TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol.12, No.4, April 201
4, pp. 3015 ~ 3
0
2
0
DOI: http://dx.doi.org/10.11591/telkomni
ka.v12i4.4774
3015
Re
cei
v
ed Se
ptem
ber 10, 2013; Revi
se
d No
vem
ber
20, 2013; Accepted Decem
ber 3, 201
3
Uncertain Attribute Graph Sub-Graph Isomorphism and
its Determination Algorithm
Yanjun Zhao
*
, Chun
y
i
ng
Zhang
Coll
eg
e of Scie
nce, Heb
e
i Unt
i
ed U
n
ivers
i
t
y
46
Xin
h
u
a
Roa
d
,
T
angsha
n 0
630
09, He
bei,
Chin
a T
e
l: 031
5-25
924
80
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: teacher_
j
sj@
126.com
A
b
st
r
a
ct
T
he exp
e
ctativ
e sub-
gra
ph
is
omor
p
h
is
m of
uncerta
in
attrib
ute gra
p
h
is b
a
sed
on t
he
a
nalysis
of
compl
e
x n
e
tw
ork structure a
n
d
the
char
acte
ristic of
u
n
cert
ain attribute
gr
aph.
T
h
e ex
pe
ctative su
b-gra
p
h
i
s
om
o
r
ph
i
s
m
o
f
u
n
c
e
r
ta
i
n
a
ttrib
u
t
e
g
r
a
p
h
i
s
on
l
y
on
e
th
re
sho
l
d
va
lu
e
a
s
con
s
tra
i
n
t
co
n
d
i
t
io
n
s
. Th
e
me
thod
is si
mp
le, b
u
t t
he c
o
mputati
o
n is
larg
e
a
m
o
unt. T
herefor
e, this p
a
p
e
r br
in
gs in
the
defi
n
i
t
ion
of
sub-
grap
h is
o
m
orp
h
is
m of
unc
ertain
attr
ibute
gr
aph,
an
d ex
pla
i
ns th
e co
nce
p
t. T
hen, this
p
aper
des
ig
ns a
n
d
comes true th
e alg
o
rith
m of
sub-gr
aph is
o
m
or
phis
m
. F
i
n
a
lly, throu
gh t
he exp
e
ri
me
nt
s proves that
sub-
grap
h
iso
m
or
phic
is
b
e
tter than
ex
pec
tative su
b-gr
ap
h, an
d
it a
naly
z
e
s
the
var
i
ati
on
in t
h
e
different thresh
old cas
e
s. T
he research of
sub-gr
aph is
o
m
orph
ism a
l
g
o
ri
t
h
m l
a
id the fo
u
ndati
on for
uncerta
in attrib
ute grap
h sub-
grap
h qu
ery an
d commun
i
ty minin
g
.
Ke
y
w
ords
:
ex
pectative su
b-g
r
aph is
o
m
orp
h
i
s
m,
sub-gr
aph
isomorp
h
is
m, uncerta
in attrib
ute grap
h
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
As a gen
eri
c
data structu
r
e, graph
ca
n modelin
g and
expre
ss the
real worl
d all kind
s of
compl
e
x data
entity and th
e relatio
n
ship
betwe
en the
entities.
The whol
e
social relation
ship can
be ab
stra
cted
as a g
r
ap
h i
n
the field of so
cial
relatio
n
netwo
rk [1-3]. Attribute graph [4
-5] is t
he
grap
h struct
ure
con
s
ide
r
i
ng the verti
c
e
s
and e
d
ges attrib
ute
s
and
relati
onship
s
bet
ween
attributes i
n
the traditio
nal
map ba
se
d on. It can be
better de
scri
bed vari
ou
s
node
s an
d th
e
relation
shi
p
b
e
twee
n attrib
utes in the
so
cial net
work,
and ea
sie
r
to
analyze the
effects of the
s
e
prop
ertie
s
. Howeve
r, vertice
s
and verti
c
e
s
relatio
n
s
and their attri
butes a
r
e p
r
o
bability in a social
netwo
rk
o
r
ot
her co
mplex netwo
rks. Un
certai
n
attr
i
b
ute graph
is
pre
s
ente
d
co
nsid
erin
g to t
h
is
factor, an
d it furthe
r de
scrib
e
s
the soci
al netwo
rk u
n
ce
rtainty.
Sub-g
r
ap
h i
s
omorphi
sm i
s
the
cl
assifi
cation
of
sub
-
graph
to n
e
twork
with t
h
e same
cha
r
a
c
teri
stics, and
re
se
a
r
ch
the
sub
-
grap
h
i
s
omo
r
phism
of dat
a graph i
s
a
more
profou
nd
study on th
e so
cial net
works.
No
w, about the sub
-
g
r
ap
h isom
orphi
sm rese
arch work is
followin
g
. Do
ng Ang
uo [6]
given alg
o
rith
ms fo
r sub-graph i
s
om
orp
h
i
sm in
graph
pattern mi
nin
g
.
The algo
rithm
s
are b
a
sed o
n
algeb
ra the
o
ry by
makin
g
use of de
gree se
que
nce
of a vertex and
eigenvalu
e
of adjace
n
cy m
a
trix. Xie Chunxin [7
] given OES: sub-g
r
aph iso
m
orph
ism verification
algorith
m
. Which
find
a
sub-g
r
a
ph i
s
o
m
orp
h
ism
g
r
aph
by
sea
r
ching
edg
e by
edg
e, a
nd it
ca
n
raise the verification effi
cien
cy by adjusti
n
g
the edge o
r
de
r.L
i
u Bo [8] given desi
gn
and
impleme
n
tation of su
b-g
r
aph isomo
r
p
h
ism d
e
tectio
n algo
rithm b
a
se
d on relat
i
onal mod
e
l. He
prop
oses a
new sub-gra
ph i
s
omo
r
p
h
i
s
m
algo
rithm
nam
ed Rel
a
tional Gra
p
h
Decompo
s
i
t
io
n
Index (RG
D
I), and it ha
s h
i
gher
dete
c
tio
n
effici
en
cy. Fred
DePie
r
o
[9] given an
algorith
m
u
s
i
ng
lengthe
r p
a
th
s to
app
roxim
a
te subg
rap
h
iso
m
or
phi
sm
. The
s
e
su
b-grap
h i
s
om
orphism
alg
o
rit
h
m
desi
gn a
r
e ba
sed o
n
tra
d
itional g
r
ap
h or uncertain
graph. But in the real
com
p
le
x networks, it is
often use
d
to
descri
be by uncertain attri
bute gr
aph [1
0]. Howeve
r, how to de
sig
n
the algo
rith
m
of unce
r
tain
attribute grap
h for un
certa
i
n attri
bute sub-g
r
a
ph iso
m
orp
h
ism, it is an imp
o
rta
n
t
issues.
Attributes fig
u
re
is the
ex
pan
sion
of traditional
gra
p
h
,
an
d it i
s
con
s
id
ere
d
t
he ve
rtex
and edg
e attribute
s
and t
he relatio
n
shi
p
of attr
ibute between the
structu
r
e of the grap
h. And
without
con
s
i
derin
g the
a
ttribute, attrib
ute gr
aph
d
egen
erate
int
o
traditio
nal
figure. So th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 4, April 2014: 3015 – 3
020
3016
traditional fig
u
re is a
spe
c
ial case of tradition g
r
ap
h. The re
se
arch on the
attribute gra
ph
st
ru
ct
ur
e,
it
s
v
a
riou
s cha
r
a
c
t
e
ri
st
ic
s and
ope
rat
i
on
i
s
the theo
ry
co
mpleme
nt an
d inn
o
vation
of
traditional g
r
aph. On un
certain g
r
ap
h sub
-
g
r
ap
h isomorphi
sm h
a
s some research. With t
he
uncertain
att
r
ibute
graph
is put fo
rward,
a
nd
on u
n
certain
attribute
g
r
aph
su
b-g
r
a
ph
isomo
r
p
h
ism
proble
m
is urgent. The
unce
r
tainty
in the data is everywh
e
re, so it is the
particula
rly importa
nt that research u
n
c
ertai
n
a
ttrib
ute gra
ph. T
h
is pa
per
prese
n
ts two
kinds
definition
s
of
sub
-
g
r
ap
h i
s
omo
r
p
h
ism
on un
ce
rtain
t
y attribute
grap
h d
a
ta. In probabili
ty,
uncertainty a
ttribute graph
expectatio
n
s sub
-
g
r
ap
h isomorphi
sm is uncertain
graph
sub
-
g
r
ap
h
isomo
r
p
h
ism
definition dire
ct extension.
The pro
babili
stic si
gnifica
n
c
e of expecta
tions sub-gra
ph
isomo
r
p
h
ism
of unce
r
tain
attribute grap
h is very
intui
t
ive, but it is
huge
comp
utational cost a
nd
matche
s re
su
lts descri
be complex. So, this pap
er
put
s forward u
n
certainty attribute sub-gra
p
h
isomo
r
p
h
ism
definition and
judgment alg
o
rithm. Un
ce
rtain attribute grap
h isom
orphism i
s
usin
g
two thre
shol
d
values to re
place un
certa
i
nty a
ttribute grap
h expe
ctativ
e sub-gra
ph isom
orphi
sm
the singl
e limit thresh
old value.
This pa
pe
r structure mainly
divided into five parts follo
wing.
2. The Unc
e
r
t
ain Attribu
t
e Graph
DEFINITIO
N
2.1. (Uncertai
n
attri
bute
g
r
aph) un
ce
rtai
n attribute
g
r
aph
and
un
certain
attribute graph
are
cal
l
ed
u
n
cert
ain
attri
but
e grap
h. It is denoted
as
((
(
,
),
(
,
)),
)
P
GA
V
V
A
L
V
E
EA
L
E
P
:
((
,
)
,
(
,
)
)
VV
A
L
V
E
E
A
L
E
is an attribute
graph
;
P
is the
prob
ability function of ed
ge, ve
rtex and their attri
bute. (
P
incl
ud
es
:[
0
,
1
]
E
PE
,
:[
0
,
1
]
V
PV
,
:[
0
,
1
]
EA
PE
A
and
:[
0
,
1
]
VA
PV
A
).
3. Expecta
t
iv
e Sub-gra
ph Isomorphis
m
of Unce
rta
i
n Attribu
t
e
Graph
Becau
s
e
of t
he u
n
ce
rtain
attribute g
r
a
p
h
is
di
vided
i
n
to two
aspe
cts to
di
scuss, we
still
discu
ss the u
n
ce
rtain attrib
ute grap
h exp
e
ctativ
e su
b-grap
h iso
m
orphism follo
wi
ng two a
s
pe
cts.
3.1. Expecta
t
iv
e the Sub-graph Isomo
r
phism of Uncertain Attr
ibute Gra
ph
First, we di
scuss the un
certain attribute grap
h
:
NATURE 1.
Acco
rdi
ng to
wheth
e
r
P
the
value is 1,
the un
ce
rtain
attribute g
r
a
ph
divided into:
''
{
,
0(
)
1
0(
)
1
(
0
(
)
1
0(
)
1
)
,
,
}
UC
ii
i
ii
i
GA
GA
G
A
GA
P
e
o
r
P
v
o
r
P
e
and
P
v
e
E
v
V
and
}
,
,
1
)
(
,
1
)
(
,
{
"
"
V
v
E
e
v
P
e
P
GA
GA
GA
GA
i
i
i
i
C
C
GA
is the certai
n
attribute gra
ph set of
GA
.
UC
GA
is the certain a
ttribute grap
h
set o
f
GA
.Two set to meet
UC
C
GA
GA
and
GA
GA
GA
UC
C
.
))
,
(
),
,
(
(
1
1
LE
EA
E
LV
VA
V
GA
C
is a
po
ssible
wo
rld
graph
the u
n
certai
n attribute
g
r
aph
)
,
)),
,
(
),
,
(
((
V
E
P
P
LE
EA
E
LV
VA
V
GA
,and
)
,
(
)
,
(
LV
VA
V
LV
VA
V
,
)
,
(
)
,
(
1
LE
EA
E
LE
EA
E
.
