Machine learning in trading: theory, models, practice and algo-trading - page 1736

New comment
 
Mihail Marchukajtes:

R do not know. For CSSA I found only that it is worse than SSA and overdraws a little.

 
Rorschach:

R do not know. The only thing I found about CSSA is that it is worse than SSA and overdraws a bit.

Well of course there will be differences. I understand that it contains a proximator with a crazy number of neurons, which can be flexible in any format. It is quite possible that the negative reviews were because people never managed to get exactly the perfect picture. I just had one just like in the first picture and I was stunned. I've never seen at the time and even now to have such a high degree of cointegration between the claws and the cycle and everything is as smooth as clockwork. I just understand that such events occur when the futures are over and the quote is purely technical. I guess I got to that point when I couldn't even imagine their existence. Moex wasn't around then, I think. Forex is a cruel joke :-( So much time wasted......
 
Maxim Dmitrievsky:

Let's go clustering... set up 3 clusters, i.e. partition the cloud of points (chips) and determine which cluster each belongs to

A simple two-dimensional example for a trace sabbet:

For the test:

We take these clusters on the trayn and non-test and construct a cumulative sum of the increments over all clusters.

On the trayne:

On the test:

Curious.

The question is which one to choose and how? Another wicked joke :-)
 
Maxim Dmitrievsky:

Lamba or Bentley ?))

 
Maxim Dmitrievsky:

Don't be stupid.

Fiches, ugh Max, you swore again. I'm going to ask what is the point of them? Is it a profit? Claws? Or what? What was the closest. I mean, they were just standing next to each other in this cloud of cordinates that had been gathered for training. So what are they standing next to each other in accordance with the trait, by the way, what is it, I keep forgetting to ask?
 
mytarmailS:

lamba or bentley ?))

go choose...

 
Maxim Dmitrievsky:

went to choose...

Why KNN?

he could have plugged in a dbscan or something.

 
mytarmailS:

Why KNN?

Could've been a dbscan or something.

decipher

 
Maxim Dmitrievsky:

decipher

well to the eye you can see that clustered the nearest neighbor knn

right?

 
mytarmailS:

you can see that you clustered the nearest neighbor knn

right?

k-means, the simplest

1...172917301731173217331734173517361737173817391740174117421743...3494
New comment
Reason: