Is there a pattern to the chaos? Let's try to find it! Machine learning on the example of a specific sample. - page 31
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
The gist of the answer is also expected. :)
Then we'll see you on the cover of Forbes or on the Nobel Prize.
Would you sell that Grail for money?
I understood correctly in this thread ML/DL/NN has triumphed and it forecasts (not trading yet) in plus with zero spread/commission but on randomly generated inputs?
and it's being seriously discussed.
I added a column: "1" buy, "-1" sell. Opening a deal in case of a signal change to the opposite one. I think I did it correctly, but I didn't check it ... :) Laziness.
On the chart without spread and commission this is the result in points:
Results: PR=157488 +trades=778 -trades=18 (profit, number of positive and negative trades)
Spread of 0.00050:
Results: PR=117688 +trades=629 -trades=167
Spread at 0.00100:
PR=77888 +trades=427 -trades=369
Spread at 200:
PR=-1712 +trades=241 -trades=555
and got the inverse function of crossing two MAs .....
again the same pitfalls, if you try to trade not virtually, but at least on demo.I propose to download a new sample - which is small in the number of predictors and strings - so it will be relatively fast to learn.
However, I have not managed to get something interesting out of it yet - please show me the level of skill!
Or just share the results.
I suggest downloading the new sample - which is small in the number of predictors and rows - so it will be relatively quick to learn.
However, I have not managed to get anything interesting out of it yet - please show me the level of skill!
Or just share the results.
Hi!
Sample train:
sample test:
Exam sample:
Calculations are done for POINT=0.00001, spread and commission = 0.
I didn't find OP,HI,LO,CL separately in the sample - maybe I wasn't looking hard enough ... :)
The results may be better for them.
The approach is very simple ... :)
Hi!
...
The approach is simple as hell ... :)
Hello. What's the approach?
Hello. What's the approach?
No catch!!! :)
Like last time, the graph that is analysed and predicted is the last column transformed by the formula target(i)=target(i-1)+Target_100_Sell(i). The resulting graph is like this:
Next, the graph data is run through the GA and the best SUPER-oscillator coefficients are obtained:
An enlargement from 1000 to 1300 counts:
Next, a neural network committee tries to predict the future value by one step. The resulting predictions are driven into patternnet and tries to get a signal from horizons + two counts ahead. The error for one step is around 5-6%, for two steps 11-13%. Well, the patternnet output and analysis of the chart itself determines either up or down. I.e. always in the market, either buy or sell.
No catch!!! :)
Just like last time, the graph that is analysed and predicted is the last column transformed by the formula target(i)=target(i-1)+Target_100_Sell(i). The resulting graph is like this:
Next, the graph data is run through the GA and the best SUPERoscillator coefficients are obtained:
Enlarged from 1000 to 1300 counts:
Next, a neural network committee tries to predict the future value by one step. The resulting predictions are driven into patternnet and tries to get a signal from horizons + two counts ahead. The error for one step is around 5-6%, for two steps 11-13%. Well, the patternnet output and analysis of the chart itself determines either up or down. I.e. always in the market, either buy or sell.
Sounds incredible!
Is it possible to get a ready-made model to reproduce the results?
That sounds incredible!
Is it possible to get a ready-made model to reproduce the results?
I suggest re-reading the answer from the last page: https: //www.mql5.com/ru/forum/434981/page30#comment_45863608.
The whole point here is that it took a lot of time and effort to get a seemingly very simple result.
But the conjuncture has changed, hence the decision to price the question (only one copy of the material in one hands): 92k$. :)
And to the haters (especially those who write about all sorts of MA, MO and the victory of pallor over reason) I would like to ask only one question: "Is it weak to repeat such a result?".
Sincerely, RomFil
I suggest re-reading the response from the last page: https: //www.mql5.com/ru/forum/434981/page30#comment_45863608
The whole point here is that it took a lot of time and effort to get a seemingly very simple result.
But the conjuncture has changed, hence the decision to price the issue (only one copy of the material in one hands): 92k$. :)
And the haters (especially those who write about all sorts of MA, MO and the victory of pallor over reason) want to ask only one question: "Is it weak to repeat a similar result?".
Sincerely, RomFil
Note, I did not ask for material to repeat the results, but I am interested to look at the result itself in the form of a model.
Do you plan to participate in the contest?