Welcome to Scribd. Sign in or start your free trial to enjoy unlimited e-books, audiobooks & documents.Find out more
Download
Standard view
Full view
of .
Look up keyword
Like this
1Activity
0 of .
Results for:
No results containing your search query
P. 1
The Concept of a Linguistic Variable and Its Applications to Approximate Reasoning III-1975

The Concept of a Linguistic Variable and Its Applications to Approximate Reasoning III-1975

Ratings:

5.0

(1)
|Views: 31|Likes:
Published by api-3717234

More info:

Published by: api-3717234 on Oct 14, 2008
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

05/09/2014

pdf

text

original

IN FO R M A T IO N SC IEN C ES9,43-80(1975)
4 3
T h e C o n c e p t o f a L in g u is tic V a ria b le
a n d its A p p lic a tio n to A p p ro x irq a te
R e a so n in g -III*
L. A. ZADEH
C O m p u ter S cien ces
d iv is io n ,
D e~ u rtm en t
o f E ~ e ~ ~ i & a ~E n g i n e e r i n g
an d
C om p u ter
S cien ces,
an d
th eE lect ro n ics
R esea rch
L a b o ra t o ry ,
U n iv ersity
o f
C a lifo rn ia ,
\u2018B e r k e l e y ,
C d ifo rn ia
9 4 7 2 0
1 . L IN G U IS T IC P R O B A B IL IT IE S
AND AVERAGES OVER FUZZY SETS
In th e c la s s ic a l a p p ro a c h to p ro b a b ility
th e o ry , a nev en t,
A ,
is d e fin e d a sa
m e m b e r o f a o -fie ld ,,d ,
of subsets of a sam ple space ft. T hus, ifP
is a norm ed
m easure over a m easurable space (a,_~#), the probabibty
of A is defined asP(A),
the m easure of.4,
a n d is a n u m b e r in th e in te rv a l [0 , I].
T h e re a re m a n y re a l-w o rld p ro b le m s in w h ic h o n e o r m o re o f th e b a s ic
a s s u m p tio n s w h ic h a re im p lic it in th e a b o v e d e fin itio n
are violated. F irst, the
e v e n t, A , is fre q u e n tly
ill-d e fin e d , a s in th e q u e s tio n , \u201c W h a t is th e p ro b a b ility
th a t it w ill b e a w arm
d ay
to m o rro w? \u201d In th is in s ta n c e , th e e v e n tw arm
d ay
i safu z z yev en t
in th e s e n s e th a t th e re is n o s h a rp d iv id in g lin e b e tw e e n
i t s
o c c u rre n c e
and nonoccurrence.
A s show n in [48], such an event m ay be char-
a c te riz e d
a s a fu z z y s u b s e t, A , o f th e s a m p le s p a c e Q , w ith tiA, th e m e m b e rs h ip
fu n c tio n
o f A , b e in g a m e a s u ra b le fu n c tio n .
S e c o n d , e v e n if A is a w e ll-d e fin e d n o n fu z z y
e v e n t, its p ro b a b ility , P (A ),
m a y b e ill-d e fin e d . F o r e x a m p le , in re s p o n s e to th e q u e s tio n , \u201c W h a t is th e
p ro b a b ility
th a t th e D o w J o n e s a v e ra g e o f s to c k p ric e s w ill b e h ig h e r in a
m o n th
fro m
n o w ? \u201d it w o u ld b e p a te n tly
u n re a s o n a b le to g iv e a n u n e q u iv o c a l
n u m e ric a l a n s w e r, e .g ., 0 .7 . In th is in s ta n c e , a v a g u e re s p o n s e lik e \u201c q u ite
p ro b a b le ,\u201d
w ould be m uch m ore com m ensurate
w ith o u r la c k o f u n d e rs ta n d in g
o f th e d y n a m ic s o f s to c k p ric e s , a n d h e n c e a m o re re a lis tic -if
less precise-
c h a ra c te riz a tio n
o f th e p ro b a b ility
in q u e s tio n .
- -
* T h is w o rk
w as supported
in p a rt b y th e N a v y E le c tro n ic
System s
C om m and
u n d e r
C o n tra c t
N 0 0 0 3 9 -7 1 -C -0 2 5 5 ,
th e A rm y
R esearch
O ffice,
D urham ,
N .C ., u n d e r G ra n t
D A -
A R O -D -3 1 -I2 4 -7 1 -G l7 4 ,
a n d
th e N a tio n a l
S c ie n c e
F o u n ~ tio n
under G rant G K -10656X 3.
T h e w ritin g
o f th e p a p e r w a s c o m p le te d
w h ile th e a u th o r
w a s p a rtic ip a tin g
in a Joint
S tu d y
P ro g ra m
w ith
th e
System s
R esearch
D e p a rtm e n t,
IB M
R esearch
L a b o ra to ry ,
San Jose,
C a lifo rn ia .
P a rts I a n d
II of this
p a p e r
a p p e a re d
i n
~ n fo r~ t io n
S cien ces
8 (3 1 , 1 9 9 -2 4 9 ;
8 (4 1 ,
3 0 1 -3 5 7 (1 9 7 5 ).
@ A m erican
E ls e v ie r P u b lis h in g
C om pany,
In c ., 1 9 7 5
4 4
L. A . ZA D EH
T h e lim ita tio n s im p o s e d b y th e a s s u m p tio n th a t A
is well-defined m ay be
re m o v e d , a t le a s t in p a rt, b y a llo w in g A to b e a fu z z y e v e n t, a s w a s d o n e in
[4 8 ]. A n o th e r a n d p e rh a p s m o re im p o rta n t
s te p th a t c a n b e ta k e n to w id e n
th e a p p lic a b ility o f p ro b a b ility
th e o ry
to ill-d e fin e d p ro b le m s is to a llo wP

