T
E
L
KO
M
N
I
KA
T
e
lec
om
m
u
n
icat
ion
,
Com
p
u
t
i
n
g,
E
lec
t
r
on
ics
an
d
Cont
r
ol
Vol.
18
,
No.
1
,
F
e
br
ua
r
y
2020
,
pp.
217
~
227
I
S
S
N:
1693
-
6930,
a
c
c
r
e
dit
e
d
F
ir
s
t
G
r
a
de
by
Ke
me
nr
is
tekdikti
,
De
c
r
e
e
No:
21/E
/KP
T
/2018
DO
I
:
10.
12928/
T
E
L
KO
M
NI
KA
.
v18i1.
14874
217
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
L
at
e
n
t
s
e
m
a
n
t
ic
a
n
al
ysi
s an
d
c
osi
n
e
si
m
il
ar
ity
f
or
h
a
d
it
h
se
ar
c
h
e
n
gi
n
e
Wah
yu
d
in
Dar
m
a
lak
s
an
a
1
,
Ce
p
y
S
lam
e
t
2
,
Wil
d
an
B
u
d
iawan
Z
u
l
f
ik
ar
3
,
I
m
am
F
ah
m
i
F
a
d
il
lah
4
,
Dian
S
a’
ad
il
lah
M
aylawa
t
i
5
,
Hap
id
Ali
6
1
D
ep
ar
t
men
t
o
f
l
l
mu
H
a
d
i
s
,
U
IN
S
u
n
a
n
G
u
n
u
n
g
D
j
a
t
i
B
an
d
u
n
g
,
In
d
o
n
e
s
i
a
2,
3,
4,
5
D
e
p
art
me
n
t
o
f
I
n
fo
rma
t
i
c
s
,
U
IN
Su
n
an
G
u
n
u
n
g
D
j
a
t
i
Ban
d
u
n
g
,
In
d
o
n
es
i
a
2
D
ep
ar
t
men
t
o
f
In
f
o
mat
i
o
n
an
d
C
o
mmu
n
i
ca
t
i
o
n
T
e
ch
n
o
l
o
g
y
,
A
s
i
a
e
U
n
i
v
er
s
i
t
y
,
Mal
a
y
s
i
a
5
Facu
l
t
y
o
f
I
n
fo
rma
t
i
o
n
an
d
Co
mm
u
n
i
cat
i
o
n
T
ech
n
o
l
o
g
y
,
U
n
i
v
er
s
i
t
i
T
e
k
n
i
k
a
l
Mal
a
y
s
i
a
Mel
ak
a,
Mal
a
y
s
i
a
6
Facu
l
t
y
o
f
T
arb
i
y
a
h
an
d
E
d
u
cat
i
o
n
,
U
IN
Su
n
an
G
u
n
u
n
g
D
j
a
t
i
Ba
n
d
u
n
g
,
In
d
o
n
e
s
i
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Aug
14
,
2019
R
e
vis
e
d
De
c
5
,
20
19
Ac
c
e
pted
De
c
22
,
20
19
Search
en
g
i
n
e
t
ech
n
o
l
o
g
y
w
a
s
u
s
e
d
t
o
fi
n
d
i
n
fo
rma
t
i
o
n
as
n
ee
d
ed
ea
s
i
l
y
,
q
u
i
ck
l
y
an
d
effi
c
i
en
t
l
y
,
i
n
c
l
u
d
i
n
g
i
n
s
earc
h
i
n
g
t
h
e
i
n
fo
rma
t
i
o
n
a
b
o
u
t
t
h
e
h
ad
i
t
h
w
h
i
ch
w
a
s
a
s
eco
n
d
g
u
i
d
e
l
i
n
e
o
f
l
i
fe
fo
r
m
u
s
l
i
m
b
es
i
d
e
s
t
h
e
H
o
l
y
Q
u
r'an
.
T
h
i
s
s
t
u
d
y
w
as
a
i
m
t
o
b
u
i
l
d
a
s
p
eci
a
l
i
ze
d
s
earch
en
g
i
n
e
t
o
f
i
n
d
i
n
f
o
rmat
i
o
n
ab
o
u
t
a
co
m
p
l
e
t
e
an
d
el
ev
e
n
h
a
d
i
t
h
i
n
In
d
o
n
e
s
i
a
n
l
an
g
u
a
g
e.
In
t
h
i
s
re
s
earch
,
s
earch
e
n
g
i
n
e
s
w
o
rk
e
d
b
y
u
s
i
n
g
l
at
e
n
t
s
ema
n
t
i
c
an
a
l
y
s
i
s
(L
SA
)
an
d
co
s
i
n
e
s
i
m
i
l
ar
i
t
y
b
as
e
d
o
n
t
h
e
k
e
y
w
o
rd
s
e
n
t
ere
d
.
T
h
e
L
SA
an
d
co
s
i
n
e
s
i
mi
l
ari
t
y
me
t
h
o
d
s
w
ere
u
s
ed
i
n
fo
rm
i
n
g
s
t
r
u
ct
u
r
ed
rep
res
e
n
t
a
t
i
o
n
s
o
f
t
ex
t
d
at
a
as
w
e
l
l
as
cal
c
u
l
a
t
i
n
g
t
h
e
s
i
m
i
l
ar
i
t
y
o
f
t
h
e
k
e
y
w
o
r
d
t
ex
t
en
t
ere
d
w
i
t
h
h
ad
i
t
h
t
ex
t
d
at
a,
s
o
t
h
e
h
a
d
i
t
h
i
n
fo
rma
t
i
o
n
w
a
s
i
s
s
u
e
d
i
n
acco
rd
a
n
ce
w
i
t
h
w
h
at
w
as
s
earc
h
ed
.
Bas
ed
o
n
t
h
e
res
u
l
t
s
o
f
t
h
e
t
e
s
t
co
n
d
u
ct
e
d
5
0
t
i
mes
,
i
t
i
n
d
i
cat
e
d
t
h
a
t
t
h
e
L
SA
an
d
co
s
i
n
e
s
i
mi
l
ari
t
y
h
ad
a
s
u
c
ces
s
rat
e
i
n
fi
n
d
i
n
g
h
i
g
h
h
a
d
i
t
h
i
n
f
o
rmat
i
o
n
w
i
t
h
an
av
erag
e
r
ecal
l
v
al
u
e
w
a
s
8
7
.
8
3
%
,
al
t
h
o
u
g
h
fro
m
al
l
i
n
f
o
rmat
i
o
n
o
b
t
a
i
n
e
d
l
e
v
e
l
o
f
p
reci
s
i
o
n
h
a
d
i
t
h
w
as
fo
u
n
d
s
ema
n
t
i
cal
l
y
n
o
t
man
y
,
i
t
w
as
i
n
d
i
ca
t
ed
b
y
t
h
e
av
erag
e
p
reci
s
i
o
n
v
al
u
e
w
as
3
6
.
2
5
%
.
K
e
y
w
o
r
d
s
:
C
os
ine
s
im
il
a
r
it
y
Ha
dit
h
L
a
tent
s
e
mantic
a
na
lys
i
s
S
e
a
r
c
h
e
ngine
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e
.
C
or
r
e
s
pon
din
g
A
u
th
or
:
Dia
n
S
a
’
a
dil
lah
M
a
ylaw
a
ti
,
D
ep
ar
t
men
t
o
f
In
f
o
rmat
i
cs
,
U
IN
Su
n
an
G
u
n
u
n
g
D
j
at
i
Ban
d
u
n
g
,
In
d
o
n
e
s
i
a
.
E
mail:
dians
m@ui
ns
gd.
a
c
.
id
1.
I
NT
RODU
C
T
I
ON
S
e
a
r
c
h
e
ngine
be
c
a
mes
one
of
f
unc
ti
ons
or
the
mo
s
t
im
por
tant
tool
on
in
f
or
mation
s
ys
tem
s
pe
c
ially
on
-
li
ne
s
ys
tem
[
1]
.
S
e
a
r
c
h
e
ngine
tec
hnology
gives
it
e
a
s
y
f
or
s
ys
tem
us
e
r
to
ge
t
the
in
f
or
mation
quickly
[
2]
.
Google
is
one
of
c
a
pa
ble
s
e
a
r
c
h
e
ngi
ne
s
but
it
s
ti
ll
ha
s
li
mi
tations
in
a
na
lyzing
the
c
o
ntent
a
nd
mea
ning
of
s
e
a
r
c
h
r
e
s
ult
s
[
3]
.
Along
with
a
dva
nc
e
d
da
te
r
e
gulation
on
the
int
e
r
ne
t,
s
e
a
r
c
h
e
ngines
r
e
quir
e
s
pe
e
d
a
nd
a
c
c
ur
a
c
y
in
r
e
lea
s
ing
r
e
s
ult
s
in
li
ne
with
e
xpe
c
tations
today.
T
he
s
e
a
r
c
h
f
unc
ti
on
be
c
omes
i
mpor
tant
thi
ng
in
ge
tt
ing
inf
o
r
mation
e
a
s
il
y
a
nd
qu
ickly.
H
owe
ve
r
,
not
a
ll
s
e
a
r
c
h
e
ngines
a
r
e
de
voted
to
f
in
d
c
e
r
tain
inf
or
mation
pr
e
c
is
e
ly
a
nd
a
c
c
ur
a
tely.
I
n
thi
s
s
tudy,
a
s
e
a
r
c
h
e
ngine
that
wa
s
buil
t
s
pe
c
if
ica
ll
y
to
ge
t
in
f
or
mation
a
bout
the
h
a
dit
h
in
a
c
c
or
da
nc
e
with
us
e
r
ne
e
ds
.
W
he
r
e
,
the
h
a
dit
h
is
the
s
e
c
ond
im
po
r
tant
s
our
c
e
of
law
f
or
M
us
li
ms
a
f
ter
the
Holy
Qur
'a
n
[
4,
5
]
.
O
f
c
our
s
e
,
the
ge
ne
r
a
ted
h
a
dit
h
inf
or
mation
mus
t
ha
nd
in
ha
nd
with
ne
e
de
d
r
e
quir
e
ments
.
T
he
r
e
f
or
e
,
s
e
a
r
c
h
e
ngines
t
ha
t
a
r
e
buil
t
ne
e
d
to
c
ons
ider
the
s
e
mantics
whe
a
ther
f
r
om
the
input
ted
ke
ywor
ds
or
the
h
a
dit
h
da
ta
whic
h
is
s
a
ve
d
in
the
s
ys
tem.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
1
,
F
e
br
ua
r
y
2020
:
217
-
227
218
Ha
dit
h
c
oll
e
c
ti
on
in
the
f
o
r
m
o
f
text
r
e
quir
e
s
c
e
r
tain
pr
oc
e
s
s
e
s
s
o
that
the
mea
ning
of
the
text
is
maintaine
d
[
6]
.
S
tar
ti
ng
f
r
om
p
r
e
pa
r
ing
uns
tr
u
c
tur
e
d
text
da
ta
int
o
s
tr
uc
tur
e
d
da
ta
[
7,
8
]
.
S
t
r
uc
tur
e
d
r
e
pr
e
s
e
ntation
of
text
c
a
n
be
us
e
d
in
the
ne
xt
pr
oc
e
s
s
e
s
both
in
inf
o
r
mation
r
e
tr
ieva
l
(
I
R
)
a
nd
text
mi
ning
[
9
]
.
I
n
the
s
tudy
of
obtaine
d
inf
o
r
mation
s
e
a
r
c
h
e
ngine,
it
us
e
s
the
inf
or
mation
r
e
tr
ieva
l
(
I
R
)
tec
hnique
by
c
ombi
ning
the
late
nt
s
e
mantic
a
na
lys
is
a
lgor
it
h
m
a
nd
c
os
ine
s
im
il
a
r
it
y
.
I
n
c
ontr
a
s
t
to
text
mi
ni
ng
whe
r
e
the
r
e
s
ult
s
obtaine
d
f
r
om
the
s
ys
tem
a
r
e
not
c
lea
r
ye
t,
I
R
pr
oduc
e
s
inf
o
r
mation
that
ha
s
a
c
tually
be
e
n
known
it
s
f
or
m,
be
c
a
us
e
it
is
the
s
a
me
a
s
the
c
oll
e
c
ti
on
of
da
ta
he
ld
[1
0
–
1
2
]
.
I
nf
or
mation
r
e
tr
ieva
l
(
I
R
)
is
us
e
d
to
c
onne
c
t
r
e
lations
hips
be
twe
e
n
lar
ge
text
da
ta
c
oll
e
c
ti
ons
a
c
c
or
ding
to
ke
ywor
ds
.
T
he
pa
r
ts
of
I
R
incl
ude
:
−
T
e
xt
o
pe
r
a
ti
ons
(
ope
r
a
ti
ons
of
text)
whic
h
include
the
s
e
lec
ti
on
of
wor
ds
in
ke
ywor
ds
o
r
doc
u
ments
(
ter
m
s
e
lec
ti
on)
in
the
t
r
a
ns
f
or
mation
of
doc
ument
s
or
ke
ywor
ds
be
c
ome
ter
m
indexe
s
(
index
o
f
wo
r
ds
)
.
−
Que
r
y
f
or
mul
a
ti
on
(
f
or
mul
a
ti
on
of
ke
ywor
ds
)
that
gives
a
s
tanda
r
d
to
the
wo
r
d
indexe
s
of
ke
ywor
d.
−
R
a
nking
(
r
a
nking)
,
look
f
or
doc
uments
that
a
r
e
r
e
leva
nt
to
ke
ywor
ds
a
nd
a
r
r
a
nge
the
doc
uments
a
c
c
or
ding
to
their
c
ompatib
il
it
y
with
ke
ywor
ds
.
−
I
nde
xing
(
indexing)
,
buil
d
a
da
ta
ba
s
e
of
indexe
s
f
r
om
doc
ument
c
oll
e
c
ti
ons
.
F
i
r
s
tl
y,
it
is
c
a
r
r
ied
be
f
or
e
s
e
a
r
c
hing
doc
uments
.
I
R
s
ys
tem
a
c
c
e
pts
ke
ywor
ds
f
r
om
us
e
r
s
,
then
r
a
nks
doc
uments
on
c
oll
e
c
ti
ons
ba
s
e
d
on
thei
r
c
ompatibi
li
ty
with
ke
ywor
ds
.
T
he
r
e
s
ult
of
r
a
nk
w
hich
is
given
to
us
e
r
s
is
doc
ument
s
ba
s
e
d
on
the
s
y
s
tem
a
r
e
r
e
leva
nt
to
ke
ywor
ds
.
B
ut
the
r
e
leva
nc
e
of
doc
uments
to
a
ke
ywor
d
is
a
s
ubje
c
ti
ve
judg
ment
a
nd
it
is
inf
luenc
e
d
by
many
f
a
c
tor
s
s
uc
h
a
s
topi
c
s
,
ti
mi
ng
,
s
our
c
e
s
of
i
nf
or
mation
a
nd
the
objec
ti
ve
o
f
us
e
r
s
.
L
a
tent
s
e
mantic
a
na
lys
is
a
lgor
it
hm
is
wide
ly
us
e
d
i
n
pr
oc
e
s
s
ing
text
da
ta
by
s
e
mantics
a
ppr
oa
c
he
s
s
o
the
mea
ning
of
the
text
is
maintaine
d.
L
a
tent
s
e
mantic
a
na
lys
is
c
a
n
be
us
e
d
not
only
f
or
text
s
umm
a
r
iza
ti
on
we
ll
[1
3
–
15
]
,
c
he
c
king
plagia
r
is
m
[1
5
]
,
a
nd
a
utom
a
ti
c
a
ll
y
e
va
luating
e
s
s
a
y
s
[1
6
]
,
of
c
our
s
e
it
c
a
n
a
ls
o
be
us
e
d
f
or
s
e
a
r
c
hing.
L
a
tent
s
e
mantic
a
na
lys
is
c
ompar
e
s
the
e
nter
e
d
text
with
owne
d
text
da
ta
c
oll
e
c
ti
on
b
a
s
e
d
on
ve
c
tor
r
e
pr
e
s
e
ntations
[
17
–
19
]
,
with
r
e
ga
r
d
to
s
e
mantics
a
ppr
oa
c
he
s
to
pr
e
s
e
r
ve
the
mea
ning
of
texts
.
I
n
a
ddit
ion
to
la
tent
s
e
mantic
a
na
lys
is
,
thi
s
ha
dit
h
s
e
a
r
c
h
e
ngine
r
e
s
e
a
r
c
h
a
ls
o
us
e
s
c
os
ine
s
im
il
a
r
it
y
to
s
e
e
the
s
im
il
a
r
it
y
of
text
da
ta
ge
ne
r
a
ted
by
s
e
a
r
c
h
e
ngines
s
o
that
it
c
a
n
b
r
ing
up
text
da
ta
s
e
que
nc
e
s
ba
s
e
d
on
popular
it
y
a
s
top
or
de
r
.
C
os
ine
s
im
il
a
r
it
y
is
one
o
f
the
mos
t
popular
s
im
il
a
r
it
y
c
a
lcula
ti
on
methods
to
be
a
ppli
e
d
to
text
doc
uments
[2
0
]
.
T
he
main
a
dva
ntage
of
the
c
os
ine
s
im
il
a
r
it
y
method
is
that
it
c
a
n’
t
be
a
f
f
e
c
t
by
the
length
a
nd
s
hor
t
of
a
doc
ument.
B
e
c
a
us
e
the
ter
m
va
lue
of
e
a
c
h
doc
ument
is
the
im
por
tant
th
ing.
B
a
s
e
d
on
the
e
xplana
ti
on
of
the
pr
oblem
f
or
mul
a
ti
on
a
bove
,
how
late
nt
s
e
mantic
a
na
lys
is
a
nd
c
os
ine
s
im
il
a
r
it
y
c
a
n
be
im
pleme
nted
in
f
indi
ng
the
h
a
dit
h
text
ba
s
e
d
on
k
e
ywor
ds
e
nter
e
d
c
or
r
e
c
tl
y
on
the
h
a
dit
h
s
e
a
r
c
h
e
n
gi
ne
?
Ar
e
late
nt
s
e
mantic
a
na
lys
is
a
nd
c
o
s
ine
s
im
il
a
r
it
y
in
th
e
s
e
a
r
c
h
e
ngine
c
a
n
f
ind
h
a
dit
h
text
da
ta
that
a
r
e
s
e
a
r
c
he
d
ba
s
e
d
on
ke
ywor
ds
that
a
r
e
e
nter
e
d
c
or
r
e
c
tl
y
a
nd
r
e
leva
nt.
2.
RE
S
E
AR
CH
M
E
T
HO
D
F
igur
e
1
de
s
c
r
ibes
a
c
ti
vit
y
f
low
o
f
th
is
r
e
s
e
a
r
c
h.
Ge
ne
r
a
ll
y
,
th
is
r
e
s
e
a
c
h
us
e
d
I
R
tec
hnique
that
im
pleme
nt
late
nt
s
e
mantic
a
na
lys
is
a
nd
c
os
ine
s
im
il
a
r
it
y
a
lgor
it
h
m
f
o
r
pr
oduc
ing
inf
o
r
mation
o
f
ha
dit
hs
ba
s
e
d
on
input
ke
ywor
ds
.
T
he
a
c
ti
vit
y
be
gin
f
r
om
input
ing
the
ke
ywor
ds
(
c
a
n
be
in
the
f
or
m
of
wor
ds
,
p
hr
a
s
e
,
or
s
e
ntenc
e
)
,
the
input
ke
ywor
d
will
be
pr
oc
e
s
s
e
d
i
n
text
pr
e
-
pr
oc
e
s
s
ing
pha
s
e
to
c
lea
n
t
e
xt
da
ta.
T
he
n,
L
S
A
a
gor
it
hm
wi
ll
be
c
onduc
ted
to
c
r
e
a
te
te
r
m
doc
ument
matr
ix
a
nd
ge
t
the
ve
c
tor
va
lue
o
f
e
a
c
h
d
oc
ume
nt.
L
a
s
t,
the
s
im
il
a
r
it
y
o
f
input
ke
ywor
ds
a
nd
ha
dit
h
da
ta
c
o
ll
e
c
ti
on
will
be
c
ounted
us
ing
c
os
ine
s
im
il
a
r
it
y
.
F
igur
e
1
.
R
e
s
e
a
r
c
h
Ac
ti
vit
ies
Start
Input
keywords
Text Pre-pr
ocessin
g:
1. Token
izing
2. Casefolding
3. Filtering/Clean
ing Data
4. Remo
ving Sto
pwords
5. Ste
mmin
g
Hadith Data
Collectio
n
Term
Diction
ary
Conduc
ting Late
nt Sem
antic A
nalysis:
1. Creating term docu
men
t matrix
2. Calculat
ing Sin
gular Value Deco
mposition
3. Calculat
ing vect
or value from each docu
men
t
Calculatin
g Cosine Similarity v
alue
Information
of Hadith
End
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
L
atent
s
e
mantic
analys
is
and
c
os
ine
s
imil
ar
it
y
for
hadit
h
s
e
ar
c
h
e
ngine
(
W
ahy
udin
Dar
malak
s
ana
)
219
2.
1
.
L
at
e
n
t
s
e
m
an
t
ic
an
alys
is
(
L
S
A)
L
a
tent
s
e
mantic
a
na
ly
s
is
is
a
n
a
lgebr
a
ic
me
thod
that
e
xtr
a
c
ts
hidden
s
e
mantic
s
tr
uc
tur
e
s
f
r
om
wor
ds
a
nd
s
e
ntenc
e
s
[2
1
]
.
L
a
tent
s
e
mantic
a
na
lys
is
a
lgor
it
hm
is
one
of
the
de
ve
lopm
e
nt
a
lgor
it
hms
in
the
f
ield
of
inf
or
mation
r
e
tr
ieva
l
that
is
a
ble
to
c
oll
e
c
t
a
lar
ge
n
umber
o
f
doc
uments
in
a
da
ta
ba
s
e
a
nd
c
onne
c
t
r
e
la
ti
ons
hips
be
twe
e
n
doc
uments
by
matc
hing
the
given
inpu
t.
T
h
e
main
f
unc
ti
on
o
f
thi
s
late
nt
s
e
mantic
a
na
ly
s
is
is
to
c
a
lcula
te
the
s
im
il
a
r
it
y
of
a
text
da
ta
by
c
ompar
ing
ve
c
tor
r
e
pr
e
s
e
ntations
f
r
om
other
text
da
ta
[
15
]
.
T
h
e
r
e
s
ult
s
of
late
nt
s
e
mantic
a
na
lys
is
r
e
pr
e
s
e
nt
text
da
ta
c
o
ntextua
ll
y
a
nd
s
e
mantic
that
g
ives
text
mea
nings
[2
1
,
2
2
]
.
T
he
e
va
luation
by
us
ing
the
late
nt
s
e
mantic
a
na
lys
is
method
f
oc
us
e
s
on
wo
r
ds
in
wr
it
ing
without
c
ons
ider
ing
to
the
or
de
r
of
wor
ds
a
nd
gr
a
mm
a
r
in
w
r
it
teng
te
xts
s
o
that
a
s
e
nt
e
nc
e
is
a
s
s
e
s
s
e
d
ba
s
e
d
on
the
k
e
y
wor
ds
include
in
the
s
e
ntenc
e
[2
3
]
.
B
a
s
ica
ll
y,
late
nt
s
e
mantic
a
na
lys
is
e
xtr
a
c
ts
inf
or
mation
f
r
om
pa
t
ter
ns
or
c
oll
e
c
ti
ons
of
wor
ds
that
of
ten
a
ppe
a
r
s
im
ult
a
n
e
ous
ly
in
dif
f
e
r
e
nt
s
e
ntenc
e
s
.
I
f
the
s
e
ntenc
e
c
ontains
a
c
oll
e
c
ti
on
of
wor
ds
that
of
ten
a
ppe
a
r
in
lar
ge
n
umber
s
,
the
s
e
ntenc
e
ha
s
s
e
mantic
or
s
a
f
e
mea
n
ing
[2
1
]
.
Ge
ne
r
a
ll
y,
the
s
teps
of
late
nt
s
e
mantic
a
na
ly
s
i
s
that
a
r
e
us
e
d
f
or
text
da
ta,
a
mong
other
s
[2
4
]
:
text
pr
e
-
pr
oc
e
s
s
ing,
c
r
e
a
ti
ng
ter
m
of
doc
ument
mat
r
ix,
c
a
lcula
ti
ng
s
ingul
a
r
va
lue
de
c
ompos
it
ion
(
S
VD
)
a
nd
c
a
lcula
ti
ng
ve
c
tor
va
lue
f
o
r
e
a
c
h
doc
ument
2.
1.
1.
T
e
xt
p
re
-
p
r
oc
e
s
s
in
g
T
he
text
pr
e
-
pr
oc
e
s
s
ing
s
tage
is
the
s
tage
to
pr
e
p
a
r
e
text
da
ta
whic
h
is
uns
tr
uc
tur
e
d
da
ta
be
c
omes
a
s
tr
uc
tur
e
d
da
ta
r
e
pr
e
s
e
ntation
[
7,
2
5
,
2
6
]
.
T
he
pr
oc
e
s
s
s
tar
ts
f
r
om
tokeniz
a
ti
on,
de
lete
s
r
e
gular
e
xp
r
e
s
s
ions
,
de
lete
s
non
letter
c
ha
r
a
c
ter
s
,
de
lete
s
s
top
wor
ds
,
a
nd
s
temmi
ng
.
I
n
f
a
c
t,
i
f
ne
e
de
d,
it
is
c
a
r
r
ied
out
a
s
pe
c
ial
pr
oc
e
s
s
to
ha
ndle
na
tur
a
l
langua
ge
s
c
ontaine
d
in
text
da
ta,
s
uc
h
a
s
;
a
bbr
e
viations
,
s
lang,
r
e
gional
langua
ge
s
,
a
nd
other
na
tu
r
a
l
langua
ge
s
.
T
he
dis
c
us
s
ion
r
e
ga
r
ding
text
pr
e
-
pr
oc
e
s
s
ing
will
be
e
xplaine
d
f
ur
ther
in
s
e
c
ti
on
3.
2.
2.
1.
2.
Cr
e
at
in
g
t
e
r
m
o
f
d
oc
u
m
e
n
t
m
at
r
ix
Af
ter
c
a
r
r
ied
ou
t
the
pr
e
-
pr
oc
e
s
s
ing
s
tage
in
the
text
da
ta,
then
the
te
r
m
of
doc
ument
matr
ix
is
c
ons
tr
uc
ted
by
plac
ing
the
wor
d
r
e
s
ult
of
the
s
temmi
ng
(
ter
m
)
pr
oc
e
s
s
int
o
the
r
ow
.
T
h
is
matr
ix
is
c
a
ll
e
d
the
ter
m
of
doc
ument
matr
ix.
E
a
c
h
r
ow
r
e
pr
e
s
e
nts
a
unique
wor
d,
whi
le
e
a
c
h
c
olum
n
r
e
pr
e
s
e
nts
the
obtaine
d
wor
d
s
our
c
e
.
T
he
s
our
c
e
o
f
the
wor
d
c
a
n
be
s
e
ntenc
e
s
,
pa
r
a
gr
a
phs
,
or
a
ll
pa
r
ts
of
the
text.
T
he
e
xa
mpl
e
s
of
the
ter
m
of
doc
ument
matr
ix
c
a
n
be
s
e
e
n
i
n
T
a
ble
1
(
that
pr
e
s
e
nted
with
I
ndone
s
ian
la
ngua
ge
)
.
On
the
T
a
ble
1
,
the
f
i
r
s
t
r
ow
r
e
pr
e
s
e
nts
the
wor
d
ha
s
pa
s
s
e
d
the
pr
e
pr
oc
e
s
s
unti
l
the
s
temmi
ng
p
r
oc
e
s
s
is
c
a
ll
e
d
s
temmed
ter
m
(
the
wor
d
a
s
ter
m
1
,
ter
m
2,
e
tc
.
)
,
a
nd
the
c
olum
n
r
e
pr
e
s
e
nts
th
e
c
ontext
,
na
mely
the
text.
T
he
va
lue
is
loca
ted
in
e
a
c
h
c
e
ll
on
the
table
s
hows
how
the
number
of
ti
mes
in
a
te
r
m
a
ppe
a
r
s
in
a
do
c
ument.
F
or
ins
tanc
e
,
the
te
r
m
1
a
ppe
a
r
s
1
ti
me
a
t
the
f
i
r
ts
doc
ument,
a
nd
a
ppe
a
r
s
2
ti
mes
a
t
the
s
e
c
ond
doc
ument,
but
the
te
r
m
1
doe
s
not
a
ppe
a
r
a
t
thi
r
d
doc
ument,
a
nd
s
o
on.
T
a
ble
1
.
M
a
tr
ix
e
xa
mpl
e
f
or
ter
m
o
f
doc
ument
W
or
d
Do
c
1
Do
c
2
Do
c
3
ja
ngan
(
do not)
1
1
0
k
al
ia
n
(
you)
1
1
0
dus
ta
(
li
e
)
1
1
1
at
as
(
on be
ha
lf
)
1
1
1
nam
a
(
na
me
)
1
1
1
ni
s
c
ay
a
(
s
ur
e
ly
)
1
0
0
m
as
uk
(
e
nt
e
r
)
1
1
0
ne
r
ak
a
(
th
e
he
ll
)
1
1
1
s
ungguh
(
a
c
tu
a
ll
y)
0
1
0
s
e
ngaj
a
(
e
xpr
e
s
s
ly
)
0
0
1
te
m
pat
(
pl
a
c
e
)
0
0
1
duduk
(
s
e
a
t)
0
0
1
he
ndak
(
s
houl
d)
0
0
1
2.
1.
3.
Calc
u
lat
in
g
s
in
gu
lar
valu
e
d
e
c
om
p
s
it
ion
a
n
d
ve
c
t
or
valu
e
f
or
e
ac
h
d
oc
u
m
e
n
t
S
ingul
a
r
va
lue
de
c
ompos
it
ion
S
VD
is
a
li
ne
a
r
a
l
ge
br
a
theor
e
m
whic
h
c
a
n
s
pli
t
ter
m
of
doc
ument
matr
ix
int
o
th
r
e
e
ne
w
matr
ice
s
,
thos
e
a
r
e
:
or
thogon
a
l
matr
ix
or
lef
t
s
ingul
a
r
ve
c
tor
mat
r
ix
(
U
)
,
diagon
a
l
matr
ix
or
s
ingul
a
r
va
lue
mat
r
ix
(
S
)
,
a
nd
tr
a
ns
pos
e
of
or
thogonal
matr
ix
o
r
r
ight
s
ingul
a
r
mat
r
ix
(
V)
[2
7
–
29]
,
f
or
mul
a
ted
by
(
1)
that
i
ll
us
tr
a
ted
in
F
igur
e
2
.
A
=
US
V
T
(
1)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
1
,
F
e
br
ua
r
y
2020
:
217
-
227
220
F
igur
e
2
.
S
VD
I
ll
us
tr
a
ti
on
of
(
1)
[3
0
]
T
he
f
or
mul
a
(
1)
is
obtaine
d
f
r
om
the
U
matr
ix
whic
h
is
a
matr
ix
of
m
x
k
s
ize
a
nd
a
matr
ix
V
of
n
x
k
s
ize
,
a
s
il
lus
tr
a
ted
in
F
igur
e
1,
U
a
nd
V
wh
ich
ha
ve
or
thogonal
c
olum
ns
s
o
that
it
c
a
be
va
li
d:
UT
U
=
VT
V
=
1
(
2)
a
nd
S
is
a
d
iagona
l
matr
ix
of
k
x
k
s
ize
.
T
he
c
on
tents
on
the
main
diagona
l
of
the
S
matr
ix
a
r
e
s
in
gular
of
the
A
matr
ix
.
T
he
r
e
s
ult
s
o
f
the
S
VD
c
a
n
be
b
e
tt
e
r
unde
r
s
tood
if
A
mat
r
ix
is
wr
it
ten
with
a
dif
f
e
r
e
nt
int
e
r
pr
e
tation.
I
f
1
,
2
,
…
,
a
r
e
c
olum
n
ve
c
tor
s
of
the
U
mat
r
ix,
1
,
2
,
…
a
r
e
e
ntr
ies
in
the
main
diagona
l
of
the
S
matr
ix,
a
nd
1
,
2
,
…
a
r
e
c
olum
n
ve
c
tor
s
of
V
matr
ix
,
A
matr
ix
c
a
n
be
w
r
it
ten
a
s
s
hown
in
(
3)
.
=
∑
=
1
(
3)
w
he
r
e
the
va
lue
o
f
σ1
is
f
o
r
1
,
f
or
i
=
1,
2,
.
.
.
,
k
,
on
(
3)
it
is
s
or
ted
f
r
om
the
la
r
ge
s
t
to
the
s
malles
t.
I
f
s
ome
big
va
lues
1
a
r
e
take
n
a
nd
a
s
mall
(
ne
a
r
z
e
r
o)
σ_
(
1)
va
lue
is
dis
c
a
r
de
d,
we
ge
t
a
n
a
ppr
oxi
mation
f
r
om
good
A
va
lue
.
S
o,
by
us
ing
S
VD
,
a
matr
ix
c
a
n
be
wr
it
ten
a
s
a
s
um
o
f
the
c
omponents
(
1
f
or
i
=
1
,
2,
…,
k)
,
a
nd
it
s
we
ight
is
the
s
ingul
a
r
va
lue
(
1
,
f
or
i
=
1,
2,
…
k
,
a
r
e
take
n
f
r
om
the
f
o
r
mul
a
of
(
4
)
[
3
0
].
A
=
[
1
,
2
,
…
,
]
[
1
0
⋯
0
0
2
⋯
0
⋮
⋮
⋱
⋮
0
0
⋯
]
[
1
2
⋮
]
(
4)
S
VD
c
a
n
identi
f
y
a
nd
a
r
a
nge
di
mens
ions
that
indi
c
a
te
whic
h
da
ta
va
r
iations
of
ten
a
ppe
a
r
.
S
VD
take
s
the
ter
m
of
doc
ument
matr
ix
whic
h
c
ons
is
ts
of
wor
ds
a
nd
doc
uments
a
s
in
T
a
ble
1
whic
h
ha
s
be
e
n
br
oke
n
down
int
o
li
ne
a
r
indepe
nde
nt
c
o
mponents
.
T
he
r
e
s
ult
of
the
S
VD
pr
oc
e
s
s
is
a
ve
c
tor
that
wil
l
be
us
e
d
to
be
c
a
lcula
ted
it
s
s
im
il
a
r
it
y
by
a
n
a
ppr
oa
c
h.
2.
1.
4.
Calc
u
lat
in
g
c
os
in
e
s
im
il
ar
i
t
y
C
os
ine
s
im
il
a
r
it
y
is
us
e
d
to
c
a
lcula
te
the
c
os
in
e
va
lue
be
twe
e
n
doc
uments
ve
c
tor
in
a
c
oll
e
c
ti
on
a
nd
the
ne
e
de
d
input
ve
c
tor
[3
1
,
3
2
]
.
T
he
s
maller
the
pr
oduc
e
d,
the
higher
the
leve
l
of
s
im
il
a
r
it
y
of
the
e
s
s
a
y
oc
c
ur
e
.
T
he
f
or
mul
a
of
c
os
ine
s
im
il
a
r
it
y
is
a
s
s
hown
in
(
5
)
:
C
os
α
=
A
.
B
|
A
|
.
|
B
|
=
∑
A
=
1
x
B
√
∑
(
A
)
2
=
1
x
∑
(
B
)
2
=
1
(
5)
with
the
s
tate
ment,
it
s
howe
d
that
A
is
a
doc
ument
ve
c
tor
,
B
is
a
n
input
ve
c
tor
,
A.
B
is
the
dot
pr
oduc
t
of
ve
c
tor
A
with
ve
c
tor
B
,
|A|
is
the
length
of
ve
c
tor
A
,
|B
|
i
s
the
length
of
ve
c
tor
B
,
|A|.
|B
|
is
a
c
r
os
s
pr
oduc
t
be
twe
e
n
|A|
a
nd
|B
|
a
nd
α
is
the
a
nge
l
whic
h
is
f
or
med
be
tw
e
e
n
ve
c
tor
A
a
nd
ve
c
tor
B
.
3.
RE
S
UL
T
S
AN
D
AN
AL
YSI
S
I
n
thi
s
s
e
c
ti
on,
it
is
e
xplaine
d
the
r
e
s
ul
ts
of
r
e
s
e
a
r
c
h
a
nd
a
t
the
s
a
me
ti
me
is
given
the
c
omp
r
e
he
ns
ive
dis
c
us
s
ion
a
bout
how
L
S
A
a
nd
C
S
a
r
e
im
ple
mente
d
in
s
e
a
r
c
hing
inf
or
mat
ion
of
h
a
dit
hs
a
nd
pr
e
s
e
nt
the
e
va
luation
r
e
s
ult
of
e
xpe
r
im
e
nt
that
c
onduc
ted.
3.
1.
P
r
e
-
p
r
oc
e
s
s
in
g
f
or
t
e
xt
d
at
a
T
e
xt
da
ta
is
uns
tr
uc
tur
e
d
da
ta
that
ne
e
ds
s
pe
c
ia
l
tr
e
a
tm
e
nt
be
f
or
e
c
a
r
ied
out
mi
ning
pr
oc
e
s
s
or
s
e
a
r
c
hing
f
or
in
f
or
mation
c
ontaine
d
in
the
text
[3
0
]
.
T
he
p
r
e
pr
oc
e
s
s
ing
s
tage
f
or
text
is
the
s
tage
of
p
r
e
pa
r
ing
text
da
ta
int
o
a
s
tr
uc
tur
e
d
da
ta
r
e
pr
e
s
e
ntation.
Ge
ne
r
a
ll
y,
two
types
of
s
tr
uc
tur
e
d
da
ta
r
e
pr
e
s
e
ntations
f
or
text
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
L
atent
s
e
mantic
analys
is
and
c
os
ine
s
imil
ar
it
y
for
hadit
h
s
e
ar
c
h
e
ngine
(
W
ahy
udin
Dar
malak
s
ana
)
221
a
r
e
ba
g
of
wor
ds
a
nd
mul
ti
ple
o
f
wor
ds
[
33
,
34
]
.
L
a
tent
s
e
mantic
a
na
lys
is
is
one
a
lgor
it
hm
that
pr
oduc
e
s
s
tr
uc
tur
e
d
text
r
e
pr
e
s
e
ntations
in
the
f
o
r
m
o
f
mul
t
ipl
e
of
wor
ds
.
W
he
r
e
,
the
text
is
not
only
r
e
pr
e
s
e
nted
by
1
wor
d
but
a
ls
o
c
a
n
be
mor
e
than
1
wor
d
or
a
ls
o
known
a
s
n
-
gr
a
m.
E
ve
n
the
late
nt
s
e
mantic
a
na
ly
s
is
wor
d
c
oll
e
c
ti
ons
c
on
s
ider
s
to
the
s
e
mantics
be
twe
e
n
one
wor
d
a
nd
a
nother
.
P
r
e
-
pr
oc
e
s
s
ing
of
text
da
ta
s
tar
ts
f
r
om
unif
or
m
it
y
of
the
s
ize
of
letter
s
to
lowe
r
c
a
s
e
,
de
leting
c
ha
r
a
c
ter
s
other
than
letter
s
a
nd
r
e
gular
e
xp
r
e
s
s
ions
,
if
it
is
n
e
c
e
s
s
a
r
y
to
c
ha
nge
a
bbr
e
viations
to
be
t
he
i
r
or
igi
n
a
l
f
or
m
,
de
lete
unim
por
tant
wor
ds
or
s
top
wor
d
r
e
moval
,
t
he
n
it
is
the
pr
oc
e
s
s
to
c
ha
nge
the
ini
ti
a
l
wor
ds
in
to
wor
ds
e
s
s
e
nti
a
ll
y
or
s
temmi
ng.
I
n
thi
s
s
tudy,
the
s
temm
ing
pr
oc
e
s
s
us
e
s
the
Na
z
ief
&
Adr
iani
a
lgor
it
hm
be
c
a
us
e
the
h
a
dit
h
text
doc
um
e
nts
a
r
e
a
r
r
a
nge
d
in
I
ndone
s
ian.
T
he
Na
z
ief
&
Adr
iani
a
lgor
i
thm
is
the
mos
t
c
o
mm
only
us
e
d
s
temmi
ng
a
lgor
i
thm
f
or
I
ndone
s
ian
be
c
a
us
e
it
is
in
a
c
c
or
da
nc
e
with
the
s
yntax
o
f
I
ndone
s
ian
[
35
–
39]
.
T
he
r
e
s
ult
s
of
the
s
temmi
ng
us
e
d
a
s
da
ta
a
r
e
e
nter
e
d
f
or
the
late
nt
s
e
mantic
a
na
lys
is
a
nd
f
or
med
the
ter
m
of
doc
ument
matr
ix
f
r
om
the
text
da
ta.
3.
2.
I
m
p
lem
e
n
t
at
ion
of
lat
e
n
t
s
e
m
an
t
ic
an
alyai
s
an
d
c
os
in
e
s
im
i
larit
y
o
n
t
h
e
h
ad
i
t
h
s
e
ar
c
h
e
n
gi
n
e
s
L
a
tent
s
e
mantic
a
na
lyais
is
a
ppli
e
d
a
f
ter
the
pr
e
pr
oc
e
s
s
of
text
is
c
ompl
e
te.
T
he
n
the
pr
e
p
r
oc
e
s
s
r
e
s
ult
s
will
be
f
or
med
to
be
ter
m
of
doc
ument
mat
r
ix.
T
he
ter
m
o
f
doc
ument
matr
ix
wil
l
be
c
omput
e
d
by
S
VD
to
pr
oduc
e
a
matr
ix
of
U
,
S
,
a
nd
V.
T
he
f
inal
s
tage
is
the
a
ppli
c
a
ti
on
of
c
os
ine
s
im
il
a
r
it
y
to
s
e
e
the
s
im
il
a
r
it
y
of
the
inf
o
r
mation
ge
ne
r
a
ted
a
s
we
ll
a
s
a
r
a
nge
it
ba
s
e
d
on
the
leve
l
of
s
im
il
a
r
it
y
.
T
he
f
low
o
f
the
late
nt
s
e
mantic
a
na
lys
is
a
nd
c
o
s
ine
s
im
il
a
r
it
y
that
im
pe
mente
d
in
thi
s
s
tudy
c
a
n
be
s
e
e
n
a
t
the
F
igur
e
1.
F
or
ins
tanc
e
,
ther
e
a
r
e
3
piec
e
s
of
the
f
o
ll
owing
h
a
dit
h
doc
uments
(
p
r
e
s
e
nt
in
I
ndone
s
ian
langua
ge
)
:
Doc
u
m
e
n
t
1:
J
anganlah
k
ali
an
be
r
dus
ta
atas
namak
u,
k
ar
e
na
s
iapa
y
ang
be
r
dus
ta
atas
namak
u
nis
c
ay
a
dia
mas
uk
ne
r
ak
a.
(
Do
not
li
e
on
be
ha
lf
of
my
na
me,
be
c
a
us
e
if
a
nyon
e
who
li
e
s
on
be
ha
lf
of
my
na
me,
he
/s
he
will
go
to
the
he
ll
s
ur
e
ly.
)
Doc
u
m
e
n
t
2:
J
anganlah
k
ali
an
be
r
dus
ta
ter
hadapk
u
(
atas
namak
u)
,
k
ar
e
na
bar
angs
iapa
be
r
dus
ta
ter
hadapk
u
dia
ak
an
mas
uk
ne
r
ak
a.
(
Do
not
l
ie
to
me
(
on
my
be
ha
lf
)
,
be
c
a
us
e
whoe
ve
r
li
e
s
on
me
he
will
go
to
the
he
ll
.
)
Doc
u
m
e
n
t
3:
B
ar
angs
iapa
y
ang
s
e
ngaja
me
lakuk
an
k
e
dus
taan
atas
namak
u,
mak
a
he
ndak
lah
dia
me
ne
mpati
tem
pat
duduk
ny
a
dar
i
ne
r
ak
a.
(
W
hoe
ve
r
de
li
be
r
a
tely
li
e
s
on
be
ha
lf
o
f
my
na
me,
he
s
hould
oc
c
upy
his
s
e
a
t
f
r
om
the
he
ll
.
)
I
n
p
u
t
Keyw
or
d
s
in
Had
it
h
S
e
ar
c
h
E
n
gi
n
e
:
J
angan
Dus
ta
M
as
uk
N
e
r
ak
a
(
Do
not
l
ie
to
go
to
the
he
ll
)
T
e
xt
da
ta
f
r
om
thes
e
thr
e
e
doc
uments
a
nd
go
to
th
e
s
e
a
r
c
h
e
ngine.
I
t
will
be
c
a
r
ied
out
pr
e
-
pr
oc
c
e
s
s
to
pr
oduc
e
text
da
ta
a
s
f
oll
ows
:
Doc
u
m
e
n
t
1:
jangan
k
ali
an
dus
ta
atas
nama
dus
ta
Doc
u
m
e
n
t
2:
jangan
k
ali
an
dus
ta
atas
nama
dus
ta
mas
uk
ne
r
ak
a
Doc
u
m
e
n
t
3:
s
e
ngaja
dus
ta
atas
nama
he
ndak
tem
pat
duduk
ne
r
ak
a
I
n
p
u
t
k
e
ywor
d
s
in
h
ad
i
t
h
s
e
ar
c
h
e
n
gin
e
:
jangan
dus
ta
mas
uk
ne
r
ak
a
T
he
n,
the
a
lr
e
a
dy
th
r
e
e
pr
e
pa
r
e
d
text
da
ta
is
pr
oc
e
s
s
e
d
to
f
or
m
matr
ixes
o
f
the
ter
m
of
doc
ument
li
ke
s
on
T
a
ble
1
a
nd
it
is
ga
ined
A
matr
ixes
a
s
f
oll
ows
:
A
=
(
1
1
0
1
1
0
1
1
1
1
1
1
0
0
0
0
0
1
1
1
0
1
1
1
0
0
0
0
1
1
1
0
0
1
0
1
1
1
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
1
,
F
e
br
ua
r
y
2020
:
217
-
227
222
T
he
main
s
tep
that
ne
e
ds
to
be
c
ompl
e
ted
is
to
de
c
o
mpos
e
A
matr
ix
to
be
3
other
matr
ice
s
us
ing
S
VD
,
s
tar
ti
ng
f
r
om
f
indi
ng
the
A
T
A
va
lue
to
c
a
lcula
te
with
c
os
ine
s
im
il
a
r
it
y.
T
he
pr
oc
e
s
s
of
a
pplyi
n
g
L
a
ten
S
e
mantics
Ana
lys
is
a
nd
C
os
ine
S
im
il
a
r
it
y
f
or
th
e
ter
m
of
doc
ument
matr
ix
is
in
the
f
oll
owing
T
a
ble
1.
