Machine learning in trading: theory, models, practice and algo-trading - page 1581
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
What kind of goals do you have for a neural network?
I was considering these options:
1) binary prediction of the next candle color;
2) change of price on the next candle, i.e. regression
3) Detecting the beginning of a trend, classification - up, down, flat.
So far only the first variant accounts for 64% of successful predictions.
You can see my last article (there will probably be another on the same topic soon). You can attach a neural network to that one.
Classification/Regression no matter what
I understand those who succeeded and left)) you can not reveal the details, but the results should be popularized
alas, the one who popularized the results publicly, the one who leaked publicly - unfortunately, it always happens that way (((
on tasks.... You can find a profitable TS with the help of the MO, the main problem is to determine when this TS stops working, to either re-train the NS or another
HH: search through the forum for articles "neural networks" - a lot of ready-made solutions, imho in runet on this forum one of the largest selections on neural networks, etc. https://www.mql5.com/ru/search#!keyword=%D0%BD%D0%B5%D0%B9%D1%80%D0BE%D1%81%D0%B5%D1%82%D0%B8&module=mql5_module_articles
I understand those who succeeded and left)) You may not reveal the details, but the results should be popularized
If no one shows any results, it means they have bad results, or no results at all. Isn't it clear.
Who will have good results, he will show the whole world :)
In short, this is the theme: who starts working, he leaves and does not tell anyone.
Are you sure they're leaving because it worked out?
The only one who left was the doc who gave a lot of good advice and who could have succeeded. But he didn't say anything about what worked.
If it had worked, there would have been an interesting signal or PAMM somewhere. But there wasn't one either.
My personal experiences were fruitless. I spent almost 2 years in vain. It was interesting to play with my brain.)
It's a bit boring now.
Are you sure they're leaving because it worked out?
Doc is the only one who left who gave a lot of sound advice and who could have made it work. But he didn't say anything about what worked.
I've been a little bit open about their secrets by implication, and I can say, like Alexander, that these people (are they people?) know something. TCs are 100% working. Not eternal, not perfect, but workers.
I've been a little bit open about their secrets by implication, and I can say, like Alexander, that these people (are they people?) know something. TCs are 100% working. Not eternal, not perfect, but working.
About 100% working strategies...
Tomorrow Trump will fart his tomahawks, all these strategies will come to naught. (This is not a prediction, that tomorrow, but somehow it comes to that).
In my opinion the same Alexander is a good storyteller and just knows how to intrigue people with a sci-fi speech.
About working 100% strategies...
Trump farts tomahawks tomorrow, all those strategies will go down.
Everything can be, but who knows... let there be intrigue. Personally, I'm pretty sure they have some good TCs. And the proportion of MO there is negligible, so just for fun.
My personal experiments were fruitless. I spent two years almost for nothing. It was interesting to stretch my brain))
It's kind of boring right now...
Try adding my research on leaf groupings to random forests.
I assume that the forests are unbalanced by votes, and after grouping the voting leaves, you can improve the results. I suggest you give it a try, since you have figured out in detail the library with which to build forests in MT5.
In short, this is the situation. Pacemaker for any trades is the average returnee shift. Knowing the shift on the interval, you can build any profitable TS, but sooner or later it changes its sign. Without the displacement indicator, all the TC will be situational.
Because of this, it is easy to predict the SB on some interval.
I don't have time yet to do an article on the mismatches, but this thing looks pretty appealing
Pacemaker for any trades is a shift in the average returnee
there is something in it!
I was testing in the 1000 and 1-st configuration of a grid of orders, the main problem is that forward is very bad on real ticks and the dynamics of the equity graph on the forward is shifted in time, in a couple of hours I will show you the screenshots of the tests, what I am writing about