Machine learning in trading: theory, models, practice and algo-trading - page 2784

New comment
 
Aleksey Vyazmikin #:

Need a ";" delimited csv file with columns: "Data", "Target".

In "Target" write the target - to encode the presence and position of the arrow.

Do you have arrows on the bar opening or what?

The arrows appear at the moment of the event.

I'll try to figure it out. If I have enough brains). I have not dealt with these questions. Haven't needed to. I can't do it fast enough. I'm busy with family matters.

 
Uladzimir Izerski #:

Arrows appear at the moment an event occurs.

I'll try to figure it out. If I have enough brains). I haven't dealt with these issues. Haven't needed to. I can't do it fast enough yet. I'm busy with family stuff.

If the arrows in the middle of the bar depend arbitrarily on the tick, it will be a bit more complicated.

You can also give me a bin file, but describe its structure.

 
sibirqk #:

Twenty years ago, when I was developing C, I also stalled at this point. Indexing from scratch, and the cycle is one unit less. But then MKL is easy to master - the same C.

If you start from scratch, it's probably easier, starting from scratch. And I have already learnt another language, it's hard to think differently....
 
Aleksey Vyazmikin #:

If the arrows in the middle of the bar, arbitrarily depend on the tick, then it will certainly be a little more complicated.

You can also give bin file, but describe its structure.

Date - bar number, and target 0 or 1 buy sell?

 
Valeriy Yastremskiy #:

The date is the bar number and the target is 0 or 1 buy sell?

The date is better, to control synchronisation.

Target "0" or "1" - arrow appeared or not, or at once -1 - arrow down, 1- arrow up, 0 - no arrow.

Although, I admit that the indicator there is a condition that the arrow is drawn only once per knee, even if there is a reason to draw it repeatedly, then it will be difficult.

 

Which of the following is the best tool to use to generate data?

https://scikit-learn.org/stable/auto_examples/cluster/plot_cluster_comparison.html


Comparing different clustering algorithms on toy datasets
Comparing different clustering algorithms on toy datasets
  • scikit-learn.org
This example shows characteristics of different clustering algorithms on datasets that are “interesting” but still in 2D. With the exception of the last dataset, the parameters of each of these dat...
 
Evgeni Gavrilovi #:

Which of the following is the best tool to use to generate data?

https://scikit-learn.org/stable/auto_examples/cluster/plot_cluster_comparison.html


One should take that clustering which is relevant to the target variable.

Is there such a clustering here?

 
Evgeni Gavrilovi #:

Which of the following is the best tool to use to generate data?

https://scikit-learn.org/stable/auto_examples/cluster/plot_cluster_comparison.html


Clustering does not generate data
 
Oh, I see, so there's only one GMM? No others?
[Deleted]  
Evgeni Gavrilovi #:
Oh, I see, so there's only one GMM? No others?

TGAN and other GANs, autoencoders, kernel density estimation, copulas

I haven't tried TGAN, the others are worse than GMM.

Maybe there are new time series GANs available.

*GMM does not converge well on large samples, you need to use not very large samples.
1...277727782779278027812782278327842785278627872788278927902791...3742
New comment