Machine learning in trading: theory, models, practice and algo-trading - page 2488

[JeeyCi]

Mihail Marchukajtes #:

In fact, you have to understand the whole philosophy of the field of IR, and without knowing a lot of information in the field it is very difficult to come to the philosophy of the question, and without it to come to the right conclusions

Yes,

In economic development, new lines of business are born out of chaotic behavior. Some of them turn out to be dead ends, but another part realizes new directions of progress. The artificial elimination of uncertainty leads to stagnation and degradation of the system, because the birth of progressive directions of development is excluded. These postulates apply to all complex systems, including economic and social systems.

In general, the main philosophy is not to eliminate uncertainty, but to exploit it...

And the "dead-end" is clearly creeping up the thread in the form of recent posts that do not know how to deal with uncertainty... MO is clearly not for them ()... as they can't prove to themselves...

The possibility of self-organization in a dynamic system.

[[Deleted]]

JeeyCi #:

Yes,

in general, the main philosophy is not to eliminate uncertainty, but to exploit it...

and a "dead end" has clearly crept up the thread in the form of recent posts by those who can't work with uncertainty... MO is clearly not for them ()... because they can't prove to themselves -

Off-topic question: What ward are you from?

[mytarmailS]

Vladimir Baskakov #:
Off-topic question: What room are you from?

ahahahaha, right on topic

[Aleksey Nikolayev]

Careful with the lady, she has daisyot)

[[Deleted]]

Aleksey Nikolayev #:
Be careful with the lady, she has daisyot)

Reminds me of Latynina on the echo. My head starts to hurt in half an hour.

[JeeyCi]

Mihail Marchukajtes #:
I, for example, organized the process of obtaining a model for 2 hours, and a set of programs is selected and I do the same steps over and over again. ONE AND THE FOLLOWING !!!! There is no need to bounce between optimization methods, network topology choices, or reshuffling of input data. The model creation algorithm is rigid and cannot be changed,

yes, discrepancies in the results of chaos ordering are BETTER, since many factors affect the system (and not all of them we can take into account)... that's why you can't predict for a long time... and have to be retrained periodically

The alternatives have a system of priorities, transition of the system into chaos is possible only in the case of unclear priorities. This corresponds to the close potential(c78) pits in the figure. If the priorities are clearly expressed and different, which corresponds to different in depth potential pits, then chaos is unlikely, the system will move into a deterministic state with the choice of a clearly expressed better alternative. It is obvious that the alternatives corresponding to the points in the figure have the following characteristics: 14 - best of all, 10 - no way, 7 - maybe, 3 - as a last resort, 4-5 "And we don't care..." (с).

The second analogy is that any complex question has at least two correct answers, corresponding to potential holes in the prototype model. And there is a long debate between these two answers. Should we allow gun sales to citizens as in the U.S.? Should we raise taxes? Should we make downtown entry a toll? etc. The talk-show discussions are reminiscent of the chaotic behavior of a balloon considered in a simple model.

A third analogy is the chaotic state of a complex system when choosing between two unequal alternatives: "To be or not to be?" (c), "Am I a creature shivering or do I have a right?" (с).

It's not hard to classify the posting trolls... - Philosophy of MO is always helpful

p.s.

what role do matrix calculus play in your modeling? do you use and from which library (what are the tools)?

[Mihail Marchukajtes]

mytarmailS #:

Yeah... but of course you can't test it ))))

And you've been using a ready-made rattle of Reshetov for several years, from it you heard the word polynom, neuron, mgua etc. for the first time, but you claim to be a master of neural network training with 20 years of experience)) and at the same time one year ago asked to be taught to program in R or Python)))

And the fools believe it, master, what the fuck, but what to do))

Surprisingly BUT you know who I am!!! Yes I am a weak programmer, but I have introduced a number of key organizational moments in the optimizer Reshetov, Yes I asked for help with the code in R and I helped but only in the organization with the library work, all the rest code dopislavl himself. And yes I now need to learn the macros Excel to start writing history on a smile. But I've been using optimizer Reshetov for 8-9 years, and before that I used Neuroshel and I was training, training and training models. And how many models did you get in your life and were you able to evaluate them? Well... How many?????

[JeeyCi]

Mihail Marchukajtes #:
with code in R... on Neuroshell and trained the same

thanks for the answer

[Mihail Marchukajtes]

JeeyCi #:

Yes, there will be differences in the results of chaos ordering, because many factors affect the system (and not all of them can be taken into account)... that's why you can't predict for a long time... and we have to periodically retrain

It's not hard to classify posters-trolls... - Philosophy of MO is always helpful

p.s.

what role do matrix calculus play in your modelling? do you use and from which library (what are the tools)?

Read Reshetov's description and much will become clear on the question of optimization.

[mytarmailS]

Mihail Marchukajtes #:
Reshetov for 8-9 years, and before that he sat tightly on Neuroshell and also trained, trained and trained models.

And that makes you a neural network expert with 20 years experience? It's so funny it's sad)

JeeyCi #:

It's not hard to classify post trolls... - Philosophy of MO is always helpful