Machine learning in trading: theory, models, practice and algo-trading - page 1462

New comment
 
Maxim Dmitrievsky:

And don't peek, we'll figure it out without you. Go teach your blots, grailer.

Start making faces and showing your tongue.) (You're just a child's garden.)
 
Well, in general, I remembered about the trends. I'll add them to the list of features to check. Thank you!
 
elibrarius:
Well, I remembered about the trends. I'll add them to the list of features to check. Thank you!

please, your Master.

 
Yuriy Asaulenko:
You make faces and show your tongue.) Kindergarten).

What kind of opponent is such a communication, though

You come to the thread, you leave - is that your joy of communication?

 

Maxim Dmitrievsky:

or read about facebook prophet how models are built there, a lot of useful stuff. I don't have it on python, it may work on R

... And how the prophet is killed - it was shown in this thread, but the Teachers slipped through the cracks)))

 
Vizard_:

... and how the prophet is doing - it was shown in this thread, but the Teachers snapped it all up)))

Pruphs, I.D.'s, passwords... it's a pleasure to read...

 
elibrarius:
Of course, the forest is not able to add and multiply, and will not create indicators inside itself, but it is good at remembering patterns from price or indicator charts.
Here you are wrong. Forest reconstructs regression as successfully as NS, and can form indicators it needs and work directly with BP. Here it is only a question of setting a real task and training. In general, MO does not like tasks of the "go there, I don't know where" type...
 
Yuriy Asaulenko:
You are wrong here. Forest is as successful as NS in restoring regression, and can also internally form the indicators it needs and work directly with BP. Here it is only a question of setting a real task and training. In general, MO does not like tasks of the "go there, I don't know where" type...
The final result - will form. But purely by memorizing input combination of predictors.
It doesn't do any multiplication or division to combine anything.
 
elibrarius:
The final result will form. But purely by remembering the input combination of predictors.
It doesn't do any multiplication or division to combine anything.
It will remember everything perfectly at logic level. If it doesn't remember, it's a problem with the problem statement.
So, earlier in this thread it was shown that the forest does an excellent job of predicting real market BPs. By the way, I didn't show it, but it was trained on my data.
 
Yuriy Asaulenko:
It will remember everything perfectly well at the level of logic. If he does not remember - this is a problem of task formulation.
So, earlier in this thread it was shown that the forest does an excellent job of predicting real market BPs. By the way, I didn't show it, but it was trained on my data.
No arguments.
In my opinion, the forest is a database with quick access.
About 3 years ago I was making a template finder on MySQL. Among a million examples, searched for the 10-20 most similar ones on 50 columns. It took me about 1 second to do the math. Forest is hundreds or thousands of times faster to find a sheet in which these 10-20 examples are collected.
1...145514561457145814591460146114621463146414651466146714681469...3495
New comment
Reason: