|
Ken,
Any non-exhaustive optimizer will tell you *ESTIMATE* of steps
required.
That estimate is absolute MAXIMUM number of tests if will
use for given number of parameters.
You may change the granularity (i.e. STEP parameter of
optimization) and you will notice
that while exhastive number of steps changes, CMAE does not.
This is because CMAE search
continuous param spaces (unlike exhaustive).
For trivial tasks as this one it would overestimate a lot
because CMAE is supposed to be applied
on problems that require way more than 10000
steps.
Best regards,
Tomasz Janeczko
amibroker.com
----- Original Message -----
Sent: Tuesday, August 12, 2008 3:19
PM
Subject: [amibroker] Optimizing
Rotational - Curious Result
I thought this was a
curious development---I wondered if it was possible to optimize a rotational
model. Of course it is.
To test, I stuck three
variables into a fairly long afl code (complex positionscore
calculations).
Raw optimization said 75
steps, estimated time ~ 4 minutes.
I added the cmae engine
statement at the beginning and the optimize screen said 9662 steps with est
time of ~ 6 hours.
I thought the "cmae"
optimization step would decrease time, not increase the time so
dramatically.
What do you think was
going on?
Have any of you optimized
a rotational strategy?
Ken
__._,_.___
Please note that this group is for discussion between users only.
To get 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
![]()
__,_._,___
|