Machine learning in trading: theory, models, practice and algo-trading - page 3588

New comment
 
Maxim Kuznetsov #:

it really is :-)

have you read the agreements and terms and conditions of the market?

"you can't guarantee or promise <here's a list of the average user's wants>"

and there are no other shops, and the user doesn't care about other things.


Not quite like that, or rather, not at all.))
Nowhere does it say, neither in the market cap nor in the rules, that it is recommended to buy a random product. On the contrary, a top of products is formed, the user goes in, sees the top, is guaranteed to go look, and with some probability buys. Anything further than the first page will not be bought with a probability close to 1.

And if you make a table of products (EA) with various indicators for the month / year, pf, the number of transactions, the factor wost, etc., then the audience of buyers expands many times.

Most importantly, the marketing factor will be reduced to 0, product descriptions will be the last thing they look at.)) Well, and there will be no monopoly.
 
Is this a thread about machine learning?
 
mytarmailS #:
Is this a thread about machine learning?

Pretty much what I wanted to do, but was already done before me :)

https://github.com/google-research/timesfm

https://github.com/moment-timeseries-foundation-model/moment

GitHub - google-research/timesfm: TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.
GitHub - google-research/timesfm: TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.
  • google-research
  • github.com
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting. This repo contains the code to load public TimesFM checkpoints and run model inference. Please visit our Hugging Face checkpoint repo to download model checkpoints. This is not an officially supported...
 
Maxim Dmitrievsky #:

Pretty much what I wanted to do, but had already been done before me :)

https://github.com/google-research/timesfm

https://github.com/moment-timeseries-foundation-model/moment

Waiting for the tests
 
mytarmailS #:
Waiting for the tests

Not enough memory, need a more powerful computer :(

because of this we won't get a long backtest, it will be very slow probably
 
Maxim Dmitrievsky #:

Not enough memory, need a more powerful comp :(

because of this you won't get a long backtest, it will be very slow probably
Whatever, do it!
It doesn't take much, after 50 bars you'll realise it's arima😁
 
mytarmailS #:
Whatever, do it!
It doesn't take much, after 50 bars you'll realise it's arima😁
Well it says it outperforms arima in some tests :)
 
Maxim Dmitrievsky #:
Well it says it outperforms arima in some tests :)
0.003%, that's super important to us).
1...35813582358335843585358635873588
New comment