A quick and free library for MT4, much to the delight of neuralnetworkers - page 12
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
Until you try it, it's hard to assess the benefits of the inputs, one area may work better for one and another for another. You can get an average result on almost any input if you try. Naked is without preprocessing, "0-1" is just a special kind of normalization, it's good, but it may not be enough... Preprocessing is a whole science, imho more complicated than neural networks themselves, and compression, and bleeding, and coding and probably a lot more. You can start by looking at articles by V.A. Krisilov, you can get them from http://neuroschool.narod.ru/. What you have in mind, to put a phase in NS, I use only as a component of a complex combination of inputs, nothing more.
Have you actually achieved any sane results? Without preprocessing in FANN to shove "bare" data is useless IMHO, non-normalized data you can try to test in STATISTICA, there at testing is given out committee of 5 grids, trained on different algorithms. And using statistics I came to not very accurate results, I wrote you a message in private.
Have you actually achieved any meaningful results? Without preprocessing in FANN to put "bare" data is useless IMHO, non-normalized data you can try to test in STATISTICA, when testing a committee of 5 grids, trained on different algorithms. And using statistics I came to not very accurate results, I wrote you a private message.
The result is just only sane is used, but I can not say that it is the merit of ZZ).
It is not a commit there, but just a few (kol'nosti set) grids selected for minimum error, but that Stathis should be fed with unprepared data is a delusion exactly. Of course in search of solutions you may try different things, but to dump everything on net and let it cook, doesn't suit you, anybody here will tell you so.
So, it turns out that 16 grids initialized with random weights from -1 to 1, after the first execution of ann_runs(...) with one InputVector[], we get (judging by logs) 16 identical outputs accurate to 8 characters??? No. There is a bug of some kind here.
You wrote yourself that the subject of neural networks does not just work. So we have to figure it out...
Bug search in this case has to be done on the spot. One can only speculate remotely. In the case you mentioned, I personally did not see how the grids were trained before it turned out from the log that the answers of all grids are the same in one survey and different in another, but still the same. And now the situation is no clearer, because ann_runs only executes the grids, but does not train them. I can only advise to put debug printers on all library calls. Only this parsing, imho, relates more to the logic of the code than to the meshes themselves.
The search for the bug in this case must be done on the spot. We can only speculate remotely.
Forgive me, but I thought we are discussing Yury's Expert Advisor posted in Codebase (post on the 3rd page).
So:
-- I did not change anything in EA, only added debug prints.
-- Clean up the ANN directory
-- Run the Expert Advisor in test mode
-- It seems to work, but looking at the log file, we see that the answers from all networks are the same! :-(
When a new grid is trained, the initial weights are initialized with random numbers. Since the initial weights are different, the training for the meshes will also be different and the final result will also be different.
There was a suspicion that weights were not randomized when creating grids. But this was not confirmed. An archive of grid files saved immediately after their creation and initialization is attached.
Look at the plz. Where to dig further do not know. If only in the depths of dll?
I have only one Expert Advisor, but everyone is different and everyone tests it on different data and under different conditions ;-). I have looked at it once, I didn't have such a problem as you have, but I understood that it makes no sense to use it in its current form and I have no time to rewrite it.
Fuck the councillor. There's no need to rewrite and deal with it.
There is a suspicion of incorrect operation of the library itself, and it is necessary to find out if this is the case or not. Otherwise there is no point in moving on.
If you don't have problems with it, don't consider it a problem, post the log of FANN-EA test with parameter DebugLevel = 3
Guys, are you sure the library isn't "crooked"?
The situation is as follows: an error message pops up regularly and that the terminal will be closed.