t o b e a lin g u is tic v a ria b le in th e s e n s e d e fin e d in P a rt II, S e c . 3 . In w h a t fo llo w s , w e s h a ll o u tlin e a w a y in w h ic h th is c a n b e d o n e a n d e x p lo re s o m e o f th e

e le m e n ta ry
consequences of allowingP
to b e a lin g u is tic v a ria b le .
L IN G U IS T IC
P R O B A B IL IT IE S
T o s im p lify o u r e x p o s itio n , w e s h a ll a s s u m e th a t th e o b je c t o f o u r c o n c e rn
is a v a ria b le , X , w h o s e u n iv e rs e o f d is c o u rs e , U , is a fin ite s e t
u = u 1
tu *
t---tu n .
(1 .1 )
F u rth e rm o re ,
w e assum e that the restriction
im p o s e d b y X c o in c id e s w ith II.
T h u s , a n y p o in t in U c a n b e a s s ig n e d a s a v a lu e to X .
W ith e a c h u i,i = 1 , . . . , n , w e a s s o c ia te alin g u is t ic p ro b a b ilit y , p i,
w h ic h
is a B o o le a n lin g u is tic v a ria b le in th e s e n s e o f P a rt II, D e fin itio n
2 .2 , w ith
p i,0 G
P i
d
1 , re p re s e n tin g th e b a s e v a ria b le fo r& @ { .
For concreteness, w e
s h a ll a s s u m e th a t V , th e u n iv e rs e o f d is c o u rs e a s s o c ia te d w ith \u2018g i, is e ith e r th e
u n it in te rv a l [0 , I]
or the finite set
V = O tO .l
t.* * to .9 +
1 .
(l-2 )
U s in g 4 8 a s a g e n e ric n a m e fo r th e .J ? ~ , th e te rm -s e t fo r @ w ill ty p ic a lly
b e
th e fo llo w in g .
T (9 )
= lik ely + not lik elyt
u n lik ely
+ v ery lik ely t m ore or les s lik ely
+ v ery u n lik ely
t
- **
+p ro b a b le
+ im p ro b a b le
+ v ery p ro b a b le
t
. . .
t
n eit h er v ery p ro b a b le
n o r v ery im p ro b a b le
+. \u2019 .
+ c l o s e t o O t c l o s e t o 0 .1
t - * .t c l o s e t o
1t . . -
+very close to 0 t
very close to 0.1 t
. . . ,
(1 .3 )
in w hichlik ely , p ro b a b le
a n d close to p la y th e ro le o f p rim a ry
term s.
T h e s h a p e o f th e m e m b e rs h ip fu n c tio n
o f lik ely
w ill b e a s s u m e d to b e lik e
th a t o ft ru e
[s e e P art II, E q. (3.2)],w ithn o t lik ely
a n d u n lik ely
d e fin e d b y
p ~ cH
lik ely
@ )
=
\u2019
-
p lik ely
@ )y
(1 .4 )
L I N G U I S T I CV A R I A B L E
III
45
and
p
u n lik ely
@ )
=
p lik ely
c 1
- p),
(1 .5 )
w here
p is a generic
nam e
fo r th e p i.
EX A M PLE
1 .1 .
A g ra p h ic
e x a m p le
o f th e
m e a n in g
a tta c h e d
t o
th e
term s
lik ely, n ot lik ely, very lik ely
a n d
u n lik ely
i s
show n
in F ig .
1 . In
num erical
term s,
if the
p rim ary
te r m
lik ely
is d e fin e d
a s
lik e ly = 0 .5 / 0 .6 + 0 .7 1 0 .7 t
0 .9 1 0 .8t
l/0 .9
t
l/l,
(1 .6 )
th e n
n ot likely =
l/( 0
t
0 .1
t
0 .2t
0 .3t
0 .4 t 0.5)t 0 .5 1 0 .6t
0 .3 1 0 .7t
0 .1 /0 .8 ,
(1 .7 )
u n lik ely =
l / O
+
l/O .1
+ 0 .9 1 0 .2
t
0 .7 1 0 .3
t
O S /O .4
(1 .8 )
a n d
v er y lik ely = 0 .2 5 1 0 .6t
0 .4 9 1 0 .7t
0 .8 1 /0 .8
t
1 1 0 .9 t
l/l.
(1 .9 )
\u2018P
U n lik e ly
N ot
lik e ly
I
F ig .
1 . C o m p a tib ility
fu n c tio n s
o f
lik ely .
n o f
lik ely ,
u n lik ely
a n d
v e ry
lik ely .
T he
te r m
p r o b a b le w ill be
assum ed
to be more
or less synonym ous
w ith
lik ely .T he
te r m
c l o s e
to a, w here
(Y is a p o in t
in
[0 , l], w ill b e a b b re v ia te d
a s
2
o r, a lte rn a tiv e ly ,
a s \u201c o ? ,l
s u g g e s tin g
th a t
(Yis a \u201cbest
e x a m p le \u201d
o f th e
fu z z y
s e t \u201c c Y \u201d . In th is
s e n s e
th e n
>
>
\u2018T h e
sym bol
\u201cor\u201d w ill
b e
e m p lo y e d
i n
p la c e
of zwhen
th e
c o n s tra in ts
im posed
b y
ty p e -
s e ttin g
d ic ta te
i t s
u s e .

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->