Machine learning in trading: theory, models, practice and algo-trading - page 2127

New comment
 
elibrarius:

Two columns must be fed into the model - both sine and cosine for the clock. And sine + cosine for the day of the week. See the link for a description of why this should be done.

pi = 3.141529 ... from school.

Okay, I'll give you two...

And about pi, so the number may be too big - who knows what accuracy is required...

3,1415926535897932384626433832795…
 
Maxim Dmitrievsky:
You have CATboost 😑

So, and? I wonder :))) Days of the week he spit out, now let's see in the new numbers wrap on the result.

 
Aleksey Vyazmikin:

So, and? I'm curious :))) Days of the week he spit out, now let's see in the new numbers wrap on the result.

I added it above. I told you this 50 years ago last week
 
Maxim Dmitrievsky:
Above added
Maxim Dmitrievsky:
You have CATboost 😑 just mark features as categorical

I can't add categoricality to MQL code :(

 
Aleksey Vyazmikin:

I have no way to load categoricality into MQL code :(

Doesn't that lib work with cuttings?
 
Aleksey Vyazmikin:

Okay, I'll give you two...

As for Pi, that number might be too high - who knows what accuracy is required...

3,1415926535897932384626433832795…

7 digits is enough

 
Maxim Dmitrievsky:
Doesn't that lib work with kat chips?

No.

 
Aleksey Vyazmikin:

No.

Teach it in python, there are 2 lines
 
Igor Makanu:

saw this book a couple of years ago

It looks... Well, yes, it fascinates, but really - why? if the purpose of writing a diploma or PhD - yes, it's a board book

if the purpose of time series - this book is about something else, about the invention of the random forest at the dawn of the computer

imho, even ensembles of NS poorly accustomed to application in practice, how to work with BP? well, as an option to mess up a bunch of a lot of NS, and in the end you get autoecoder? - I doubt that even a convolutional network can be obtained with the help of this book


Old knowledge, Vorontsov is more relevant, and data processing - I'm chewing on some online courses on BP - there's something in it ;)

I should have read it a couple of years ago, when I first posted it here )). These approaches are still in use, including for time series, and have a number of advantages over deep learning. For example Zircon works better than Lstm on time series, and the principle is the same as for MGUA. Autoencoders and convolutions are a different story. Is there anything on time series to look at? Usually they are all about seasonality and autoregression. And in fact, these components in the market only get in the way.
 
Maxim Dmitrievsky:
Train in python, there are 2 lines

I must have misunderstood your question.

There is no model interpreter on MT5 with categorization predictors and CatBoost with command line is able to do everything that python version does, except purely python things, such as visualization.

1...212021212122212321242125212621272128212921302131213221332134...3548
New comment
Reason: