Machine learning in trading: theory, models, practice and algo-trading - page 3025

 
Aleksey Vyazmikin #:

Good - the important thing is to get started! :)

I couldn't find any ready-made solutions, I have to parse the tree structure. It'll take a bit longer.
 

MO for context, mashka for trading )))

Trading is all about risk/profit, not about akurasi... IMHO


a, MO is just a tool to describe your understanding of the market, your model ...

If there is no model and you think that MO is AI and it is all by itself, then you are in trouble for a long time....

 
mytarmailS mashka for trade )))

Trading is all about risk/profit, not akurasi... IMHO


a, MO is just a tool to describe your understanding of the market, your model....

If there is no model and you think that MO is AI and it is all by itself, then you are in trouble for a long time....

The yellow one is adaptive? It's in the centre, it's horizontal.

 
Ivan Butko #:

The yellow one is adaptive? It's the one in the centre that goes horizontal.

No, it's just a normal one.

 
Maxim Dmitrievsky #:
That's 100 years.
From what date to what date of the oos? I was thinking a couple of months.
 
Maxim Dmitrievsky #:
I haven't found any ready-made solutions, I have to parse the tree structure. It will take a bit longer.

Yes, it's the same for me in R - it saves the tree structure in its own weird way. Then I have a separate parser to pull out leaves.

 
mytarmailS #:

And what library do you use that you immediately get inequalities as leaf rules?

 
Aleksey Vyazmikin #:

Yes, that's how I have it in R too - it saves the tree structure in its own weird way. Then I have a separate parser to pull out leaves.

Something like this, you can sort by maximum importance in the model, by probability of belonging to a class and by frequency of use.

That's enough for today.


 
Maxim Dmitrievsky #:

Something like this would be, you can sort by maximum importance in the model, by probability of class membership and by frequency of use

That's enough for today.


That's effective!

Do you code predictors into numbers from the master sample?

 
Aleksey Vyazmikin #:

It's effective!

Do you code predictors into numbers from the master sample?

any names