|
PureBytes Links
Trading Reference Links
|
Hi Al –
I completely agree. Both of us would
prefer a set of parameter values that are robust. So both of us will add
a robustness component to our objective function. That means that either AB,
or our own code, needs to evaluate the shape and smoothness of the objective
function response surface. That will be pretty far down Tomasz$B!G(Bs list,
but we can do ourselves.
Howard
-----Original Message-----
From: Al Venosa
[mailto:advenosa@xxxxxxxxxxxx]
Sent: Sunday, October
19, 2003 <span
>7:00 AM<font size=2
face=Tahoma>
To: amibroker@xxxxxxxxxxxxxxx
Subject: Re: Objective functions
(was RE: [amibroker] Re: Optimization -- again)
[HB] If I look at the list of alternative
models sorted according to my objective function and decide, for some reason,
that I prefer one other that the one at the top of the list, then I have not
included something important in the definition of my objective function.
I need to revise my objective function, and rerun earlier test runs and model selections.
<span
>
[AV] Howard, I can envision one reason for
rejecting the top parameter value or model of an objective function without
resorting to the need to improve the definition of the objective function.
Suppose, as an example, your objective function were simply net % profit, which
AB's opt currently sorts on. Suppose the net% profit in the top model were
100%. Then, looking at the next 3 or 4 models just below the top, suppose the
net % profit were 85.14, 84.26, 83.89, and 81.99. I would choose one of the
next 3 rather than the top one because choosing the top one imposes a non-robustness
to your system results, a local optimum that may have occurred by chance.
Changing a parameter just a small bit changes the profitability by 15%, whereas
changing that same parameter by the same amount from there only changes
profitability a few tenths of a percent. So, the system is more robust in that
area of 2- or 3-dimensional space than the top one.
<span
>
Al Venosa
<span
>
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.525 / Virus Database: 322 - Release Date: 10/9/2003
<font size=3
face="Times New Roman">
<font size=2
color="#003399" face="MS PGothic">Send
BUG REPORTS to bugs@xxxxxxxxxxxxx<font
face="Courier New">
Send SUGGESTIONS to
suggest@xxxxxxxxxxxxx<span
>
-----------------------------------------<font
face="Courier New">
Post AmiQuote-related messages ONLY
to: amiquote@xxxxxxxxxxxxxxx<font
face="Courier New"> <span
>
(Web page: <a
href="">http://groups.yahoo.com/group/amiquote/messages/)<font
face="Courier New">
--------------------------------------------<font
face="Courier New">
Check group FAQ at: <a
href="">http://groups.yahoo.com/group/amibroker/files/groupfaq.html<font
face="MS PGothic">
Your use of Yahoo! Groups is subject
to the Yahoo! Terms of Service.<font
face="MS PGothic">
Yahoo! Groups Sponsor
ADVERTISEMENT
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
|