[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [amibroker] Re: Optimization Question



PureBytes Links

Trading Reference Links

I don't know how long 40k steps take to perform on your machine but a solution to the problem using a single processor took less than 5 minutes on mine.  I wouldn't think AB's internal engine would take much more than this i.e. 40k steps @ 100 / sec = 400 secs ~ 7 minutes ?!
 
Correct ... But I have a good reason to get the complete grid and not just a small part of it. I am looking for an optimalization of area's. Not just points. But you're right, CMAE is about 5 minutes on my computer ... That's why I said 'even with CMAE' ... So yes, it's acceptable ...
 
Regards, Ton.
 
 
----- Original Message -----
Sent: Thursday, January 08, 2009 5:33 PM
Subject: Re: [amibroker] Re: Optimization Question

Obviously problems like this aren't going to be solved with exhaustive search ...
 
This is why there are artificially intelligent optimizaion engines whether they be imbedded in AB or not.
 
I can't say I have used the ones inside AB much since my initial investigation of them and long ago I posted my thoughts on the particular AB artificially intelligent optimization engine you reference.
 
I don't know how long 40k steps take to perform on your machine but a solution to the problem using a single processor took less than 5 minutes on mine.  I wouldn't think AB's internal engine would take much more than this i.e. 40k steps @ 100 / sec = 400 secs ~ 7 minutes ?!

----- Original Message -----
From: Ton Sieverding
Date: Thursday, January 8, 2009 6:04 am
Subject: Re: [amibroker] Re: Optimization Question
To: amibroker@xxxxxxxxxps.com

> Fred my problem with all kinds of combinations like in yr code
> is the following ...
>
>
>
> Even with CMAE ...
>
>
>
> Regards, Ton.
>
>
> ----- Original Message -----
> From: ftonetti@xxxxxxxxxxnet
> To: amibroker@xxxxxxxxxps.com
> Sent: Thursday, January 08, 2009 7:55 AM
> Subject: Re: [amibroker] Re: Optimization Question
>
>
>
> One way to code this is as below ...
>
> One could of course code in separate "Use" optimizable
> variables for each of the indicators to allow them to be
> independently used or not on the buy / sell side.
>
> The default values in the optimization statements below were
> arrived at from the best results based on CAR - MDD over the
> entire life of a common Index using daily data. Notice that the
> optimization process determined NOT to use the MACD indicator
> i.e. MACDUse = 0 but did employ the other two.
> RSILen = Optimize("RSILen", 43, 1, 100, 1);
>
> RSIUse = Optimize("RSIUse", 1, 0, 1, 1);
>
> RSIx = RSI(RSILen);
>
> RSIBuy = RSIx > Ref(RSIx, -1) OR RSIUse == 0;
>
> RSISell = RSIx < Ref(RSIx, -1) OR RSIUse == 0;
>
> StoLen1 = Optimize("StoLen1", 16, 1, 100, 1);
>
> StoLen2 = Optimize("StoLen2", 25, 1, 100, 1);
>
> StoLen3 = Optimize("StoLen3", 99, 1, 100, 1);
>
> StoUse = Optimize("StoUse", 1, 0, 1, 1);
>
> StoK = StochK(StoLen1, StoLen2);
>
> StoD = StochD(StoLen1, StoLen2, StoLen3);
>
> StoBuy = StoK > StoD OR StoUse == 0;
>
> StoSell = StoK < StoD OR StoUse == 0;
>
> MACDLen1 = Optimize("MACDLen1", 32, 1, 100, 1);
>
> MACDLen2 = Optimize("MACDLen2", 60, 1, 100, 1);
>
> MACDLen3 = Optimize("MACDLen3", 48, 1, 100, 1);
>
> MACDUse = Optimize("MACDUse", 0, 0, 1, 1);
>
> MACDx = MACD(MACDLen1, MACDLen2);
>
> MACDs = Signal(MACDLen1, MACDLen2, MACDLen3);
>
> MACDBuy = MACDx > MACDs OR MACDUse == 0;
>
> MACDSell = MACDx < MACDs OR MACDUse == 0;
>
> Buy = RSIBuy AND StoBuy AND MACDBuy;
>
> Sell = RSISell AND StoSell AND MACDSell;
>
>
>
> ----- Original Message -----
> From: longarm61
> Date: Thursday, January 8, 2009 12:49 am
> Subject: [amibroker] Re: Optimization Question
> To: amibroker@xxxxxxxxxps.com
>
> > Thanks, appreciate it. If it's not too much trouble, could
> you
> > (or
> > someone) give me a quick example of how exactly that would
> be coded?
> >
> >
> > --- In amibroker@xxxxxxxxxps.com, ftonetti@xxx wrote:
> > >
> > > One way to do this is to have an on/off ( 1/0 )
> optimizable
> > variable for each indicator so that individual indicators
> are
> > either
> > used or not and then let optimzation make the determination.
> > >
> > > ----- Original Message -----
> > > From: longarm61
> > > Date: Wednesday, January 7, 2009 8:25 pm
> > > Subject: [amibroker] Optimization Question
> > > To: amibroker@xxxxxxxxxps.com
> > >
> > > > Is there a way to optimize a group of indicators so that
> the
> > > > results
> > > > not only give you the optimized values, but also tell
> you
> > which
> > > > indicators you'd have better of f not using at all?
> Something
> > > > like
> > > > this:
> > > >
> > > > Best return: Indicator1=80, Indicator2=NO,
> Indicator3=50,
> > > > Indicator4=NO
> > > >
> > > > Thanks in advance,
> > > >
> > > > Grant
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> >
> >
> >
>
>
>

__._,_.___

**** IMPORTANT ****
This group is for the discussion between users only.
This is *NOT* technical support channel.

*********************
TO GET TECHNICAL SUPPORT from AmiBroker please send an e-mail directly to
SUPPORT {at} amibroker.com
*********************

For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/

For other support material please check also:
http://www.amibroker.com/support.html

*********************************




Your email settings: Individual Email|Traditional
Change settings via the Web (Yahoo! ID required)
Change settings via email: Switch delivery to Daily Digest | Switch to Fully Featured
Visit Your Group | Yahoo! Groups Terms of Use | Unsubscribe

__,_._,___