Machine learning in trading: theory, models, practice and algo-trading - page 3093

[Forester]

mytarmailS #:
Is there anything about forests/NS in the article?

No

[mytarmailS]

Forester #:

no

then why are you looking for it in the code? )))

[Forester]

mytarmailS #:

then why are you looking for it in the code? )))

I don't get it. What am I looking for?

[mytarmailS]

Forester #:
Where are the forest/NS hyperparameters? No - so it's not training. Predictors aren't fed either.
I think it's just evaluating the stability of the external model predictions.

here

[Forester]

mytarmailS #:

here

I have the code for this package to look for anything in it. I read the article, that's all.

[mytarmailS]

Forester #:
I have the code for this package to look for something in it. I read the article, that's all.

Would you like to feel the package?

[Forester]

mytarmailS #:

Would you like to feel the bag?

R uninstalled it about 5 years ago and have no desire to install it. I have enough ideas for me. If I like it, I will do it myself to test it.
In this variant - there is nothing unique. Metrics in MT are roughly similar (Bal*RF, Bal*EP, Bal*PF), especially the composite metric is interesting. It is interesting to see its algorithm.

More precisely - unique, not sure which is better.
If you compare - it would be interesting to know the result

[mytarmailS]

Forester #:

R uninstalled about 5 years ago and have no desire to install. I have enough ideas for me. If I like it, I'll do it myself to test it out.
In this variant - there is nothing unique. Metrics in MT are roughly similar (Bal*RF, Bal*EP, Bal*PF), especially the composite metric is interesting. It is interesting to see its algorithm.

More precisely - unique, not sure which is better.
If you compare - it would be interesting to know the result

I won't be able to compare it to MT metrics.

I'm writing code now just to test this algorithm for usefulness.

but the code will be impressive by the standards of R.

[Forester]

mytarmailS #:

I can't compare it to MT's metrics.

I'm writing code now just to test this algorithm for usefulness

but the code will be impressive by R standards.

MT formulas are simple, everything can be taken from here and repeated

https://www.mql5.com/ru/articles/1486

https://www.metatrader5.com/ru/terminal/help/algotrading/testing_report#recovery_factor

https://www.mql5.com/ru/forum/4485

[Forester]

Forester #:

MT formulas are simple, everything from here can be pulled out and replicated

https://www.mql5.com/ru/articles/1486

MT gave a strange formula:

Expected Payoff =(ProfitTrades / TotalTrades) * (GrossProfit / ProfitTrades) -(LossTrades / TotalTrades) * (GrossLoss / LossTrades) reduce the highlighted colour, you will be left with

Expected Payoff = GrossProfit / TotalTrades - GrossLoss / TotalTrades = (GrossProfit - GrossLoss / TotalTrades)

(GrossProfit - GrossLoss) / TotalTrades = Balance / TotalTrades

And that's it!!! ))

Bottom line

Bal*EP = balance * fabs( balance / number of trades)

Other 2 indicators (with ProfitFactor and MaximalDrawDown ) by analogy.