(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 5, May 2011
Where
:
{ }
1()1()()(),2
jjqj
hxSignhxSignhhx
ω β α α β
⎡ ⎤= + − ∈⎣ ⎦
Definition 2.3
Invoke
fm
is fuzziness measure of hedgealgebra X,
f
: X > [0, 1].
∀
x
∈
X, denoted by
I(x)
⊆
[0, 1]and I(x) is measure length of I(x).A family J = {I(x):x
∈
X} called the partition of [0, 1]if:(1): {I(c
+
), I(c

)} is partition of [0, 1] so thatI(c) =
fm
(c), where c
∈
{c
+
, c

}.(2): If I(x) defined and I(x) = fm(x) then{I(hix): I = 1...p+q}is defined as a partition of I(x) so thatsatisfy conditions: I(h
i
x) =
fm
(h
i
x) and I(h
i
x) is linear ordering.Set {I(h
i
x)} called the partition associated with theterms x. We have
1
()()()
pqii
IhxIxfmx
+=
= =
∑
Definition 2.4
Set X
k
=
{ }
:
xXxk
∈ =
, consider P
k
=
{ }
():
k
IxxX
∈
is a partition of [0, 1]. Its said that u equalv at k level, denoted by u =
k
v, if and only if
I(u)
and
I(v)
together included in fuzzy interval k level. Denote
∀
u, v
∈
X, :()
kkk k
uvPIu
= ⇔ ∃Δ ∈ ⊆ Δ
and ()
k
Iv
⊆ Δ
.III.
D
ATA SEARCH METHOD
Let fuzzy class C = ({a
1
, a
2
, …, a
n
}, {M
1
, M
2
, …,M
m
});
o
is object of fuzzy class C. Denoted o.a
i
isattribute value of
o
on attribute a
i
(1
in
≤ ≤
) and o.M
j
isvalue method of
o
(
1
jm
≤ ≤
).In [2] we presented the attribute values are 4 cases: precise value; imprecise value (or fuzzy); object;collection. In this paper, we only interested in handingcase 1 and 2: precise value and imprecise value (fuzzyvalue) and to see precise value is particular case of fuzzyvalue. Fuzzy value is complex and linguistic label is oftenused to represent the value of this type. Domain fuzzyattribute value is the union two components:
Dom(a
i
)
=
CDom(a
i
)
∪
FDom(a
i
)
(
1
in
≤ ≤
).Where:
CDom(a
i
)
: domain crisp values of attributea
i
.
FDom(a
i
)
: domain fuzzy values of attributea
i
.
A.
Neighborhood level k
We can get fuzzy interval of terms length k as thesimilarity between terms. It means that the term thatrepresentative value of them depending on fuzzy interval
level k
is similar
level k
. However, to build the fuzzyinterval
level k
, representative value of terms
x
havelength less than k is always in the end of fuzzy interval
level k
. Hence, when determining
neighborhood level k,
we expect representative value it must be inner point of
neighborhood level k
.Based on fuzzy interval level
k
and
k+1
we constructa partition of the domain [0, 1] following as [8]:(1)
Similar level 1
: with
k
= 1, fuzzy interval
level 1
including
I
(
c
−
) and
I
(
c
+
). fuzzy interval
level 2
on interval
I
(
c
+
) is
I
(
h
q
c
+
)
≤
I
(
h
q+
1
c
+
) ...
≤
I
(
h
2
c
+
)
≤
I
(
h
1
c
+
)
≤
υ
A
(
c
+
)
≤
I
(
h
1
c
+
)
≤
I
(
h
2
c
+
)
≤
...
≤
I
(
h
p
1
c
+
)
≤
I
(
h
p
c
+
). Meanwhile,we construct partition at
similar level 1
include theequivalence classes following:
S
(
0
) =
I
(
h
p
c
−
);
S
(
c
−
)=
I
(
c
−
) \ [
I
(
h
q
c
−
)
∪
I
(
h
p
c
−
)];
S
(
W
) =
I
(
h
q
c
−
)
∪
I
(
h
q
c
+
);
S
(
c
+
) =
I
(
c
+
) \ [
I
(
h
q
c
+
)
∪
I
(
h
p
c
+
)] and
S
(
1
) =
I
(
h
p
c
+
).We see that except the two end points
υ
A
(
0
) = 0 and
υ
A
(
1
) = 1, representative values
υ
A
(
c
−
),
υ
A
(
W
) and
υ
A
(
c
+
)are inner point corresponding of classes
similar level 1
S
(
c
−
),
S
(
W
) and
S
(
c
+
).(2)
Similar level 2
: with
k
= 2, fuzzy interval
level 2
including
I
(
h
i
c
+
) and
I
(
h
i
c

) with q
≤
i
≤
p. We haveequivalence classes following:
S
(
0
) =
I
(
h
p
h
p
c
−
);
S
(
h
i
c
−
) =
I
(
h
i
c
−
) \ [
I
(
h
q
h
i
c
−
)
∪
I
(
h
p
h
i
c
−
)];
S
(
W
) =
I
(
h
q
h
q
c
−
)
∪
I
(
h
q
h
q
c
+
);
S
(
h
i
c
+
) =
I
(
h
i
c
+
) \ [
I
(
h
q
h
i
c
+
)
∪
I
(
h
p
h
i
c
+
)] and
S
(
1
) =
I
(
h
p
h
p
c
+
), with q
≤
i
≤
p.By the same, we can construct partition equivalenceclasses
level k
at any. However, in fact,
k
≤
4 and it meansthat there is maximum 4 hedges consecutive action onto primary terms
c
−
and c
+
. Precise and fuzzy values will beat the
similar level k
if the representative value of their inthe same class
similar level k
.Hence,
neighborhood level k of
f
uzzy concept
isdetermining following: Assuming partition the class
similar level k
is intervals
S
(
x
1
),
S
(
x
2
), …,
S
(
x
m
).Meanwhile, every fuzzy value
fu
is only and only belongto a similar class. Instance for
S
(
x
i
) and called
neighborhood level k
of
fu
and denoted by ()
k
FRNfu
.
B.
Relation matching on domain of fuzzy attribute value
Based on the concept neighborhood, we give thedefinition of the relation matching between terms in thedomain of the fuzzy attribute value.
Definition 3.1
Let fuzzy class C determine on the set of attributes Aand methods M, a
i
⊆
A. o
1
, o
2
∈
C. We say that
12
..
iki
oaoa
=
and equal
level k
if:(1)
If
12
.,.()
iii
oaoaCDoma
∈
then
12
..
ii
oaoa
=
or existence
()
k
FRNx
such that
12
.,.()
iik
oaoaFRNx
∈
.(2)
If
1
.
i
oa
or
2
.
i
oa
()
i
FDoma
∈
, instance for
1
.
i
oa
then we have to
21
.(.)
iki
oaFRNoa
∈
.(3)
If
1
.
i
oa
,
2
.
i
oa
()
i
FDoma
∈
then
21
(.)(.)
kiki
FRNoaFRNoa
=
.
Definition 3.2
Let fuzzy class C determine on the set of attributes Aand methods M, a
i
⊆
A. o
1
, o
2
∈
C. We say that
12
..
iki
oaoa
≥
if:(1)
If
12
.,.()
iii
oaoaCDoma
∈
then
12
..
ii
oaoa
≥
.(2)
If
1
.
i
oa
and
2
.
i
oa
()
i
FDoma
∈
then we have to
38http://sites.google.com/site/ijcsis/ISSN 19475500