S
e
a
r
c
h
the
va
lue
of
AT
A:
A
T
A
=
(
1
1
1
1
1
1
1
1
0
0
0
0
0
1
1
1
1
1
0
1
1
1
0
0
0
0
0
0
1
1
1
0
0
1
0
1
1
1
1
)
(
1
1
0
1
1
0
1
1
1
1
1
1
1
1
1
1
0
0
1
1
0
1
1
1
0
1
0
0
0
1
0
0
1
0
0
1
0
0
1
)
=
(
8
7
4
6
7
4
4
4
8
)
s
e
a
r
c
h
de
ter
mi
na
nt
of
A
T
A
r
e
s
ult
,
s
o
|AT
A
-
λ
I
|=
0
:
−
ƛ
=
(
8
7
4
6
7
4
4
4
8
)
−
(
ƛ
0
0
0
ƛ
0
0
0
ƛ
)
=
(
8
−
ƛ
7
4
6
7
−
ƛ
4
4
4
8
−
ƛ
)
|A
T
A
−
ƛ
|
=
(
8
−
ƛ
)
de
t
(
7
−
ƛ
4
4
8
−
ƛ
)
−
(
7
)
(
6
4
4
8
−
ƛ
)
−
(
4
)
(
6
7
−
ƛ
4
4
)
|A
T
A
−
ƛ
|
=
[
(
7
)
(
8
−
ƛ
)
−
(
4
)
(
4
)
]
−
(
7
)
[
(
6
)
(
8
−
ƛ
)
−
(
4
)
(
4
)
]
+
(
4
)
[
(
6
)
(
4
)
[
−
(
7
−
ƛ
)
(
4
)
]
|
−
ƛ
|
=
3
+
23
2
−
102
+
80
=
0
s
e
a
r
c
h
e
igen
va
lue
a
nd
e
igen
va
c
tor
:
E
igen
Va
lue:
E
igen
Va
c
tor
:
λ
1=
17.
40312
V1
=
1.
24704
,
1
.
10373,
1
λ
2=
4.
59687
V2=
-
0.
54366,
-
0.
30712,
1
λ
3=
1
V3=
-
1,
1,
0
s
e
a
r
c
h
s
ingul
a
r
matr
ix
ba
s
e
d
on
the
va
lue
of
e
igen
va
lue
whic
h
ha
s
be
e
n
ga
ined:
S
1 =
√17.
403
12
=
4.
171
7
S
2 =
√ 4.59687 =
2.14403
S
3 =
√1 =
1
S
=
(
S1
0
0
0
S2
0
0
0
S3
)
=
(
4
.
1717
0
0
0
2
.
14403
0
0
0
1
)
S
-
1 =
(
0
.
23971
0
0
0
0
.
46641
0
0
0
1
)
s
e
a
r
c
h
V
matr
ix
va
lue
by
us
ing
va
lue
no
r
maliza
ti
o
n
of
e
igen
va
c
tor
whic
h
ha
s
be
e
n
ga
ined:
|
V1
|
=
√
1
.
24704
2
+
1
.
10373
2
+
1
2
=
1.94251
|
V2
|
=
√
−
0
.
54366
2
+
−
0
.
30712
2
+
1
2
=
.17894
|
V3
|
=
√
−
1
2
+
1
2
+
0
2
=
1.41421
V1
=
1
.
24704
1
.
94251
,
1
.
10373
1
.
94251
,
1
1
.
94251
=
0.
64197
,
0
.
56819,
0.
51479
V2
=
−
0
.
54366
1
.
17894
,
−
0
.
30712
1
.
17894
,
1
1
.
17894
=
-
0.
46114,
-
0.
26051,
0.
84822
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
L
atent
s
e
mantic
analys
is
and
c
os
ine
s
imil
ar
it
y
for
hadit
h
s
e
ar
c
h
e
ngine
(
W
ahy
udin
Dar
malak
s
ana
)
223
V3
=
−
1
1
.
41421
,
1
1
.
41421
,
0
1
.
41421
=
-
0.
70711,
0.
70711
,
0
f
or
mul
a
te
V
matr
ics
with
ga
ined
va
lue
f
r
omt
he
r
e
s
ult
of
nor
maliza
ti
on
c
a
lcula
ti
on
of
e
igen
va
c
tor
:
V
=
(
−
0
.
64197
0
.
56819
0
.
51479
0
.
46114
−
0
.
26051
0
.
84822
−
0
.
70711
0
.
70711
0
)
V
T
=
(
0
.
64197
−
0
.
46114
−
0
.
70711
0
.
56819
−
0
.
26051
0
.
70711
0
.
51479
0
.
84822
0
)
s
e
a
r
c
h
U
matr
ix
va
lue
wi
th
the
f
or
mul
a
of
U=
AV
S
-
1:
U
=
(
1
1
0
1
1
0
1
1
1
1
1
1
1
1
1
1
0
0
1
1
0
1
1
1
0
1
0
0
0
1
0
0
1
0
0
1
0
0
1
)
x
(
−
0
.
64197
0
.
56819
0
.
51479
0
.
46114
−
0
.
26051
0
.
84822
−
0
.
70711
0
.
70711
0
)
x
(
0
.
23971
0
0
0
0
.
46641
0
0
0
1
)
U
=
(
0
.
04335
0
.
14351
1
.
36301
0
.
04335
0
.
14351
1
.
36301
−
0
.
12615
−
0
.
12615
−
0
.
12615
0
.
15389
0
.
04335
−
0
.
12615
−
0
.
11054
−
0
.
16590
−
0
.
16590
−
0
.
16590
−
0
.
16590
0
.
47331
0
.
47331
0
.
47331
0
.
26501
0
.
14351
0
.
47331
−
0
.
12150
0
.
32980
0
.
32980
0
.
32980
0
.
32980
1
.
36301
1
.
36301
1
.
36301
0
.
51479
1
.
36301
1
.
36301
0
.
84822
0
0
0
0
)
Af
ter
be
ing
obtaine
d
the
va
lue
o
f
the
USV
T
matr
i
x,
the
ne
xt
s
tep
is
to
r
e
duc
e
the
r
a
nk
of
the
matr
ix
.
T
his
wa
s
done
in
or
de
r
to
r
e
duc
e
c
omput
ing
ti
me.
I
t
is
a
n
e
xa
mpl
e
of
a
r
a
nk
r
e
duc
ti
on
of
k
=
2
f
r
om
the
USV
T
matr
ix
a
s
f
oll
ows
:
U
k =
(
0
.
04335
0
.
14351
0
.
04335
0
.
14351
−
0
.
12615
−
0
.
12615
−
0
.
12615
0
.
15389
0
.
04335
−
0
.
12615
−
0
.
11054
−
0
.
16590
−
0
.
16590
−
0
.
16590
−
0
.
16590
0
.
47331
0
.
47331
0
.
47331
0
.
26501
0
.
14351
0
.
47331
−
0
.
12150
0
.
32980
0
.
32980
0
.
32980
0
.
32980
)
S
k =
(
4
.
1717
0
0
2
.
14403
)
;
S
k
-
1 =
(
0
.
23971
0
0
0
.
46641
)
V
k =
(
0
.
64197
0
.
56819
−
0
.
46114
−
0
.
26051
−
0
.
70711
0
.
70711
)
;
V
kT
=
(
0
.
64197
−
0
.
46114
−
0
.
70711
0
.
5681
−
0
.
26051
0
.
70711
)
T
he
las
t
s
t
e
p
is
to
c
a
lcula
te
a
ngle
c
os
ine
va
lue
be
t
we
e
n
doc
ument
va
c
tor
(
A
)
a
nd
input
va
c
tor
(
B
)
a
s
f
oll
ows
:
D
i
=
D
iT
U
k S
k
-
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
1
,
F
e
br
ua
r
y
2020
:
217
-
227
224
D
i
=
D
iT
=
(
0
.
04335
0
.
14351
0
.
04335
0
.
14351
−
0
.
12615
−
0
.
12615
−
0
.
12615
0
.
15389
0
.
04335
−
0
.
12615
−
0
.
11054
−
0
.
16590
−
0
.
16590
−
0
.
16590
−
0
.
16590
0
.
47331
0
.
473
31
0
.
47331
0
.
26501
0
.
14351
0
.
47331
−
0
.
12150
0
.
32980
0
.
32980
0
.
32980
0
.
32980
)
(
0
.
23971
0
0
0
.
46
641
)
D
M
=
(
−
0
.
03970
0
.
57538
)
D
1 =
(
0
.
64197
0
.
56819
)
D
2 =
(
−
0
.
46114
−
0
.
26051
)
D
3 =
(
−
0
.
70711
0
.
70711
)
C
os
α =
A
.
B
|
A
|
.
|
B
|
C
o
s
α
1
=
(
−
0
.
03970
)
(
0
.
64197
)
+
(
0
.
57538
)
(
0
.
56819
)
√
(
−
0
.
03970
)
2
+
(
0
.
57538
)
2
√
(
0
.
64197
)
2
+
(
0
.
56819
)
2
C
o
s
α
2
=
(
−
0
.
03970
)
(
−
0
.
46114
)
+
(
0
.
57538
)
(
−
0
.
26051
)
√
(
−
0
.
03970
)
2
+
(
0
.
57538
)
2
√
(
−
0
.
46114
)
2
+
(
−
0
.
26051
)
2
C
os
α
1
=
0
.
71113
C
os
α
2
=
0
.
43739
C
os
α
3
=
0
.
70542
F
r
om
the
r
e
s
ult
s
of
the
a
bove
c
a
lcula
ti
on,
it
c
a
n
be
c
onc
luded
that
the
a
r
a
nge
ment
of
doc
uments
t
ha
t
ha
ve
the
c
los
e
s
t
s
im
il
a
r
it
y
with
the
input
doc
uments
is
d
oc
ument
1,
doc
ument
3,
a
nd
doc
ument
2.
