Suppose we have a more or less stationary BP and have its frequency decomposition. Question: what is better for the model - 1 feature or 5, and why?
If you happen to have such decomposition and it is correct and reversible, then increasing the number of features for the model is only good and increases its flexibility.
Ivan Negreshniy: Well, if you happen to have such decomposition and it is correct and reversible, then increasing the number of features for the model only benefits and increases its flexibility, the main thing is that the MO tools do not get pumped up, by the way now there are new, progressive neural networks that make deep learning achievements at a higher rate.
Well, + the most noisy elements removed. And what are the ns, what are they called? and what about xgboost, do they make it too? )
Well + the noisiest elements removed. And what are the ns, what are they called? and what about xgboost, do they make it too? )
This is a new yet little-known n-network architecture, without any gradient descents, everything is primitively simple, just like the brain of the above mentioned comrade).
Ivan Negreshniy: This is a new yet little-known n-network architecture, without any gradient descents, everything is primitively simple, like the brain of the above mentioned comrade).
Where to read? I can't even find anything to google.
Yuriy Asaulenko: I believe that the daytime will be even better. Judging by the model and preliminary debugging runs. But it is too early to say anything, of course.
Allow the use of cookies to log in to the MQL5.com website.
Please enable the necessary setting in your browser, otherwise you will not be able to log in.
You are signed in. Two-factor authentication has been disabled. Set it up again and scan a new QR code, or continue with two-factor authentication disabled.
Suppose we have a more or less stationary BP and have its frequency decomposition. Question: what is better for the model - 1 feature or 5, and why?
Well, if you happen to have such decomposition and it is correct and reversible, then increasing the number of features for the model only benefits and increases its flexibility, the main thing is that the MO tools do not get pumped up, by the way now there are new, progressive neural networks that make deep learning achievements at a higher rate.
Well, + the most noisy elements removed. And what are the ns, what are they called? and what about xgboost, do they make it too? )
there is a training of water ..... so that it "remembers" and leaks itself
Well + the noisiest elements removed. And what are the ns, what are they called? and what about xgboost, do they make it too? )
This is a new yet little-known n-network architecture, without any gradient descents, everything is primitively simple, like the brain of the above mentioned comrade).
Where to read? I can't even find anything to google.
here are all the options https://tproger.ru/translations/neural-network-zoo-2/
Where to read? I can't even find anything to google.
it seems that all options are here https://tproger.ru/translations/neural-network-zoo-2/
not much is written about it in the public domain, but research is underway https://creatime.me/MediaLibrary/Zanimatelno/NovayaEra
I believe that the daytime will be even better. Judging by the model and preliminary debugging runs. But it is too early to say anything, of course.
How's it going?
not much is written about it in the public domain, but research is going on https://creatime.me/MediaLibrary/Zanimatelno/NovayaEra
not much is written about it in the public domain, but research is going on https://creatime.me/MediaLibrary/Zanimatelno/NovayaEra
ah, well, there are no libraries yet... I can't write them myself)