Machine learning in trading: theory, models, practice and algo-trading - page 874

[Aleksey Vyazmikin]

Yuriy Asaulenko:

How do we know if they fit? Do you work with NS (neural networks)?

And if not, what do you do? And that would be closer to the topic of MO).

I in the process of preparation, so to speak, I want to target normalize and give them an estimate - very bad - complete crap, average - ok, very good - as garbage (for rarity).

[Yuriy Asaulenko]

Aleksey Vyazmikin:

I'm in the process of preparing, so to speak, I want to normalize the targets and give them a grade - very bad - complete crap, average - okay, very good - the same nonsense (because rare).

So what do you do? What is your field?

[Maxim Dmitrievsky]

Aleksey Vyazmikin:

In which state should I look for them and check if they fit? I don't have the knowledge - stupid, so I'm asking the smart ones.

http://www.flong.com/texts/code/shapers_exp/

You need a Logistic Sigmoid (at the very bottom of the page)

substitute your data normalized from 0 to 1 into the function.

[Aleksey Vyazmikin]

Yuriy Asaulenko:

So what do you do? Area?

I am fantasizing - the night before last I invented a method for selecting (evaluating) predictors (for me it is the calculation of filters), and I want to try it out.

In addition to natural indicators, I want to use some evaluation coefficients - what will be better - do not know yet. May be nothing and I am delirious, but I will check...

Maxim Dmitrievsky:

http://www.flong.com/texts/code/shapers_exp/

You need a Logistic Sigmoid (at the bottom of the page)

In this function, you put your data normalized from 0 to 1.

Thanks - I'll look into it!

[Maxim Dmitrievsky]

[[Deleted]]

Maxim Dmitrievsky:

What do you do with it?

[Maxim Dmitrievsky]

Oleg avtomat:

And what to do next with it?

I don't know, you could file an NS

need to do for the ticks yet

[[Deleted]]

Maxim Dmitrievsky:

I don't know, you can file it in the NS.

we need to do for the ticks.

I see...

Well, what was the original purpose / idea?

[Maxim Dmitrievsky]

Oleg avtomat:

I see...

Well, what was the original goal/concept?

Well, the sifting of the flows that Alesandre wrote about

to achieve a Poisson distribution of returns or something

[[Deleted]]

Maxim Dmitrievsky:

well, sifting the flows that Alesander wrote about

achieve a Poisson distribution of returns or something like that

Actually, that's a load of crap.

To understand it, all you have to do is ask the question "Why?" and try to give a reasonable answer to it.