A purely theoretical question for mathematicians. With the possibility of moving to the practical plane. - page 5

 
Nikolai Semko:

No, you didn't write that. It's just that, like Yusuf, you are looking for something that cannot be found where it does not exist.

And in doing so, you bring this nonsense to everyone's attention, creating a flood in this intellectual forum.

Yusuf's thread is an intellectual flubber, as is yours.

Nicholas and other opponents, before you accuse, you too show your result on the TC topic:



For calculation formulas, seehttps://www.mql5.com/ru/articles/250
 
Сергей Таболин:

Dear comrades of higher mathematics (or at least above average :)), I am sure that mathematics can describe almost everything. And on the basis of this confidence I want to ask:

  1. Isn't it possible, but how do you calculate patterns in a series of values? For example +165, -240, +18, -378, +681, -115....
  2. How to calculate the same pattern in consolidation with another row (possibly more than one)?

I will indicate the practical application later, after answering the questions posed.

Thanks.

I've dealt with a similar question and even asked on the forum, looking for information. But the result is not very good. All I've come to is that you can make an equation and solve it. But it is very difficult to say something about the form of the equation. That is, yes you can solve the problem, but you don't know the equation, and how to find it is a mystery.
The only option I've come up with is by brute force. If you take your series of numbers, all numbers except the last one are put into the equation, define coefficients, degrees, maybe on the contrary some number is a degree and put one last number after equal... but this is unreal, the number of combinations is simply absurd.
That's why people invented neural networks. It's essentially a ready-made equation of some complexity, where you substitute values and look for weighting coefficients.
If you find a way, let me know, it would be interesting to see.
 
Maxim Romanov:
I have dealt with a similar issue and even asked on the forum, looking for information. But the result is not very good. All I've come to is that you can make an equation and solve it. But it is very difficult to say something about the form of the equation. That is, yes you can solve this problem, but you don't know the equation, and how to find it is a mystery.

Eureqa is good at matching formulas to numbers, I like it. For some things it worked better than with neural networks.

 
Yousufkhodja Sultonov:

Nikolai and other opponents, before you accuse me in any unsubstantiated way, you too should show your result on the TC topic:

No, Yusuf - I am not accusing you of anything, nor do I feel I am your opponent.

My youngest daughter is 2 years old and she is a disaster right now. She paints everything (!!!) with crayons - walls, furniture, carpet, clothes, etc etc. etc.

Can you blame a child for being too young?

And when I say to her, "Olivia, aye-aye-aye!"- I'm not being her opponent, I'm just admonishing her because I shouldn't leave it at that, even though I realise she's just learning about the world and learning.

And with your pictures you remind me of my eldest daughter (age 6) who is always bringing me her drawings and waiting for me to praise her. And I of course tell her - wow!!!

:)) - no offence

 
Yousufkhodja Sultonov:

Understand, finally, once and for all, only I have been able to extend the MOC to the domain of nonlinear dependencies, absorbing, among other things, the Gaussian MOC for the linear domain.

This is a poorly framed delusion of grandeur.

Your quote:
"You may ask, why from 4? The fact is, I, so far, know how to solve this equation up to 4 variables."

Sorry, but I was able to solve SLAEs with any number of variables back in school. (with the help of a computer, of course, in the form of a small function, written by my own hand according to my own algorithm).

 
Yousufkhodja Sultonov:

Understand, finally once and for all, I am the only one who has managed to extend the MOC to the domain of non-linear dependencies, absorbing, among other things, the Gaussian MOC for the linear domain.

Congratulations on the gold ranking!!!

I had a real funny story about the number 666.
A good friend of mine once called me (he was a military commissar in one of the districts of St. Petersburg at the time) to his house to see what was wrong with his computer, because he knew that I sort of know my way around computers.
I came in, and on the system unit there was a new number plate with the number 666 on it.
I said, "Wow, those are some new plates."

And he said to me: "Can you imagine, I bought a car and asked a cop I knew to pick up a nice number. And I got it, damn it!"

When I found out what was wrong with the computer, I found out that it was infected with the DOOM666 virus. And the trouble happened the day they put those bad numbers on it.
Mystery or coincidence?! But it's an absolutely true story.

 
Nikolai Semko:

Congratulations on your gold rating!!!

Thank you.

 
Yousufkhodja Sultonov:

Understand, finally, once and for all, only I have managed to extend the MOC to the domain of nonlinear dependences, absorbing, among others, the Gaussian MOC for the linear domain.

We got that, we didn't get it - then why doesn't it converge?))) I do, although I didn't invent or propagate anything)))


 
Vizard_:

We got that, we didn't - then why doesn't it add up?))) I do, although I didn't invent or distribute anything))))


Give me the formula for calculating it, otherwise it's a fit. Topikstarer will be asked to add some more figures, and the formula you provide should also nicely describe them. I don't believe in such a perfect match between actual and calculated, arbitrarily given, values. So far, I know that only the best estimate of the calculated values is possible, ensuring that not every value, but their sums, coincide with each other.

 
Yousufkhodja Sultonov:

Provide, the formula for the calculation, otherwise, a fitting. Topikstarer will be asked to add a few more figures, and the formula you provide should also describe them nicely. I don't believe in such a perfect match between actual and calculated, arbitrarily given, values. So far, I know that only the best estimate of the calculated values is possible, ensuring that not every value, but their sums, coincide with each other.

Docent, you can't be serious))) Of course the fit. Of course, by disguising the "truth" in the sums, you arrive
in an illusion, and the apparent ease of predicting MAK is misguided. It's unclear how you're looking for anything at all, if you can't even get a hint
from the cartoon. The next value of vr with a different sign is an obvious thing. You should take abs(x) instead of x, and the prediction
"zz ratios" will eventually boil down to volatility prediction(with variations). And the right answer to all questions has already been sounded...


Reason: