Machine learning in trading: theory, models, practice and algo-trading - page 3663

New comment
 
Maxim Dmitrievsky #:

I am not ready to fully describe my approach yet. Here is an example of dividing a dataset into groups, where observations with labels 0 and 1 will be maximally far away in the feature space, e.g. by Euclidean distance.

You can develop this topic and write something interesting based on it.

    data = filter_samples_by_clusters(pr, n_clusters=2, threshold_percentile=100, min_distance_percentile=90)

youdyshch, euros, education up to 21 years of age


 
Maxim Dmitrievsky #:

You know, euros, tuition up to 21.

I guess it's logical to post with the charts and the interpretation of them.

The way I see it is that to the left of the right vertical line is IN. To the right is OOS. Judging by the comment, the OOS is 2021-2024. Then why is it so short on the chart?

Anyway, some confusion is being posted.

 
fxsaber #:

I guess it's logical to post the graphs and the interpretation of them.

The way I see it, to the left of the right vertical line is IN. To the right is OOS. Judging by the comment, the OOS is 2021-2024. Then why is it so short on the chart?

Anyway, some confusion is being posted.

I've written 100500 times about my charts, there is even an article on the tester.

short can occur for various reasons, for example, there are fewer patterns on new data.

Or on new data clusters are shifted. The first follows from the second.
 
Maxim Dmitrievsky #:

short can occur for various reasons, e.g. there are fewer patterns on new data

Or the clusters are shifted on the new data. The former follows from the latter.

Forum on trading, automated trading systems and testing trading strategies

Machine Learning in Trading: Theory, Models, Practice and Algorithm Trading

Maxim Dmitrievsky, 2024.11.30 12:02 AM

strive for such that IS and OOS were similar, it means that there is no drift of patterns on new data.

If the frequency of trades on OOS is noticeably lower (if higher - no opinion) than on IN, I don't know how to treat it.
 
fxsaber #:
If the frequency of trades on OOS is noticeably lower (if higher, no opinion) than on IN, I don't know how to feel about it.
With understanding. The general case was described first, then the new code was tested. The point of taking it out of context?

The question is specifically what?)
 
Maxim Dmitrievsky #:
The question is specifically what kind of question?)

Please display on the charts which (from->to) intervals IN, OOS, and how many trades on each of them.

For better understanding.

 
fxsaber #:

Please display on the charts which (from->to) intervals are IN, OOS, and how many trades are on each of them.

For better understanding.

Number of trades at the bottom of the chart 😵‍💫
IN/OS vertical stick 😱

Considered as trolling

I'm interested in discussing MO algorithms.
 
Maxim Dmitrievsky #:
IS/OS vertical stick 😱
from->to.
 
fxsaber #:
from->to.
Earlier in the text, in the context of the discussion.
10 to 25, OS to 21
 
Maxim Dmitrievsky #:
Earlier in the text, in the context of the discussion.

I had a question like this. If Python graphics had this information, you could avoid being distracted by dumb questions.

1...365636573658365936603661366236633664366536663667366836693670...3674
New comment