Here is the text of the message:
There has been a critical error
Time : 2009.12.23 22:55
Program : Client Terminal
Version : 4.00 (build: 225, 10 Jul 2009)
OS : Windows XP Professional 5.1 Service Pack 3 (Build 2600)
Processors : 2 x X86 (level 6)
Memory : 1047532/161248 kb
Exception : C0000005
Address : 02086AD8
Access Type : read
Access Addr : FFFFFFFFF
Registers : EAX=FFFFFFFFFFFF CS=001b EIP=02086AD8 EFLGS=00010206
: EBX=FFFFFFFFFFFF SS=0023 ESP=0291B8E8 EBP=0291B8E8
: ECX=09104830 DS=0023 ESI=00000000 FS=003b
: EDX=09104830 ES=0023 EDI=0000015C GS=0000
Stack Trace : 02011D13 00463E17 00455691 FFFFFFFFF
: 00000000 00000000 00000000 00000000
: 00000000 00000000 00000000 00000000
: 00000000 00000000 00000000 00000000
Modules :
1 : 00400000 002B1000 c:\program files\metatrader - alpari\terminal.exe
2 : 01750000 00327000 c:{windows\system32\xpsp2res.dll
3 : 02010000 0000A000 c:program files\metatrader - alpari\experts\libraries\fann2mql.dll
4 : 02030000 00039000 c:{windows\system32\tbb.dll
5 : 02070000 00034000 c:\windows\system32\fanndoublemt.dll
6 : 14240000 00020000 c:\program files\drweb\drwebsp.dll
7 : 5B260000 00038000 c:{windows\system32\uxtheme.dll
8 : 61EC0000 0000E000 c:{windows\system32\mfc42loc.dll
9 : 698B0000 00058000 c:{windows\system32\hnetcfg.dll
10 : 71A30000 00040000 c:{windows\system32\mswsock.dll
11 : 71A70000 00008000 c:{windows\system32\wshtcpip.dll
12 : 71A80000 00008000 c:{windows\system32\ws2help.dll
13 : 71A90000 00017000 c:{windows\system32\ws2_32.dll
14 : 72CD0000 00008000 c:{windows\system32\msacm32.drv
15 : 72CE0000 00009000 c:{windows\system32\wdmaud.drv
16 : 73D90000 000FE000 c:{windows\system32\mfc42.dll
17 : 746E0000 0004C000 c:\windows\system32\msctf.dll
18 : 75310000 0002E000 c:{windows\system32\msctfime.ime
19 : 76350000 00005000 c:{windows\system32\msimg32.dll
20 : 76360000 0001D000 c:\windows\system32\imm32.dll
21 : 76380000 00049000 c:\windows\system32\comdlg32.dll
22 : 76770000 0000C000 c:\windows\system32\cryptdll.dll
23 : 769A0000 000B5000 c:{windows\system32\userenv.dll
24 : 76B20000 0002E000 c:\windows\system32\winmm.dll
25 : 76C20000 0002E000 c:\windows\system32\wintrust.dll
26 : 76C80000 00028000 c:{windows\system32\imagehlp.dll
27 : 76D50000 00019000 c:{windows\system32\iphlpapi.dll
28 : 77110000 0008B000 c:{windows\system32\oleaut32.dll
29 : 773C0000 00103000 c:\windows\winsxs\x86_microsoft.windows.common-controls_6595b64144ccf1df_6.0.2600.5512_x-ww_35d4ce83\comctl32.dll
30 : 774D0000 0013D000 c:\windows\system32\ole32.dll
31 : 77A70000 00096000 c:{windows\system32\crypt32.dll
32 : 77B10000 00012000 c:{windows\system32/msasn1.dll
33 : 77BC0000 00007000 c:{windows\system32\midimap.dll
34 : 77BD0000 00015000 c:{windows\system32\msacm32.dll
35 : 77BF0000 00008000 c:{windows\system32/version.dll
36 : 77C00000 00058000 c:{windows\system32\msvcrt.dll
37 : 77C60000 00025000 c:{windows\system32/msv1_0.dll
38 : 77DC0000 000AC000 c:{windows\system32\advapi32.dll
39 : 77E70000 00092000 c:{windows\system32\rpcrt4.dll
40 : 77F10000 00049000 c:{windows\system32\gdi32.dll
41 : 77F60000 0007C000 c:{windows\system32/shlwapi.dll
42 : 77FE0000 00011000 c:{windows\system32\secur32.dll
43 : 78480000 0008E000 c:\windows\winsxs\x86_microsoft.vc90.crt_1fc8b3b9a1e18e3b_9.0.30729.4148_x-ww_d495ac4e\msvcp90.dll
44 : 78520000 000A3000 c:\windows\winsxs\x86_microsoft.vc90.crt_1fc8b3b9a1e18e3b_9.0.30729.4148_x-ww_d495ac4e\msvcr90.dll
45 : 7C800000 000F8000 c:\windows\system32\kernel32.dll
46 : 7C900000 000B3000 c:\windows\system32\ntdll.dll
47 : 7C9C0000 01979000 c:{windows\system32\shell32.dll
48 : 7E360000 00091000 c:{windows\system32\user32.dll
Call stack :
This never happened before using the Expert Advisor on FANN.
Is everything running like clockwork for you?
The situation is as follows: regularly an error message pops up saying that the terminal will be closed.
Regularly in relation to what? After a news release?
I've had a couple of terminal crashes while the EA was running, but they are directly related to the EA code.
For example: call of deinit() at the end of init()
The library is designed for a user who understands what he does and why. And it doesn't detract from its merits...
Regularly regarding its use. When testing and installing the Expert Advisor in a trade.
I wrote the Expert Advisor on FANN myself using a totally different principle of input-output organization and optimization from Reshetov's one.
Therefore I regard your last sentence as nonsense.