A quick and free library for MT4, much to the delight of neuralnetworkers - page 38
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
Guys, has anyone heard of this advisor NEURON-SYSTEMS? I know it's pretty old, but I can't find any neural network EAs on the web ...........
It's about time I had a "birth" of one myself ... :-)))
It's about time you 'gave birth' to a chen... :-)))
I can't "give birth" to it, I'm not a coder:))) I can run it or proofread it, but coding would be a feat for me :)). I'm not a technician, I'm an economist by education:))
Well, this week FANN with a trained network seems to trade better, today he decided to proopt it during the trade, it turned out some nonsense .....stopped the same, after optimization just reloaded the terminal and everything, left it the same, I think maybe to attach a trail, because in the bu sometimes need to move and it is stupid waiting either SLT or SL....
Well, this week FANN with a trained network seems to trade better, today he decided to proopt it during the trade, it turned out some nonsense .....stopped the same, after optimization just reloaded the terminal and everything, left it the same, I think maybe to attach a trail, because in the bu sometimes need to move and it is stupid waiting either SLT or SL....
there's a lot to do... "everything is simple", but "here" is too simple... Don't bet on the real in this form... These are just basic "points" in the EA (inputs, outputs), explaining the principle of neurofilter itself... I'm sure, as you wrote, "five losses in a row" is not the limit :-))
there's a lot to be done... "Everything is simple," but "here" is too simple... Don't bet on real in this form... These are just the basic "points" in the EA (inputs, outputs), explaining the principle of neurofilter itself... I'm sure, as you wrote, "five losses in a row" is not the limit :-))
No, of course I won't bet, although it has rehabilitated today and seems to be even 5 profitable trades in a row, but it's a trend :)) I think maybe a couple of trivial averages and trawl.
Yeah no, won't bet of course, although as of today it has rehabilitated itself and seems to have even had 5 profitable trades in a row, but that's the trend :)) I think maybe we should use some simple averages and trawl.
On small TF the averages do not work well, too much "noise", trail also need, of course, we need to look at what kind of trail... Take a look here, if it's like you say,
screw it on... - http://fms.1free.ws/2.html. I haven't fully come to grips with these issues myself yet. "Digging in other directions, I'm getting there...
On small TFs mediums work poorly, too much "noise", trawl is of course needed too, need to look - what kind... Take a look here, if it is, as you say,
screw it on... - https://www.mql5.com/go?link=http://fms.1free.ws/2.html. I haven't fully come to grips with these issues myself yet. "Digging" in other directions, I'm approaching...
If I read it, the guy has a buy picture (according to his rules) and he selects a seed, it may be something I don't understand, but I don't know if it will work and how to connect it with the library or will the library calculate everything automatically?
I read, the man has a picture to buy (according to his own rules) and he settles, strange as that, maybe I do not understand what, but not the point, it is probably possible to attach, will it all work and how to link with the library, I do not know, or in the library everything will be calculated by machine?
Well, of course, it's just using the library as a neurofilter - no more, i.e. - training, etc. (we need a more qualitative approach to the question of preparing the input data - their normalization), "dude" has it right (on the strategy) - re-read more. And in the EA, replace the MACD (RSI) signal input with CCI,BB, MA - like he has. Although, maybe it will work fine without neurofilter... To properly connect the neurofilter you should have about 1000 trades...
Well, of course, this is just using the library as a neurofilter - nothing more, i.e. also training, etc. (you need a better approach to input data preparation - their normalisation), the "dude" has everything right (by strategy) - re-read more. And in the EA, replace the MACD (RSI) signal input with CCI,BB, MA - like he has. Although, maybe it will work fine without neurofilter... To properly connect the neurofilter you should have about 1000 trades...
About 1000 - this is of course tinny.... as Uneo can do, but not on a minute, but on an hour, but I just need to bring the TF we are using, but I can't do it myself.
About 1000 is of course a tantrum .... as ineo can be done, but not on a minute, but on an hour, but I can't do it myself.
Well, how is the tin? I tried to optimise, including forward "library" IAD etc. if there are few trades - the network is not trained and no use, if there are too many trades > 1000, then
The net is poorly trained again as it cannot find dependencies in term of price movement. It turns out that 700 to 1000 сd is just enough - at least in my tests. If we make H1 in TF.
the number of deals would fall dramatically... If I want to use neurofilter exactly for scalping strategy variant to filter entries, not to mention creation of complete neuro-advisor - these questions should be solved by myself as soon as possible.
These matters should be dealt with, such things cannot be solved by hastening...