GA
contai
ns
C
GA
,a
nd
s
i
gn
C
GA
GA
.So the probability
of “
GA
cont
ains
C
GA
”
comp
utationa
l formula is:
))
(
1
)(
)
(
1
(
)
(
)
(
)
(
i
GA
GA
i
i
i
GA
C
v
P
e
P
v
P
e
P
GA
GA
P
C
UC
C
(1)
The probability of “
GA
sub-g
r
a
ph isom
orphi
c to
C
GA
” is the fol
l
owin
g formul
a:
12
12
11
2
2
1
2
()
()
(
)
(
,
)
GA
GA
PG
A
G
A
P
GA
G
A
P
G
A
G
A
G
A
G
A
(2)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Un
certai
n Attribute Gra
ph
Sub-G
r
ap
h Isom
or
phi
sm
and its Dete
rm
ination… (Y
a
n
jun Zha
o
)
3017
Functio
n
)
,
(
2
1
GA
GA
va
lues in
{0,1}
. If
1
GA
and
2
GA
sub
-
gra
ph i
s
omo
r
p
h
i
s
m,
)
,
(
2
1
GA
GA
value is 1,
or 0. Appare
n
tly,
)
(
2
1
GA
GA
P
is the
probability expectation of
)
,
(
2
1
GA
GA
. Acco
rdin
g t
o
the exp
e
ct
ative value
we
ca
n b
e
defi
ned fo
r the
e
x
pectation
su
b
-
grap
h iso
m
orphism of un
certain attrib
ute grap
h
.
DEFINITIO
N
3.1. Known
uncer
tainty a
ttribute g
r
ap
h
1
GA
and
2
GA
, and
expe
ctative
threshold val
u
e
]
1
,
0
(
, if and only if
)
(
2
1
GA
GA
P
,
1
GA
is sub-g
r
ap
h isom
orphi
c to
2
GA
. Sign
2
1
GA
GA
.
3.2 The exp
e
c
ta
tiv
e
sub-graph isomo
r
phic of unc
ertain a
ttribu
t
e grap
h
Similarly, we
can
define
the expe
ctative su
b-gra
ph i
s
omo
r
p
h
ism
of uncertai
n
attribute
grap
h
.
NATURE 2.
Acco
rdi
ng to
the a val
u
e
wh
ether
P
is
1, the un
ce
rt
ain attrib
ute
grap
h
divides into mutually disjoint
subsets of
''
{
,
0(
)
1
0(
)
1
(0
(
)
1
0
(
)
1
)
,
,
}
UC
ii
ii
i
i
GA
GA
GA
GA
P
e
a
o
r
P
v
a
or
P
e
a
and
P
v
a
e
a
E
A
v
a
V
A
and
}
,
,
1
)
(
,
1
)
(
,
{
"
"
VA
va
EA
ea
va
P
ea
P
GA
GA
GA
GA
i
i
i
i
C
. Where
C
GA
is th
e
certain
attribute grap
h of
GA
, and
UC
GA
is the uncertain
attribute gra
ph of
GA
,and Two set to meet
UC
C
GA
GA
and
GA
GA
GA
UC
C
.
))
,
(
),
,
(
(
1
1
LE
EA
E
LV
VA
V
GA
C
is a po
ssible
worl
d graph
o
f
unce
r
tain att
r
ibute g
r
ap
h
,and
)
,
(
1
LV
VA
V
LV
VA
V
)
,
(
,
)
,
(
)
,
(
1
LE
EA
E
LE
EA
E
.
GA
contai
ns
C
GA
, and si
gn
s
C
GA
GA
.So the probability of “
GA
contains
C
GA
” is the fo
llowing fo
rmul
a:
()
()
()
(
1
()
)
(
1
(
)
)
UC
CC
C
ii
i
i
G
A
GA
GA
PG
A
G
A
P
e
a
P
va
P
e
a
P
va
(3)
The probability of “
GA
sub-g
r
a
ph isom
orphi
c to
C
GA
” is the fol
l
owin
g formul
a:
)
,
(
)
(
)
(
)
(
2
1
2
2
1
1
2
1
2
1
GA
GA
GA
GA
P
GA
GA
P
GA
GA
P
GA
GA
.
(4)
Functio
n
)
,
(
2
1
GA
GA
va
lues in
{0,1}
. If
1
GA
and
2
GA
sub-graph isomo
r
p
h
ism,
)
,
(
2
1
GA
GA
value is
1, o
r
0. Appa
rent
ly,
)
(
2
1
GA
GA
P
is the
probability exp
e
ctation
of
)
,
(
2
1
GA
GA
. Acco
rdin
g t
o
the exp
e
ct
ative value
we
ca
n b
e
defi
ned fo
r the
e
x
pectation
su
b
-
grap
h iso
m
orphism of un
certain attrib
ute grap
h
.
DEFINITIO
N
3.2. Known u
n
ce
rtainty attribute g
r
ap
h
1
GA
and
2
GA
, and expectative
threshold val
u
e
]
1
,
0
(
. If and only if
)
(
2
1
GA
GA
P
,
1
GA
is su
b-g
r
aph isomo
r
ph
ic to
2
GA
. Sign
2
1
GA
GA
.
3.3. The Expecta
t
iv
e Sub
-
grap
h Isomorphic of Un
certain Attri
bute Gr
aph
DEFINITIO
N
3.3. (the exp
e
ctat
ive sub-grap
h isomo
r
phism
of
un
certain attri
but
e gra
p
h
)
1
P
GA
,
1
P
GA
and expe
ct
ation thre
shold
)
1
,
0
(
are
kn
own
qua
ntity. If and only if
)
(
)
(
2
1
2
1
GA
GA
P
GA
GA
P
,
1
P
GA
is
su
b
-
graph
iso
m
orp
h
ic to
2
P
GA
. Sign
2
1
P
P
GA
GA
.
The alg
o
rith
m of expectat
i
ve sub
-
graph
isomo
r
phi
sm
judging i
s
co
stly. Expectative sub
-
graph
i
s
om
orphi
sm still exist
the problem of
m
a
tching resul
t
s describe
difficulty in t
h
e
appli
c
ation. T
hese pro
b
lem
s
limit the applic
atio
n of expectative sub
-
graph i
s
omo
r
phi
sm.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 4, April 2014: 3015 – 3
020
3018
4. The
-
Sub-g
r
aph Isomor
phism of Un
certain Attri
bute
4.1. The
-
Sub-grap
h Isomorphism of
Unce
rtain Attribute
Grap
h
DEFINITIO
N
4.1. The u
n
ce
rtain attri
but
es
set of
uncertain
a
ttribute grap
h
GA
is.Definition
]
1
,
0
(
2
UC
GA
:
is
er
r
o
r
fu
nc
tio
n
.
F
o
r th
e
an
y s
u
b-
gr
a
p
h
'
GA
of
UC
GA
, error f
u
nction
value is the followin
g
:
0
))
(
1
))(
(
1
(
1
)
(
'
'
GA
GA
i
i
v
P
e
P
GA
'
'
GA
GA
(5)
DEFINITIO
N
4.2. The
un
certain
attribut
es set
of
u
n
c
ertai
n
attrib
ute
g
r
ap
h
UC
GA
is
kno
w
.
GA
is any sub-graph of
GA
. Definition
]
1
,
0
(
2
:
I
GA
is stre
ngth fun
c
tion. For the
any
s
u
b-
gr
a
p
h
GA
of
GA
, strength fun
c
tion is the fol
l
owin
g:
GA
GA
GA
i
i
e
P
v
p
GA
)
(
)
(
)
(
(6)
DEFINITIO
N
4.3. Un
ce
rtain attribut
e
g
r
ap
h
1
GA
and
2
GA
are
kno
w
n
qua
ntity.
Hypothe
si
s t
he u
n
certai
n
attribute
of
1
GA
is
UC
GA
, if the p
r
e
s
en
ce
UC
GA
GA
'
and
mappin
g
function
f
, while s
a
tis
f
ying the following:
(1)
'
1
max
)
(
GA
GA
I
sub
-
grap
h isomo
r
p
h
ic
to
)
(
2
max
GA
I
and
f
is map
p
ing fun
c
tion;
(2)
max
'
)
(
GA
, Consta
nt
max
called erro
r thre
shol
d valu
e,
]
1
,
0
[
max
;
(3)
min
1
1
))
(
)
(
(
V
f
E
f
, Consta
nt
min
called
stren
g
t
h threshold value,
]
1
,
0
[
So unce
r
tain
attribute graph
1
GA
-
sub-g
r
aph isomo
r
p
h
ic to un
cert
ain attribute
grap
h
2
GA
. Sign
2
,
1
GA
GA
.
THEO
RERN 1. If the uncertain attribut
e grap
h
1
GA
is isomorphi
sm to the unce
r
ta
in
attribute gra
p
h
2
GA
in limiting
condition
max
and
min
, then
)
,
(
)
(
min
max
2
1
GA
GA
P
,and
min
max
min
max
)
1
(
)
,
(
.
If the unce
r
tain attribute
grap
h
1
GA
is
-
sub-g
r
a
ph iso
m
orp
h
ism to
uncertai
n
attribute grap
h
2
GA
, then the prob
ability is
)
,
(
)
(
min
max
2
1
GA
GA
P
that the uncertai
n
attribute g
r
ap
h
1
GA
is
-
sub
-
gra
ph isomo
r
phi
sm to un
ce
rt
ain attribute
grap
h
2
GA
in the
real
wo
rld. A
s
sho
w
ing
in
definition
3.1, ex
pectativ
e
su
b-gra
ph isomo
r
p
h
ism
use
s
a sing
le
con
s
tant a
s
t
he limit thre
shold valu
e, a
nd
-
sub-graph
isomo
r
p
h
ism
use
s
two co
nstant
s
max
and
min
as the li
mit thre
shol
d
value to
re
place
. Sub-graph i
s
om
orphi
sm
has
probability
signifi
can
c
e.
It can be expre
s
sed a
s
followin
g
. If the un
certai
n
attribute gra
p
h
1
GA
is sub-
grap
h i
s
omo
r
phic to
a the
uncertain
attri
bute g
r
ap
h
2
GA
and
)
,
(
min
max
is not le
ss than the
desi
r
ed
threshold
, then, u
n
ce
rtain
attrib
ute graph
1
GA
is expectative sub-g
r
a
ph
i
s
o
m
orp
h
ic
to unce
r
tain a
ttribute gra
p
h
2
GA
.
4.2. The
-
S
ub-graph Isomo
r
phism of Uncertain Attr
ibute Gra
ph
DEFINITIO
N
4.4. The un
certain attri
but
es
set of un
certai
n attrib
ute gra
ph
GA
is
}
1
0
,
1
0
,
{
EA
VA
UC
P
P
GA
GA
GA
GA
.Definite
]
1
,
0
(
2
UC
GA
:
is e
r
ror fu
nction.
For the
any sub
-
g
r
ap
h
GA
of
UC
GA
, error func
tion valu
e is
the following:
0
))
(
1
))(
(
1
(
1
)
(
GA
GA
i
i
va
P
ea
P
GA
(
GA
GA
(7)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Un
certai
n Attribute Gra
ph
Sub-G
r
ap
h Isom
or
phi
sm
and its Dete
rm
ination… (Y
a
n
jun Zha
o
)
3019
DEFINITIO
N
4.5. The
un
certai
n attrib
ut
es
set of u
n
ce
rtain
attribute g
r
ap
h
GA
is
kno
w
.
GA
is an
y sub
-
graph
of
GA
. Definition
]
1
,
0
(
2
:
GA
is
s
t
rength func
tion.
For the
any sub
-
g
r
ap
h
GA
of
GA
, s
t
rength func
tion is
the following:
GA
GA
GA
i
i
UC
ea
P
va
p
GA
)
(
)
(
)
(
(8)
DEFINITIO
N
4.6. Unce
rtain attribut
e grap
h
1
GA
and
2
GA
are kno
w
n
quantity.
Hypothe
si
s t
he u
n
certain
attribute
of
1
GA
is
UC
GA
, if th
e
pr
es
en
c
e
UC
GA
GA
'
and
ma
pping
function
f
, while s
a
tis
f
ying the following:
(1)
'
1
max
)
(
GA
GA
I
sub
-
grap
h isomo
r
p
h
ic
to
)
(
2
max
GA
I
and
f
is map
p
ing fun
c
tion;
(2)
max
'
)
(
GA
, Consta
nt
max
called erro
r thre
shol
d valu
e,
]
1
,
0
[
max
;
(3)
min
1
1
))
(
)
(
(
VA
f
EA
f
, Consta
nt
min
called
stren
g
th threshold value,
]
1
,
0
[
So un
certai
n
attribute g
r
aph
1
GA
-
sub
-
g
r
aph i
s
om
orp
h
ic to
un
cert
ain attrib
ute
grap
h
2
GA
. Sign
2
,
1
GA
GA
.
4.3. The Semantics o
f
-
Sub-gra
ph Iso
m
orphism of Uncer
tain Attribu
t
e Gr
ap
h
DEFINITIO
N
4.7. Unce
rtain attribute
graph
1
P
GA
and
2
P
GA
are known quantity.
