Machine learning in trading: theory, models, practice and algo-trading - page 2558

New comment
 
mytarmailS #:

why?

you just need to look for not one correct cycle (sinusoid) which does not exist

but several simple cycles (sinusoids) the sum of which will create our complex non-sinusoidal cycle (complex from the word add)

Well, it's a variation of that, sort of, in any variation
 
mytarmailS #:

Alexei, do you have a good understanding of hmm?

I understand the essence more or less, but did not go into detail in the practice, because I do not consider them suitable for prices.

 
elibrarius #:

Why do we need to detrend? It's like we need to look for trends. It's more like de-trending.)

In order not to miss the moment when noise becomes the opposite trend.)

 
Perhaps we should go back to simple definitions generally accepted.
Decomposition into trend, seasonality, cycles, and noise. You can try to predict any of the above, with varying degrees of success.
Stationarity - the constancy of the average and dispersion, it is not observed in the market.
Regularity - the presence of repeatability, some analogue of 2D cycles or persistence, a certain level of signal. It is kind of like an opportunity for speculation. It differs from SB for the better.
 
Aleksey Nikolayev #:

I more or less understand the essence, but did not go into detail in the practice, because I do not think they are suitable for prices.

I want to teach SMM, but in an unusual way, through a fitness function, genetic or other...

I want to do state transition matrices myself... I have a package, these matrices, but what and where to change, I do not really understand, can you help with this?

 
Maxim Dmitrievsky #:
Regularity - the presence of repetition, some analogue of 2D cycles, or persistence, a certain level of signal. Sort of like carrying opportunities for speculation. It differs from SB.

Why should regularity not be called a pattern?

Because there's already a definition of regularity and it doesn't coincide with yours...

the first law of logic is violated)

 
mytarmailS #:

Why can't regularity be called - a pattern?

But there is already a definition of "regularity" and it's not the same as yours.
It's not mine, it's from some article. That many people confuse stationarity and regularity, giving stationarity some sort of predictive ability that a non-stationary series does not have. But stationarity has no effect on predictive ability, other than returning to the mean. But this is if the original series is such, when it is differentiated, it is no longer the original series.

Regular - repeating in time and space. A normal definition. Or carrying in itself some sort of ordered information.
 
Maxim Dmitrievsky #:
It's not mine, it's from some article.

Some pseudoscientific article since the author violates the first law of logic.

You can't call a regularity a regularity, even though the concept of regularity is already occupied...

What a fool...

Maxim Dmitrievsky #:
But regularity does not affect the predictive ability

Predictive ability of what?

if the row is stationary, of course it does.

 
mytarmailS #:

It's like a pseudoscientific article since the author violates the first law of logic.

You can not call a regularity - regularity, even though the concept of regularity is already occupied...

What kind of a moron...

Predictive powers of what?

If it's a series that's regular, of course it affects...

A stationary series can be regular and singular. Perhaps this will help in the search :)
 
Maxim Dmitrievsky #:
Stationary series can be regular and singular. Perhaps this will help in the search :)

In search of the road to the nuthouse? :)

1...255125522553255425552556255725582559256025612562256325642565...3574
New comment