Re: fuzzy finance


Subject: Re: fuzzy finance
From: Pramit Sarma (pramits@bom8.vsnl.net.in)
Date: Sat Mar 04 2000 - 07:02:23 MET


All,

I too have been watching this parley with some interest. Besides
interest in things AI, real-world stuff (process engineering) definitely
indicates the "steal with pride" thing. The way engineers put it, is
that which works well (demonstrably) shall be put to work!

It is, intuitively, clear that all this Theory of Possibility vs.
Probability is more like one multi-wrench vs. another. The possible
(probable?) problem being here, that here the two can and do have some
similar "heads". Systems common sense, rather than wrangling, seems to
indicate both tools are not only useful, the similarities need to
highlighted, leading to a clearer way of looking at the critical and
*useful* differences. Synergy as the key, and all that. That, and the
pretty clear systems axiom: "There is no Universal Model"!

It is possible that everyone in the discussion is not also involved in
fuzzy logic control (FLC), or rather in control systems. So, despite
being a strong proponent of FLC, nevertheless there are many quasi-crisp
controller models which have been working for decades, very well. One of
this set is the Optimal/Stochastic Controllers and Estimators (LQG,
Kalman, etc.)**. They are all variously based on probabilistic Gaussian
noise models. So indeed there are, and several, probabilistic control
applications. The fact that some of the most successful FLC fuzzy terms
tend to be Gaussian (an observation of my own FLC work, too) ... appears
to me as one of those synergies.

** (For an interesting debate on crisp/optimal control vs. fuzzy
control, with Athans vs. Zadeh, have a look at the IEEE Control Systems
Mag., June 1999)

Finally, I think that in systems the phrase "The end justifies the
means" is particularly attractive and devoid of most of its negative
connotations, for (intelligent) systems engineers ... and to the most
important person at the end of the loop: the end-user.

Pramit

Paul Victor Birke wrote:
>
> "Philip E. Bennett" wrote:
> >
> > Paul, et al
> >
> > I've been monitoring this thread with muted interest. The discussion has
> > bounced around and touched on many topics but the main interest seems to be
> > in fuzzy vs probabilistic methods and their application to financial
> > modelling/trading.
> >
> > As a private trader, former institutional derivatives trader, and a bit of a
> > financial engineer I am concerned only with how well these techniques work
> > to help me create workable models. I define a workable model has one that
> > provides information so as to reduce the uncertainty regarding the
> > interpration of a specific situation or the course of action to take. In
> > this regard, IMHO, possibilistic (i.e. fuzzy logic) and probabilistic (i.e.
> > neural nets, datamining, etc.) methods are complementary and HIGHLY
> > synergistic! The use of GA's and SOM's to help fine tune the application of
> > these methods are also benefical.
> >
> > Would it be appropriate to suggest that if you're interested in gaining more
> > success in modelling you focus more on what can be gained from by using the
> > methods together?
> **********************
> Dear Philip
>
> Well in fact that is what I had done at the time back in about 1996.
> While I appear to be a bit one sided in fact I am not so.
>
> I would agree with you fully here. There a loads of insights to be
> obtain be sticking to formal Probability Theory and its offshoots. The
> University of Guelph library has a number of books on subjective
> probability assessment of uncertain prediction that derived in part from
> the Cold War as just a for instance. I was most interested modelling
> uncertainty in the marketplace. I became interested in the idea of
> using or borrowing from Fuzzy Set Theory the UltraFuzzy Function. You
> can roam around the UltraFuzzy using a Geometric Mean definition between
> the lower and upper bounds. These bounds can also be interpreted as
> probability limits.
>
> So as an engineer Steal With Pride anywhere I can get ideas to obtain
> some advantage. I am not stuck in the ground fuzzy person. I try to
> use each where advantage.
>
> In discussing this problem at work today I was reminded that Fuzzy has
> shown advantage in Control Applications wherein a loop or control loop
> is the key idea here. I will pass this along to Peter Hamer in a few
> minutes. Probability Theory seems lacking in the control context. And
> maybe why is failed should be highlighted.
>
> all for now,
>
> Paul
>
> ############################################################################
> This message was posted through the fuzzy mailing list.
> (1) To subscribe to this mailing list, send a message body of
> "SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at
> (2) To unsubscribe from this mailing list, send a message body of
> "UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com"
> to listproc@dbai.tuwien.ac.at
> (3) To reach the human who maintains the list, send mail to
> fuzzy-owner@dbai.tuwien.ac.at
> (4) WWW access and other information on Fuzzy Sets and Logic see
> http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info
> (5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html

-- 

========================================================================== Pramit 'Jake' Sarma

E-mail: PramitS@vsnl.com [Res] PSarma@ChE.IITB.ernet.in [Lab] W-mail: pramit_s@india.com

Processes + Control, Intelligent & Simulation Systems | P R O C I S S | {Systems + Mathematics}---{Real World Processes} ==========================================================================

############################################################################ This message was posted through the fuzzy mailing list. (1) To subscribe to this mailing list, send a message body of "SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at (2) To unsubscribe from this mailing list, send a message body of "UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com" to listproc@dbai.tuwien.ac.at (3) To reach the human who maintains the list, send mail to fuzzy-owner@dbai.tuwien.ac.at (4) WWW access and other information on Fuzzy Sets and Logic see http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info (5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html



This archive was generated by hypermail 2b25 : Thu Apr 06 2000 - 15:59:47 MET DST