Machine learning in trading: theory, models, practice and algo-trading - page 961
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
Then it should all work only as long as historical prices are repeated...
Maybe that's the problem ) I'll re-do it and see
Exactly.
A certain Asaulenko does just that. Although he is trying to loop like a hare, he is still a physicist and I have confidence in his model.
And it is as follows - it looks whether the price has left the level of confidence probability, and the NS additionally gives the permission / refusal to enter the trade. I have the same thing, only instead of NS I use Pearson's asymmetry coefficient. But, his is better - I want to do it that way too.
Yes, I understand that he has a breakdown of ranges, but the demon is buried in the nuances, as usual.
Suggest, which currency pair to choose (EURUSD-H1 or XAUUSD-H1), and what periods to set as a training and testing? 10000/1000? 30000/1500? I can also set the forest parameters as I wish. The package is randomForest(https://cran.r-project.org/web/packages/randomForest/randomForest.pdf). I want my work to be as socially useful as possible :)
any, better an article with research :)
History/Future = 10000/1000
On the graph we can clearly see the beginning of the OOS period)) Input - indicator RSI (100 components)
formula for calculating the j-th input iRSI(NULL, 0, 2+j, 4, i+1+Shift) - 50) / 50, the period of RSI is 2+j where j is from 0 to 99.
Yes, we also know how to rebate them )) to beat off the spread through the affiliate
yes, we also know how to use such rebates ))
Yes, the first pancake, but now I'm doing a new run. Now the input is a body of candles - the output also only as a signal (0.1). the strategy without filtering.
The one at the bottom is a test one. I got sad result with 2 pips spread, i.e. -398 pips. I will try to filter it. I think I don't need the training period.
History/Future = 10000/1000
On the graph we can clearly see the beginning of the OOS period)) Input - indicator RSI (100 components)
formula for calculating the j-th input iRSI(NULL, 0, 2+j, 4, i+1+Shift) - 50) / 50, the period of RSI is 2+j where j is from 0 to 99.
Only a segment with feedback (without learning and validation)
LevelOpen > 0.1(< -0.1) Profit 73 pips
LevelOpen > 0.15(< -0.15) Profit 21 pips
A tiny profit but still a profit. At levels lower than 0.1 this is a loss. Now we should try to outrun these results)