3.
3.
E
xp
e
r
im
e
n
t
an
d
r
e
s
u
lt
e
valu
at
ion
T
e
s
ti
ng
is
c
a
r
ied
out
by
t
r
ying
a
ll
the
ha
dit
h
que
r
i
e
s
on
the
s
ys
tem.
R
e
c
a
ll
a
nd
p
r
e
c
is
ion
va
lues
a
r
e
s
e
a
r
c
he
d
by
us
ing
f
or
mul
a
s
(
6
)
a
nd
(
7)
[
38
,
39
]
.
=
(
6)
=
(
7)
w
he
r
e
,
R
is
R
e
c
a
ll
,
s
o
the
R
va
lue
is
obtaine
d
b
y
c
ompar
ing
the
Numbe
r
o
f
r
e
leva
nt
it
e
ms
r
e
tr
ie
ve
d
with
the
t
otal
numbe
r
of
r
e
leva
nt
it
e
ms
in
the
c
oll
e
c
ti
on
.
R
e
c
a
ll
is
a
doc
ument
that
is
c
a
ll
e
d
f
r
om
the
s
ys
tem
ba
s
e
d
on
the
us
e
r
r
e
q
ue
s
ts
that
f
ol
low
the
pa
tt
e
r
n
of
the
s
ys
tem.
T
he
gr
e
a
ter
R
e
c
a
ll
va
lue
c
a
nnot
be
s
a
id
a
s
a
good
s
ys
tem
or
not.
And
,
P
is
p
r
e
c
is
ion.
S
o
,
the
P
va
lue
is
obtaine
d
by
c
ompar
ing
the
n
umber
o
f
r
e
leva
nt
it
e
ms
r
e
tr
ieve
d
with
the
T
otal
number
of
i
tems
r
e
tr
ieve
d.
P
r
e
c
is
ion
is
the
number
of
doc
uments
that
a
r
e
c
a
ll
e
d
f
r
om
the
r
e
leva
nt
da
taba
s
e
a
f
ter
be
ing
a
s
s
e
s
s
e
d
by
the
us
e
r
with
ne
e
de
d
inf
or
mation
.
T
he
gr
e
a
ter
the
va
lue
of
a
s
ys
tem
pr
e
c
is
ion,
the
s
ys
tem
c
a
n
be
s
a
id
we
ll
.
T
he
pur
pos
e
of
the
r
e
c
a
ll
a
nd
p
r
e
c
is
ion
tes
t
i
s
to
obtain
inf
or
mation
on
the
s
e
a
r
c
h
r
e
s
ult
s
obtaine
d
by
the
s
ys
tem.
S
e
a
r
c
h
r
e
s
ult
s
c
a
n
be
judged
by
it
s
r
e
c
a
ll
a
nd
pr
e
c
is
ion
leve
l
.
P
r
e
c
is
ion
c
a
n
be
c
o
ns
ider
e
d
a
mea
s
ur
e
of
a
c
c
ur
a
c
y
while
r
e
c
a
ll
is
pe
r
f
e
c
ti
on.
T
he
va
lue
of
p
r
e
c
is
ion
is
the
leve
l
o
f
a
c
c
u
r
a
c
y
be
twe
e
n
the
inf
or
mation
r
e
que
s
ted
by
the
us
e
r
a
nd
the
a
ns
we
r
s
given
by
the
s
ys
tem.
W
hil
e
the
R
e
c
a
ll
va
lue
is
the
s
uc
c
e
s
s
leve
l
of
the
s
ys
tem
in
r
e
dis
c
ove
r
ing
in
f
or
mation
.
As
f
or
the
r
e
s
ult
s
of
the
r
e
c
a
ll
a
nd
pr
e
c
is
ion
t
e
s
ts
a
nd
the
ti
me
whic
h
is
s
pe
nt
on
s
e
a
r
c
hing
the
tes
ted
h
a
d
it
h,
it
c
a
n
be
s
e
e
n
in
T
a
ble
2
,
F
igur
e
s
3
a
nd
4.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
L
atent
s
e
mantic
analys
is
and
c
os
ine
s
imil
ar
it
y
for
hadit
h
s
e
ar
c
h
e
ngine
(
W
ahy
udin
Dar
malak
s
ana
)
225
F
igur
e
3
.
R
e
s
ult
of
r
e
leva
nt
inf
o
r
mation
F
igur
e
4
.
R
e
s
ult
of
pr
e
c
is
ion
a
nd
r
e
c
a
ll
va
lue
T
a
ble
2
.
T
e
s
ted
r
e
s
ult
o
f
late
nt
s
e
mantics
a
na
lys
is
a
nd
c
os
ine
s
im
il
a
r
it
y
No
K
e
yw
or
ds
A
ppe
a
r
e
d
r
e
le
va
nt
H
a
di
th
A
ppe
a
r
e
d
ir
e
le
va
nt
H
a
di
th
T
he
t
ot
a
l
numbe
r
of
r
e
le
va
nt
H
a
di
th
R
e
c
a
ll
(%)
P
r
e
c
is
io
n
(%)
1
J
angan
be
r
dus
ta
at
as
nam
ak
u
m
as
uk
ne
r
ak
a
(
D
on’
t
li
e
in
be
ha
lf
of
my
na
me
to
go
to
th
e
h
e
ll
)
2
2
2
100
50
2
M
e
ndi
r
ik
an
s
hal
at
m
e
nunaik
an
z
ak
at
dan
be
r
puas
a
di
bul
an
r
am
adl
an
(
C
a
r
y
out
pr
a
yi
ng,
a
lm
s
a
nd pa
s
t
in
r
a
ma
da
n
M
ont
h)
2
4
2
100
33.33
3
I
s
la
m
di
bangun
at
as
li
m
a
das
ar
y
ai
tu
pe
r
s
ak
s
ia
n,
s
hal
at
,
z
ak
at
,
puas
a
dan
k
e
bai
tu
ll
ah
(
I
s
la
m
w
a
s
f
or
me
d
in
f
iv
e
pi
la
r
s
na
me
ly
;
w
it
hne
s
s
,
pr
a
yi
ng,
a
lm
s
,
pa
s
ti
ng
a
nd
pi
lg
r
im
a
ge
t
o me
c
c
a
)
3
2
3
100
60
4
B
ar
angs
ia
pa
y
ang
be
r
puas
a
di
bul
an
r
am
adl
an
de
ngan
k
e
imanan
dan
ik
hl
as
di
am
puni
dos
a
-
dos
any
a
(
W
hoe
ve
r
f
a
s
ts
in
th
e
mont
h
of
R
a
ma
da
n w
it
h f
a
it
h a
nd s
in
c
e
r
it
y i
s
f
or
gi
ve
n of
hi
s
s
in
s
)
2
3
4
50
40
5
M
al
u
s
e
bagi
an
dar
i
iman
(
S
ha
me
is
pa
r
t
of
f
a
it
h)
1
1
3
33.33
50
30
A
k
u
pe
r
nah
m
andi
be
r
s
am
a
N
abi
s
hal
la
ll
ahu
'
al
ai
hi
w
as
al
la
m
dar
i
s
at
u
be
ja
na,
dan
ta
ngan
k
am
i
s
al
in
g be
r
s
e
nt
uhan
(
I
ha
d ba
th
e
d w
it
h t
he
P
r
ophe
t
s
a
ll
a
ll
a
a
hu
'
a
l
a
ih
i
w
a
s
a
ll
a
m
f
r
om
one
ve
s
s
e
l
a
nd our
ha
nd
s
t
ouc
he
d e
a
c
h ot
he
r
)
1
1
1
100
50
31
Se
ti
ap
N
abi
m
e
m
il
ik
i
doa
y
ang
di
a
panj
at
k
an
unt
uk
um
at
ny
a
(
E
ve
r
y
P
r
ophe
t
ha
s
a
pr
a
ye
r
th
a
t
he
pr
a
ye
d f
or
hi
s
pe
opl
e
)
1
3
2
50
25
32
J
ik
a
dat
ang
hai
d
ti
nggalk
an
s
hal
at
dan
bi
la
be
r
ak
hi
r
b
e
r
s
ik
an
dar
ah
la
lu
s
hal
at
la
h
(If
me
ns
tr
ua
ti
on
c
ome
s
le
a
ve
pr
a
ye
r
a
nd
w
he
n
it
e
nds
, c
le
a
n bl
oody the
n pr
a
y)
3
1
3
100
75
33
T
uj
uh
pul
uh
r
ib
u
o
r
ang
da
r
i
um
at
k
u
ak
an
m
as
uk
s
ur
ga,
w
aj
ah
m
e
r
e
k
a
s
e
m
ua
s
e
pe
r
ti
r
e
m
bul
an
(
S
e
ve
nt
y t
hous
a
nd of
my
pe
opl
e
w
il
l
go t
o he
a
ve
n, t
he
ir
f
a
c
e
s
l
ik
e
t
he
moon)
1
4
1
100
20
47
J
adi
k
anl
ah
(
s
e
bagi
an
da
r
i)
s
hal
at
k
al
ia
n
ada
di
r
um
ah
k
al
ia
n
dan
ja
ngan
ja
di
k
an
k
ubur
an
(
M
a
ke
(
s
ome
of
)
your
pr
a
ye
r
s
in
your
hous
e
a
nd
do not m
a
ke
i
t
a
gr
a
ve
)
1
2
1
100
33.33
48
B
ar
angs
ia
pa
m
e
ni
nggal
dal
am
k
e
adaan
m
e
ny
e
k
ut
uk
an
A
ll
ah
de
ngan
s
e
s
uat
u,
m
ak
a
ia
m
as
uk
ne
r
ak
a
(
W
hoe
ve
r
di
e
s
in
a
s
ta
te
th
a
t
a
s
s
oc
ia
t
e
s
G
od
w
it
h
s
ome
th
in
g,
he
goe
s
to
he
ll
)
2
3
2
100
40
49
C
uk
upl
ah
s
e
s
e
or
ang
(
di
anggap)
be
r
bohong
apabil
a
di
a
m
e
nc
e
r
it
ak
an
s
e
m
ua
(
I
t
is
e
nough
f
or
s
ome
one
(
c
ons
id
e
r
e
d)
t
o l
ie
i
f
he
t
e
ll
s
a
ll
)
1
0
1
100
100
50
Se
or
ang
m
us
li
m
y
ang
pal
in
g
bai
k
adal
ah
k
am
bi
ng
y
ang
di
ge
m
bal
ak
anny
a
di
punc
ak
gunung
dan
te
m
pat
-
te
m
pat
te
r
pe
nc
il
(
T
he
b
e
s
t
M
us
li
m
is
th
e
goa
t
th
a
t
he
f
e
e
ds
on
mount
a
in
to
ps
a
nd r
e
mot
e
pl
a
c
e
s
)
1
2
1
100
33.33
A
ve
r
a
ge
(
%
)
87.83
36.25
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
1
,
F
e
br
ua
r
y
2020
:
217
-
227
226
4.
