed and sold.

Thanks for the detailed answer...

But do people need one protected *.ex5 file? You need it for your services, but not the people....

People need convenience, less change, etc...

People are ready to sacrifice speed, security in favour of convenience.

[Sergey Gridnev]

mytarmailS #:

People want convenience, less change, etc....

People are willing to sacrifice speed, safety for convenience.

That's exactly right.

[Maxim Dmitrievsky]

mytarmailS #:
Thanks for the detailed answer...

But do the people need one, protected *.ech5? You need it for your services, but not for the people....

People need convenience, less change, etc...

People are ready to sacrifice speed and security for convenience.

Write your own tester in 15 minutes, it's inconvenient to test in different applications and there's no point.

[mytarmailS]

Maxim Dmitrievsky #:
Write your own tester in 15 minutes, it is inconvenient to test in different applications and there is no point

Yes, I have, and not just one...

a simple tester can be written, but something serious with multiple orders, multi-instruments, stops, takes, etc.... it's hard to write normally, and the main thing is to work fast.

[Serqey Nikitin]

Sergey Gridnev #:
That's right.

Limited perception of reality!

If you consider TRADING as a BUSINESS, then the issue of protecting the business and its secrets comes to the forefront....

And there is nothing to argue about: defence of one's original methods of making a Profit - above all!

[mytarmailS]

Serqey Nikitin #:

If you think of TRADING as a BUSINESS.....

if you think of grandma as grandpa.

You count how many people are engaged in research, and how many people have a real TC that must be protected by rights, make up the ratio, God willing 1000 to 1.

Then you wouldn't write such nonsense