How to use a strategy tester in the stock market? - General

[Deleted] 2006.11.22 15:16 #21

Why not use external NN systems

Something like NeuroShell, TS or the open source Joone? You can call the external dll from Metatrade. The time and effort should be spent on selecting the inputs(indicators or good strategies) and training rather than writing a NN system in MQL4. I would go this way.

Exchange Execution - Opening Exchange Execution - Opening Creating an indicator draft

Tommy 2006.11.23 00:47 #22

Thats a good idea, ill look into that right now. Thanks.

-Tommy

Tommy 2006.11.26 23:24 #23

GP2X:
Something like NeuroShell, TS or the open source Joone? You can call the external dll from Metatrade. The time and effort should be spent on selecting the inputs(indicators or good strategies) and training rather than writing a NN system in MQL4. I would go this way.

Joone looks really good but do you know how to write an imput file that woudl incorporate an indicator and teach to predict its signals? I cant make heads or tales of it, and i cant find anything about it in the forums. thanks.

Market App Store Relative Strength Index - MQL Wizard and program

[Deleted] 2006.11.27 15:59 #24

I think there are two ways, one is to use a Java indicator library, and feed Joone with arrays of numbers, like what this guy did: http://www.jooneworld.com/wiki/tiki-index.php?page=FinancialForecastTutorial#comments

Another is my idea, we can write an EA, which creates the input files(probably also the validation files) when running inside the strategy tester, then train Joone with these files.

Finally, after training, when using it for real, we have to use an EA to talk to Joone. All of these are not hard to me(will just take some time). The hard one is what to be used as inputs(definately not raw data, should be a combination of indicators). The structure and number of hidder neurals are also unknown, but we could find by repeating the training process.

Cyclesurfer:
Joone looks really good but do you know how to write an imput file that woudl incorporate an indicator and teach to predict its signals? I cant make heads or tales of it, and i cant find anything about it in the forums. thanks.

Mechanism for describing the Creating an indicator draft OOP fundamentals: Abstraction

[Deleted] 2006.11.27 16:05 #25

For example, if you want to use MA as the input, just write an EA, which saves the MA values to a CSV file for each bar, run the EA in tester, if you get two years of history data, the CSV file would contain two years of MA values. Then you can use it as the input file.

GP2X:
I think there are two ways, one is to use a Java indicator library, and feed Joone with arrays of numbers, like what this guy did: http://www.jooneworld.com/wiki/tiki-index.php?page=FinancialForecastTutorial#comments

Another is my idea, we can write an EA, which creates the input files(probably also the validation files) when running inside the strategy tester, then train Joone with these files.

Finally, after training, when using it for real, we have to use an EA to talk to Joone. All of these are not hard to me(will just take some time). The hard one is what to be used as inputs(definately not raw data, should be a combination of indicators). The structure and number of hidder neurals are also unknown, but we could find by repeating the training process.

Additional Features - Price Sending data frames from Uninstalling the Platform -

fx4_ever 2006.12.24 19:19 #26

Artificial_Intelligence EA

Hi:

I found this EA, but not sure what strategies does this EA based on..

Can someone take a look and explain here please?

I attached EA and backtesting result.

Files:

strategytesterai1.htm 30 kb

strategytesterai1.gif 7 kb

artificialintelligence.mq4 6 kb

witchazel 2006.12.24 20:09 #27

its all based on williams accelerator/decelerator, which the formula is

AO = SMA(median price, 5)-SMA(median price, 34)

AC = AO-SMA(AO, 5)

here is the code in EA

extern int x1 = 135;

extern int x2 = 127;

extern int x3 = 16;

extern int x4 = 93;

double w1 = x1 - 100;

double w2 = x2 - 100;

double w3 = x3 - 100;

double w4 = x4 - 100;

double a1 = iAC(Symbol(), 0, 0);

double a2 = iAC(Symbol(), 0, 7);

double a3 = iAC(Symbol(), 0, 14);

double a4 = iAC(Symbol(), 0, 21);

return(x1* a1 + x2 * a2 + x3 * a3 + x4 * a4);

so it ends up being

return(35 * ac(this bar) + 27 * ac(7 bars ago) + -84 * ac(14 bars ago) ) + -7 * ac(21 bars ago) );

if it ends up being greater then 0 then it buys and if less then it sells. looks like its alway in the market as well. It only opens 1 order at a time though.

It looks like a variation of the alligator by Williams but it uses 4 ACs instead of 3.

how they pick the x1,x2,x3,x4 numbers? i dont know, seems random to me

why they -100 from each of them? no idea, should have just enterd a smaller number

Accelerator Oscillator - Bill Accelerator Oscillator - Bill Accelerator Oscillator - Bill

Craig 2006.12.24 20:46 #28

My 2 cents, Neural Networks don't really work as far as I can see, I have spent countless hours coding and testing such beasts with nothing really to show.

When evaluating such methods it is important to understand what is actually happening, to cut though the BS as it were. Neural Networks are not magic, just non-linear regression, as such they can do an amazing job of fitting test data but tell you nothing of the future. The other mistaken assumption which is applied to NN & Forex is that there is some type of 'hidden pattern' in the data which the NN will deduce, I have not found this to be the case. In my experence the greater the degrees of fredom of a system the more prone it is to overfitting the data, hence giving no value in data ranges outside the test data.

Regularization Statements Neural network training

fx4_ever 2006.12.24 20:46 #29

Thanks!

WOW.

Received reply so quickly.

Thanks a lot. Have a good holiday!

witchazel:
its all based on williams accelerator/decelerator, which the formula is