Machine learning in trading: theory, models, practice and algo-trading - page 3306

 
Aleksey Vyazmikin #:

It's

Still don't understand, state the question directly if you can.
 
Maxim Dmitrievsky #:

A homeless algorithm is needed that digs through the rubbish

From dirt to princes, you could call this series of articles

For some reason I immediately thought of Renat Akhtyamov)
 
Aleksey Nikolayev #:
For some reason I immediately thought of Renat Akhtyamov))

:))

 
Aleksey Nikolayev #:
I still don't understand, formulate the question directly if possible.

How do you propose to prove the statement that"series close to the SBappear to have a period". I asked about the opposite way - if the series have a period, how it can be revealed.

 
mytarmailS #:
Thinking Fourier is all about periodicity is like thinking music is all about rap.

You're not giggling at radio amateurs, you're giggling at your illiteracy.
That's the thing, only those who think that Fourier is about periodicity of prices keep popping up) If anyone was looking for periodicity of volatility or persistence-antipersistence cycles, they would be the first to applaud.
 
Aleksey Vyazmikin #:

How do you propose to prove the statement that"series close to the SBappear to have a period". I asked about the opposite way - if the series have a period, how it can be revealed.

In econometrics, for example, there are many different statistical tests for seasonality.
 
Probably, we can simplify the understanding by imagining that not only prices, but also volatility are inputted
 
Aleksey Nikolayev #:
In econometrics, for example, there are many different statistical tests for seasonality.

So no one has thought of doing these tests?

 
Aleksey Nikolayev #:
That's the thing, only those who think that Fourier is about periodicity of prices keep popping up) If someone was looking for periodicity of volatility or cycles of change from persistence to antipersistence, they would be the first to applaud.

I did )

https://www.mql5.com/ru/forum/86386/page1056#comment_8678465

Машинное обучение в трейдинге: теория, модели, практика и алготорговля - Сейчас еще на рэндомных трансформациях через косинусы посмотрю
Машинное обучение в трейдинге: теория, модели, практика и алготорговля - Сейчас еще на рэндомных трансформациях через косинусы посмотрю
  • 2018.09.14
  • www.mql5.com
Пусть миклуха маклай портянку с gmdh скинет - будет не ожидаемо. Сейчас еще на рэндомных трансформациях через косинусы посмотрю. матожидание этого ВР в любой момент времени при любой выборке должно быть const
 
Aleksey Vyazmikin #:

So no one has thought to run these tests?

Why is it assumed that these tests have always existed? Most of all statistical tests were invented (and most importantly - justified) only in the 20th century, after the creation of the theorist. Fourier decomposition, on the other hand, was invented a hundred years earlier.