How to form the input values for the NS correctly. - page 22
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
Why take indicators? Metastock has an interesting tool, Maximum Profit System (MPS), designed to compare the profitability of systems. The MPS is supposed to calculate all probable trades with a positive result. It is very convenient to create training arrays for MLP.
Where is lib.mqh ? If you can post it...
Where is lib.mqh ? If you can post...
Sorry ...
By the way, don't be fooled, but you have yet to do the tambourine dance. At first glance, it looks like this is it.
And it's just a markup on the size of the tracks... BUT, its bad quality that it grabs MPS not only tracks, but also jumps in price ("fat tails").
And here it's more a question for Matemat, how would this series, which is obtained in the course of MPS work, be handled so as to exclude jumps. (MA and its
The best thing I've got working so far is heuristic algorithm written in Prolog.
To analyse the initial price range.
Seemingly good, but again there is a nuance, the network requires regular pre-learning. And the heuristics have to be fine-tuned. Also, trads, for NS training are easier.
marking by hand ;) It's a vicious circle, and if you holograph a clever one, a poorly formalizable task.
The topic must not be abandoned! Who might have something new? Or have something to share...
MPS - you can modify it to avoid the problems mentioned above, I'll post it as I see fit...
The topic must not be abandoned! Who might have something new? Or have something to share...
MPS - it is possible to modify, so that there were no above-mentioned problems, as I shall do correctly from my point of view, I shall lay out...
Don't hurry :), I won't abandon this theme.
The topic must not be abandoned! Who might have something new? Or have something to share...
What do you develop your NS on?
After running a couple of variants, just to check and compare the speed between MQL and VC++, I decided to go with the latter.
At the moment I'm writing classes for VC. I only need to get the input and output from mql. C++ does the rest.
If you want, I can give you this set of classes as soon as I debug everything.
What do you develop your NS on?
After running a couple of variants, just to check and compare the speed, between MQL and VC++, I decided to go with the latter.
At the moment I'm writing classes for VC. All I need from mql is to get the input and output. C++ does everything else.
If you want, I can give you this set of classes as soon as I debug everything.
Halt!!! I already have a ready-made VC++ lib.
Except there are 2 problems there:
1. binding to Boost, I want to get rid of it, it's better to serialize it manually, it glitches anyway.
2. something with adaptive pitch.
Why make a bicycle? Especially there.
1. MLP with possibility to create a tree structure.
2. std::valarray + aggressive optimization of operations for faster counting.
3. there is an adaptive step
4. patterns with auto-normalization.
5. wide opportunities for expansion.
Ы ?
What do you use to develop your NS?
After running a couple of variants, just to check and compare speed, between MQL and VC++, I decided to go with the latter.
At the moment I'm writing classes for the NS. I only need to get the input and output from mql. C++ does everything else.
If you want, I can give you this set of classes as soon as I've debugged it.
I'm not working out NS, now I'm searching for optimal inputs and outputs for training sample, I think proper sampling is more important than NS, there are a lot of variants of NS in different languages in the net...
What do you develop your NS on?
After running a couple of variants, just to check and compare the speed, between MQL and VC++, I decided to go with the latter.
At the moment I'm writing classes for VC. All I need from mql is to get the input and output. C++ does everything else.
If you want, I can give you this set of classes as soon as I've debugged it.
I use my own classes. Data preparation in MQL + C#, .dll - C++ ... So far it's the most convenient.
Most optimal to use for matlab testing.
Freeze!!! I already have a ready-made VC++ libc.
Only there are 2 problems:
1. binding to Boost, I want to get rid of it, it's better to serialize it manually, it glitches anyway.
2. something with adaptive pitch.
Why should we make a bicycle? Especially there
1. MLP with possibility to create tree structure.
2. std::valarray + aggressive optimization of operations for faster counting.
3. there is an adaptive step
4. patterns with auto-normalization.
5. wide opportunities for expansion.
Ы ?
What's a lib?