Machine learning in trading: theory, models, practice and algo-trading - page 2022
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
This week it is holding on until Thursday), but there have been no market reversals yet. It is unclear how quickly the market will react
Have you used lstm? Try it, it's cool
I don't use ltsm, but the point is not in ltsm but in RL ... isn't it?
No, I didn't use ltsm, but the point is not in ltsm, but in RL... isn't it?
I'm just talking about recurrent meshes... the essence is the same, but the approaches may be different
you can build something like that, you can do it without RL, you can just teach something
I'm just talking about recurrence grids... the essence is the same, just the approaches may be different
you can build something like that, you can just teach something without rl
Maxim, tell me, did you write the grid yourself? Or did you use a ready-made one?
Ready-made, mostly in Python
Check out the sequel. It's working this week. The next one still needs a test, with new logic.
I think if market trades, you can add slippage of 10-20% of a minute candle height if it is too high (i.e. if there was a gap).
If trading by pending orders, then it is necessary to skip deals at very high minute candlesticks (gaps).
It is enough to take an average spread (in it to include the commission, if any) *1.5, for stress tests, if it is not a wild hft. I.e. depends on the number of deals.
If the slippage is big, then the problem is not in the Trading System. If large, then the problem is not in the TS
If you want to check how big the slippages are, then the problem is not with your TSO. Your own tester is much more flexible and faster, without overheads. Better suited for optimization, for example.
Check out the sequel. It's working this week. The next one still needs a test, with new logic.
cool!
cool!
there are cool materials on the internet on predicting the behavior of different robots in an unknown environment. From control theory.
there are cool materials on the Internet on predicting the behavior of different robots in an unknown environment. From control theory.
In Russian? Send me the link.
So, do you use your own stuff or is it the same dude's script?
in Russian? Throw me a link.
So are you already using your own, or is it the same dude's script?
English... it's closed access on the medium, you need a paid subscription
I'm making another bot now... it's just in testing
put lstm or gru and some reinforcements in it... then go on to the new NS architectures for time series