**Subject: **Re: BISC: resend

**From: **H. Mark Hubey (*HubeyH@mail.montclair.edu*)

**Date: **Sun Jul 23 2000 - 21:19:52 MET DST

**sorted by:**[ date ] [ thread ] [ subject ] [ author ]**Next message:**Grigorios Tsoumakas: "FuzzyCLIPS and Delphi"**Previous message:**Michelle T. Lin: "BISC: resend"**Maybe in reply to:**Michelle T. Lin: "BISC: resend"**Maybe reply:**H. Mark Hubey: "Re: BISC: resend"

"Michelle T. Lin" wrote:

*>
*

*> *********************************************************************
*

*> Berkeley Initiative in Soft Computing (BISC)
*

*> *********************************************************************
*

*>
*

*> To: BISC Group
*

*> From: L. A. Zadeh <zadeh@cs.berkeley.edu>
*

*>
*

*> A Challenge to Data Miners: The Soccer Problem
*

*> ----------------------------------------------
*

*>
*

*> Recently, I was watching the soccer match between France and
*

*> Portugal. I noticed that most of the time the ball was in the vicinity
*

*> of the goal of Portugal. This observation suggested the following
*

*> hypothesis, call it H.
*

*>
*

*> For generality, let the opposing sides be labeled A and B. Let
*

*> r be the fraction of time the ball spends in the vicinity of the goal
*

*> of A. The hypothesis is that the closer the value of r is to 1, the
*

*> higher the probability that B will win.
*

*>
*

*> To make the hypothesis more concrete, assume that r is
*

*> measured as follows. Let the playing field be partitional into zones
*

*> R1,...Rn, with R1 being nearest to the goal of A and Rn the farthest.
*

*> Let ri be the fraction of time the ball spends in Ri, i=1,..,n. Let
*

*> wi, i=1,..,n be weights ranging in magnitude from 0 to 1. Then
*

*> r=w1r1+...+wnrn.
*

*>
*

*> Does there exist a choice of the Ri and this wi such that H is
*

*> true? This is the crux of the problem. The assumption is that we
*

*> analyze N games, with r computed at the end of each game. The result
*

*> for game j, j=1,..,N, will be W(j) (win), D(j) (draw) and L(j) (lose),
*

*> with r being r(j), j=1,..,N. These data, then, would serve as a basis
*

*> for testing H.
*

*>
*

*> The soccer problem is an instance of a problem in data mining
*

*> in which a hypothesis, H, is (a) generated; (b) tested; and (c)
*

*> modified. In my view, it is a challenging problem because how to
*

*> choose and adjust the Ri and wi is not a simple matter.
*

It seems like this can be solved using the method in my paper

named Curse.pdf.

It's just that instead of a weighted average kind of a solution

we might get a nonlinear product-sum kind of a solution.

-- Regards, Mark /\/\/\/\/\....I love humanity. It's people I can't stand...../\/\/\/\/\ ==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-== hubeyh@mail.montclair.edu =-=-=-=-=-= http://www.csam.montclair.edu/~hubey############################################################################ 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

**Next message:**Grigorios Tsoumakas: "FuzzyCLIPS and Delphi"**Previous message:**Michelle T. Lin: "BISC: resend"**Maybe in reply to:**Michelle T. Lin: "BISC: resend"**Maybe reply:**H. Mark Hubey: "Re: BISC: resend"

*
This archive was generated by hypermail 2b25
: Sun Jul 23 2000 - 21:42:13 MET DST
*