Hypothe
si
s t
he u
n
certain
attribute
of
1
P
GA
is
UC
P
GA
1
, if the presence
UC
P
P
GA
GA
'
and
mappin
g
function
f
, while s
a
tis
f
ying the following:
(1)
'
1
max
)
(
P
P
GA
GA
I
sub
-
grap
h isomo
r
p
h
ic
to
)
(
2
max
P
GA
I
and
f
is map
p
ing fun
c
tion;
(2)
max
'
'
)
(
)
(
GA
GA
, Consta
nt
max
called erro
r thre
shol
d valu
e,
]
1
,
0
[
max
;
(3)
min
1
1
))
(
)
(
)
(
)
(
(
V
f
E
f
VA
f
EA
f
, Cons
tant
min
calle
d
stren
g
th thre
sh
old
value,
]
1
,
0
[
So un
ce
rtain
attribute
graph
1
P
GA
-
sub-graph i
s
om
orp
h
ic to
un
ce
rtain attrib
ute
grap
h
2
P
GA
. Sign
2
,
1
P
P
GA
GA
.
5. Experimental An
aly
s
is
5.1. The Experimental Da
ta
and the Parameter Sett
ing
Un
certai
n attribute gra
ph d
a
ta inclu
d
e
s
t
he que
ry gra
ph and u
n
certain attribute grap
h.
The expe
rim
ent uses
4 g
r
oup
s q
u
e
r
y grap
h
GA5,
GA8, GA11
and GA1
4
, a
nd ea
ch
gra
ph
numbe
r of vertice
s
is
n
. the
sub
-
g
r
ap
h iso
m
orp
h
ism of
uncertain attri
bute gra
ph d
a
ta use
d
para
m
eters mainly
inclu
d
e
s
)
,
(
max
maz
, and set
t
ing up thre
e
of paramet
ers
are
)
7
.
0
,
2
.
0
(
,
)
7
.
0
,
4
.
0
(
and
)
3
.
0
,
4
.
0
(
.It main
observes
the error threshold a
nd the time of
c
a
lc
ulation how to
cha
nge i
n
th
e sa
me
stre
n
g
th functio
n
circum
stan
ce
s, and
the
strength
th
re
sh
old an
d the ti
me
of calculation
how to ch
an
ge in the sam
e
error thresh
old circu
m
sta
n
ce
s.
5.2. Analy
s
is
of Experime
ntal Re
sults
Thro
ugh
the
experim
ent, it ca
n be
seen
that
the
erro
r threshold i
s
small
e
r
and
sho
r
ter
cal
c
ulatio
n time in the sa
me stren
g
th thre
shol
d circumstan
ce
s from the
analytical Figure 1
;
th
e
intensity thre
shol
d incre
a
ses a
nd the ti
me of
com
p
u
t
ational imple
m
entation th
e longe
r in t
he
same
error t
h
re
shol
d und
er ci
rcum
stan
ce
s. Figur
e 2
is avera
ge e
x
ecution time
of before a
n
d
after optimiza
t
ion sub
-
g
r
ap
h isom
orp
h
ism judgme
n
t algorith
m
. Th
e paramete
r
of experime
n
t is
the second
grou
ps b
e
ca
use it is a repre
s
e
n
ta
tive sample. Th
e execution
time of decision
algorith
m
sh
o
r
tens in the
search o
r
de
r o
p
timizing afte
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 4, April 2014: 3015 – 3
020
3020
Figure 1. The
Time of Com
putational
Implementati
on und
er Th
resh
old
Circum
stan
ce
s
Figure 2.
Average Exe
c
utio
n Time of before
and after O
p
timization Su
b-grap
h
Isomo
r
phi
sm Jud
geme
n
t
Algorithm
6. Results a
nd Prospec
t
s
Un
certai
n attribute gra
ph e
x
pectation
s sub-g
r
a
ph iso
m
orp
h
ism i
s
a kind of ide
a
l state.
But it is
often difficult to
achi
eve in
th
e prac
ti
cal
problem
s. Thi
s
pap
er fo
cu
ses
on
define
d
of
uncertain attri
bute
graph
sub-g
r
a
ph i
s
o
m
orp
h
ism
an
d its dete
r
min
a
tion metho
d
. In setting
the two th
re
shold valu
e,
sub-g
r
a
ph i
s
o
m
orp
h
ism i
s
more
clo
s
e to
the actu
al isomorphi
sm.
