wdr@wang.com (William Ricker) (06/04/90)
Book reports on Fuzzy stuff, as promised.
[I've dropped comp.realtime and added sci.math to the Newsgroups:.
Hardcore mathematicans should note the Appendix in Negotia on The Category
of Fuzzy Sets.]
Both of these were written as texts, but neither includes problems.
Negoita: "The annotated Readings pose sufficient real problems, and it
seemed inappropriate to add contrived ones. The reader is warned that
real expert systems are tailor-made."
I have read and thoroughly recommend Schmucker; I believe it is still in
print. (Note that CSP has changed ISBN prefixes since it was published;
later editions may have a newer ISBN than the one cited here.)
I have only recently acquired Negoita, and so can recommend it on only
the basis of the leafing through it, the full results of which are reported
here. I'm glad I bought it (used); I don't know if it is still available.
/s/ Bill Ricker
Amateur Mathematician & Alleged "Software Engineer"
----------------
Schmucker, Kurt J.
/Fuzzy Sets, Natural Language Computations, and Risk Analysis/
Computer Science Press. LC# QA248.S345 1894; ISBN* 0-91894-83-8. 192+xvpp.
[ * Old ISBN; CSP has new prefix, and may have renumbered this book? ]
Forward by Lofti A. Zadeh. 130pp in 6 chapters: Review of Set Theory,
Fuzzy Set Theory, Natural Language Computation, Psychological Considerations
of Fuzziness, The Fuzzy Risk Analyzer, Future Research; 24pp in 3 appendices:
Formal Definition of a Linguistic Variable, The Extension Principle,
Implementation of Fuzzy Sets; 38pp of bibliography and index.
Lofti Zadeh's Foreward says "This book ... serves to introduce the
reader to the theory of fuzzy sets and explains clearly and with many
examples the use of the linguistic approach. Mr. Schmucker derserves to
be complimented for presenting a coherent and self-contained account of
a body of concepts and techniques which are of considerable relevance
to risk analysis and natural language computations, and for contributing
many insigts which facilitate their application to the solution of
practical problems."
This book was reviewed at length (1.5pp) in the annotated Readings
of the book below; after paraphrasing Zadeh's praise above,
Negoita continues: "Schmucker observes that the very core of fuzzy set
theory, the degree of membership, is difficult to grasp and that,
fortunately, the linguistic variable -- a notion built on top of fuzzy
set theory -- is an alternative. He observes also that the terms
/linguistic variable/ and /fuzzy set/ are not interchangeable; having
precisely manipulable natural language expressions is the goal, and
fuzzy set theory (and in particular its use to represent linguistic
variables) is relatively new, the goal of having something like a
linguistic variable is rather old. As Schmucker observes, Leibnitz once
said: 'If we could find characters or signs appropriate for expressing
all our thoughts as definitely and as exactly as arithmetic expresses
lines, we could in all subjects, insofar as they are amenable to
reasoning, accomplish what is done in arithmetic and geometry.'"
Schmucker did this work as (part of?) his doctoral thesis at George
Washington U. under Prof. Lance Hoffman, separate from his better-known
work in Object-Oriented Macintoshes etc.
----------------
Negoit,a^ , C.V. (Constantin Vergil) [t-cedilla, a-hacheck(inverted ^)]
/Expert Systems and fuzzy systems/
(c)1985 Benjamin/Cummings Publ. Co.,Inc. QA76.9.E96N44 1984 (sic);
ISBN 0-8053-6840-X. 190+x pp.
Seven chapters (each ending with a summary and readings):
Introduction, Exact and Inexact Reasoning in Knowledge Engineering,
Fuzzy Sets, Knowledge Representation, Approximate Reasoning, Knowledge
Engineering in Decision Support Systems, Knowledge Engineering in Management
Support Systems; plus Appendix: The Categorical Analysis of Logic;
and a substantial bibliography and index.
The readings sections are annotated, and thus should be a useful
guide to the literature from 1965 to 1983. (Only two references to '84
made it: a paper of the authors, and the Schmucker book (above), which
was in-press for a year at least, since the forward was dated April '82
and the Preface, November '82. 1983 appears to be better represented
in the annotated Readings than in the bibliography.)
The book's examples focus on DSS & Management Expert Systems
because the author sees them as more ill-defined and thus more needing
the semantic approach. "This text describes ... semantic manipulations
[of fuzzy systems] and gives the reader the mathematical background
necessary to understand the algorithms used in approximate reasoning.
With this background, and with the of knowledge that practical results
have demonstrated the reasonableness of this approach, the reader can
begin to use algorithms in decision support systems knowledgeably."
Two criticisms: (1) the bibliography seems to be light on mainstream
expert systems papers; I spot only Shortliffe & Buchanan with one paper on
Mycin in the bibliography and Winston, The Handbook of AI, and Davis & Lenat
in the Intro's readings. (2) These entries in the readings do not
appear in the bibliography (although they are indexed). For this
scholarly a work, it is inexcusable not to cross reference the
annotated readings in the non-annotated bibliography.
The author's early self-references are from Budapest & eastern european
journals; his address in 1984 was Hunter College, CUNY.
----------------
To comment on the "theory of plausibility", which I am informed is
different from fuzzy <anything>, the confusion may result from the following
appearing in bibliographies without reports of later deveopments:
Zadeh, L.A. 1978, "Fuzzy Sets as a basis for a theory of possibility."
/Fuzzy Sets and Systems/ 1:3-28.^^^ ^^^
and,
various papers in the collection /Fuzzy set and possibility theory/,
ed. R. Yager. London: Pergamon Press
--
/bill ricker/
wdr@wang.com a/k/a wricker@northeastern.edu
*** Warning: This account not authorized to express opinions ***