Machine learning in trading: theory, models, practice and algo-trading - page 518

 
Yuriy Asaulenko:
If you do not feel sorry for the NS inputs, you can cram the delta between the maxima there. The delta can be preliminary passed through the sigmoid and subtracted by 0.5 to bring it to zero.

By the way, there is no sense to feed the derivatives of the moving average, you can simply feed the increments with other lags, they will describe the trends.

 
Maxim Dmitrievsky:

By the way, there is no point in feeding the derivatives of the muwings, you can simply feed the increments with other lags, they will describe the trends.

The derivatives show the direction of the trend. Derivatives from 2 MAs and the difference between them completely describe the state of the system. You yourself asked for the thread).

However, it is up to your business.)

 
Yuriy Asaulenko:

The derivatives show the direction of the trend. The derivatives of 2 MAs and the difference between them completely describe the state of the system.

However, it's up to the owner).


Yes, but I want to feed 50 inputs, for example, on each input shift the increment by 1 bar backwards. I.e. the model will just take into account changing of the trend component

The more increments with different periods the more they will describe all movements, I think 3 is enough

I.e. it will be 50*3 inputs in total... or 250*3... what a trifle :D

 
Maxim Dmitrievsky:

Yes, but I want to feed 50 inputs, for example, on each input shift the increment by 1 bar back. I.e. the model will just take into account change of trend component

I adhere to the opinion, that it is better to prepare something simple by yourself and apply the result to the National computer.

The NS itself becomes simpler, and its resources are freed up for more interesting tasks, than those solved by elementary logic. And the regular MAs are a nice trend indicator, and let the National Computer process its results, instead of inventing something.

 
Yuriy Asaulenko:

I am of the opinion that it is not necessary to load unnecessary things on NS, and what is done elementary, it is better to prepare yourself and already the result is submitted to the NS.

The NS itself becomes simpler, and its resources are freed up for more interesting tasks, than those solved by elementary logic. And the regular MAs are a nice trend indicator, and let the NS process its results, instead of inventing the unknown.


I'll try it both ways, then I'll show you the results

i think that these inputs and outputs correlate well when i have homogeneous data

 
Maxim Dmitrievsky:

here's another thing that inputs and outputs correlate well when you feed homogeneous data

I don't know what that means?

On an instrument, everything correlates with everything. And it's inevitable, because everything is obtained by transforms from the same time series, from the same data.

 
Yuriy Asaulenko:

I don't understand what this is about?

Everything correlates with everything on the instrument. And this is inevitable, since everything is obtained by transformations from the same time series, from the same data.


Well essentially yes, the question is probably the strength of the correlation

 

Guys, Don't forget the causal relationship of the data. Expectation+Volume+Delta+OI. And then any TC will be true. Very much I ask to pay attention to this link. I would be extremely grateful :-)

FXlab лаборатория форекс | ВКонтакте
FXlab лаборатория форекс | ВКонтакте
  • vk.com
Показать все 6 комментариев Turbo, тут есть чёткая грань, мм - всего лишь часть ТС, система подразумевает точное описание условий для входа, в случае гармоники - точка Д паттерна, в случае тс на машках - пересечение ма на разных периодах, в случае уровневой тс - отбои от уровней, в случае прайсэкшн - комбинации свечей, а так же описывает тс...
 
Mihail Marchukajtes:

Guys, Don't forget the causal relationship of the data. Expectation+Volume+Delta+OI. And then any TC will be true. Very much I ask to pay attention to this link. I would be very grateful :-)

This is nonsense. It has nothing to do with anything. It is definitely off-topic).
 
Yuriy Asaulenko:
This is nonsense. It has nothing to do with anything. (It is off-topic, for sure.)

What topic???? Friend... Who are you?

Just maybe you're new here and don't know me. I'm kind of an AI guy in general,too..... What are you... You're from around here. :-)

Although I'm kind in principle, and everything you do here will work ONLY when the data will be the reason for the price. Then any TS will work fine, even if it's 1 bar or 15 bars ahead of the forecast (of course 15 bars will be worse than 1, but not the point). Not the point... The point... RTS index which has OI. As the meaning of volume.... And the problem is SOLVED. ANYTHING, even a prediction, even a classification.

And what did you want to say with your phrase, dear.....