The dete
r
min
i
ng algo
rithm
efficiency h
a
s
bee
n great
l
y
improved i
n
the pre
m
ise that thresh
old
can b
e
mod
u
l
ated by.
Sub-graph i
s
om
o
r
phi
sm lay a
foundatio
n for su
b-g
r
a
ph q
uery an
d
comm
unity mining work in-depth devel
o
p
ment in
co
mplex network (So
c
ial
Net
w
ork). Ho
we
ver,
how to
set th
e thre
shol
d to make su
b-grap
h isomo
r
phism i
s
mo
re rea
s
o
nable
and the a
c
tual
situation mo
re matchin
g
, it is
the further
resea
r
ch pro
b
lems.
Referen
ces
[1]
Z
hang Sh
uo,
Gao Hon
g
, Li Jianz
hon
g. Efficient
Quer
y Pr
ocessi
ng on U
n
certai
n Graph
Databas
es.
Chin
ese Jo
urn
a
l of Co
mp
uter
s
. 2009; 32(
10)
: 2066-2
0
7
9
.
[2]
Z
hang
Yin
an,
Z
ou Z
h
a
oni
an,
Li Ji
anz
hon
g.
Algor
ithm for
α
-
β
sub-gr
ap
h Isomorp
h
ism
Probl
em o
n
Uncerta
i
n Grap
h.
Intellig
ent C
o
mputer a
nd A
pplic
atio
ns
. 20
11; 10: 1-3.
[3] Z
hang,
Ch
un
yi
ng,
Guo,
Jin
g
feng, Ch
en, Xi
a
o
.
Res
earc
h
o
n
ra
nd
om w
a
lk
rou
g
h
matchi
n
g
a
l
gor
ith
m
o
f
attribute s
ub-
grap
h.
Interna
t
iona
l Co
nfere
n
ce o
n
Adv
a
nced M
a
teri
al
s and
Com
p
uter Scie
nce
,
ICAMCS. 2011
; 5: 297-30
2.
[4]
JK Che
ng, T
S
Huan
g. A sub
g
rap
h
isom
orp
h
ism al
gorit
hm using r
e
sol
u
ti
on ori
g
i
nal r
e
s
earch
article
.
Pattern Reco
g
n
itio
n
. 198
1; 13(5): 371-
37
9.
[5]
Don
g
An
gu
o,
Gao L
i
n, Z
h
a
o
Jian
ban
g. Al
go
rithms
for su
b-
grap
h is
omorp
h
ism i
n
gr
ap
h
pattern M
i
ni
ng.
Mathe
m
atics i
n
Practice and T
heory
. 20
11; 7:
105-1
11.
[6]
Xi
e Ch
un
xi
n, W
ang W
e
i. OES: sub-gr
ap
h i
s
omorp
h
ism verificati
on al
go
rithm.
Comput
er Engi
neer
in
g
.
201
1; 3: 46-51.
[7]
Liu B
o
, F
a
n
g
Bin, Z
h
a
ng S
h
i
y
on
g, Li
Z
h
il
in
. Desig
n
and
i
m
pleme
n
tatio
n
of sub
g
ra
ph
i
s
omorp
h
ism
detectio
n
al
gori
t
hm based o
n
r
e
lati
ona
l mod
e
l
.
Comp
uter En
gin
eeri
n
g
. 20
1
1
; 11: 39-4
3
.
[8]
F
r
ed DePi
ero,
David Kr
out. An alg
o
rithm us
i
ng
le
ngth-r p
a
ths to appr
o
x
im
ate sub
g
rap
h
i
s
omorp
h
ism.
Pattern Recognition Letters
. 200
3; 24(1
–3): 33-4
6
.
[9]
Z
hou Ao
yi
ng, J
i
n Ch
eqi
ng, W
ang Gu
oren. A
Su
rve
y
on th
e
Manag
eme
n
t of Uncertai
n D
a
ta.
Chin
es
e
Journ
a
l of Co
mputers
. 200
9; 1
:
1-16.
0.00
0.50
1.00
1.50
2.00
2.50
5
8
11
14
查询
图规模
判定时间
(秒)
第
1组参数
第
2组参数
第
3组参数
0
.00
0
.50
1
.00
1
.50
5
8
11
14
查询
图规
模
判定时间(
秒)
未改
进算
法
改进
算法
Evaluation Warning : The document was created with Spire.PDF for Python.