CONC
L
USI
ON
B
a
s
e
d
on
50
ti
mes
tes
ti
ng
of
the
r
e
c
a
ll
a
nd
p
r
e
c
is
ion
va
lues
that
ha
ve
be
e
n
c
a
r
r
ied
out
(
c
ontaine
d
in
T
a
ble
2)
,
it
s
howe
d
that
the
s
e
a
r
c
h
e
ngine
ha
dit
h
pe
r
f
or
manc
e
c
a
n
a
pply
the
late
nt
s
e
mantics
a
na
lys
is
a
lgor
it
hm
a
nd
c
os
ine
s
im
il
a
r
it
y
quit
e
we
ll
.
Ha
dit
h
inf
or
m
a
ti
on
whic
h
is
obtaine
d
ba
s
e
d
on
ke
ywor
ds
,
phr
a
s
e
s
,
or
s
e
ntenc
e
s
e
nter
e
d
s
uc
c
e
s
s
f
ull
y
f
ound
we
ll
,
it
wa
s
indi
c
a
ted
by
a
r
e
c
a
ll
va
lue
of
87.
83%
.
Although
th
e
ove
r
a
ll
inf
or
mation
whic
h
is
ge
ne
r
a
ted
onl
y
ha
s
a
va
lue
of
a
c
c
ur
a
c
y
or
c
ompl
ianc
e
with
us
e
r
input
only
36.
25
%
whic
h
is
indi
c
a
ted
by
the
va
lue
of
the
pr
oduc
e
d
pr
e
c
is
i
on.
Ge
ne
r
a
ll
y,
the
late
nt
s
e
mantics
a
na
lys
is
a
lgor
it
hm
a
nd
c
os
ine
s
im
il
a
r
it
y
that
a
r
e
us
e
d
a
r
e
a
ble
to
pr
oduc
e
the
h
a
dit
h
inf
or
mation
we
ll
.
T
he
r
e
we
r
e
s
e
ve
r
a
l
f
a
c
tor
s
that
inf
luenc
e
d
the
s
e
a
r
c
h
r
e
s
ult
s
other
than
the
pos
s
ibi
li
ty
of
a
n
e
r
r
or
in
us
ing
the
a
lgor
i
thm
,
including
inc
ompl
e
te
da
ta
a
nd
too
much
nois
e
.
T
he
r
e
f
or
e
,
the
pr
e
pr
oc
e
s
s
ing
s
tage
is
ve
r
y
im
por
tant
to
be
a
ble
to
pr
od
uc
e
mo
r
e
a
c
c
ur
a
te
inf
or
mation.
B
e
c
a
us
e
the
pr
e
pr
oc
e
s
s
ing
s
tage
pr
oduc
e
s
text
da
ta
that
give
s
a
n
input
int
o
the
late
nt
s
e
mantics
a
na
ly
s
is
a
lgor
it
hm
whic
h
will
c
e
r
tainly
a
f
f
e
c
t
the
s
e
a
r
c
h
r
e
s
ult
s
.
F
or
f
ur
ther
r
e
s
e
a
r
c
h,
the
c
oll
e
c
ti
on
of
s
a
ve
d
Ha
dit
h
da
ta
ne
e
ds
to
be
c
ompl
e
ted
s
o
that
s
e
a
r
c
h
e
ngines
c
a
n
lea
r
n
a
nd
ge
t
mor
e
pr
e
c
is
e
d
inf
or
mation.
I
n
a
ddit
ion
,
the
inf
or
mation
obtaine
d
c
a
n
be
de
ve
loped
not
only
s
or
ted
by
s
im
il
a
r
it
y
but
a
ls
o
c
a
n
be
gr
oupe
d
a
c
c
or
ding
to
their
mea
nings
.
AC
KNOWL
E
DGE
M
E
NT
Author
s
wis
hing
to
a
c
knowle
dge
R
e
s
e
a
r
c
h
a
nd
P
ubli
c
a
ti
on
C
e
ntr
e
of
UI
N
S
una
n
Gunung
Dja
ti
B
a
ndung
that
s
uppor
ts
a
nd
f
unds
thi
s
r
e
s
e
a
r
c
h
publ
ica
ti
on.
RE
F
E
RE
NC
E
S
[1
]
J
.
M.
K
as
s
i
m
an
d
M.
Rah
man
y
,
“In
t
ro
d
u
c
t
i
o
n
t
o
s
ema
n
t
i
c
s
earch
en
g
i
n
e,
”
P
r
o
ceed
i
n
g
s
o
f
t
h
e
2
0
0
9
In
t
er
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
E
l
ec
t
r
i
ca
l
E
n
g
i
n
eer
i
n
g
a
n
d
In
f
o
r
m
a
t
i
cs
,
I
CE
E
I
2
0
0
9
,
v
o
l
.
0
2
,
2
0
0
9
.
[2
]
D
.
K
u
r
n
i
a
d
i
an
d
A
.
M
u
l
y
an
i
,
“
T
h
e
E
ff
ect
o
f
G
o
o
g
l
e'
s
Search
E
n
g
i
n
e
T
ec
h
n
o
l
o
g
y
o
n
t
h
e
D
ev
e
l
o
p
men
t
o
f
St
u
d
en
t
Cu
l
t
u
re
an
d
E
t
h
i
c
s
(i
n
Bah
a
s
a:
Pen
g
aru
h
T
ek
n
o
l
o
g
i
Mes
i
n
Pen
cari
G
o
o
g
l
e
T
er
h
ad
a
p
Perk
emb
a
n
g
a
n
Bu
d
a
y
a
d
a
n
E
t
i
k
a
Mah
as
i
s
w
a
)
,
”
Ju
r
n
a
l
A
l
g
o
r
i
t
m
a
S
ek
o
l
a
h
Ti
n
g
g
i
Tek
n
o
l
o
g
i
G
a
r
u
t
,
v
o
l
.
1
4
,
n
o
.
1
,
2
0
1
7
.
[3
]
P.
W
.
H
an
d
ay
an
i
,
I.
M.
W
i
ry
an
a,
an
d
J
.
T
.
Mi
l
d
e,
“
Seman
t
i
c
Bas
ed
Search
E
n
g
i
n
e
Fo
r
In
d
o
n
e
s
i
a
n
(i
n
Bah
as
a:
Mes
i
n
Pen
car
i
Berb
a
s
i
s
k
a
n
Seman
t
i
k
U
n
t
u
k
Ba
h
as
a
I
n
d
o
n
es
i
a)
,
”
Ju
r
n
a
l
S
i
s
t
em
In
f
o
r
m
a
s
i
M
TI
-
U
II
,
v
o
l
.
4
,
n
o
.
2
.
p
p
.
1
1
0
-
1
1
4
,
2
0
1
2
.
[4
]
A
.
K
ari
m,
“
D
es
i
g
n
an
d
D
e
t
ect
i
o
n
o
f
t
h
e
T
rad
i
t
i
o
n
o
f
H
ad
i
t
h
as
an
In
f
o
rmat
i
o
n
Ret
ri
e
v
a
l
i
n
t
h
e
Bo
o
k
s
o
f
H
a
d
i
t
h
(i
n
Bah
as
a
:
Ran
ca
n
g
Ba
n
g
u
n
Pe
n
d
e
t
ek
s
i
a
n
K
e
s
h
a
h
i
h
an
H
a
d
i
t
s
Se
b
ag
a
i
Seb
u
ah
I
n
fo
rma
t
i
o
n
Re
t
ri
e
v
al
Pa
d
a
K
i
t
a
b
-
K
i
t
ab
H
ad
i
t
s
)
,”
Ju
r
n
a
l
Tek
n
i
k
In
f
o
r
m
a
t
i
ka
,
v
o
l
.
5
,
p
p
.
1
–
2
0
,
2
0
1
2
.
[5
]
R.
N
.
E
d
i
,
“
AS
-
SU
N
N
A
H
(H
A
D
I
T
S)
(A
n
In
g
k
ar
Su
n
n
ah
Fl
o
w
St
u
d
y
)
(i
n
Bah
a
s
a:
AS
-
S
U
N
N
A
H
(H
A
D
I
T
S)(S
u
at
u
K
aj
i
an
A
l
i
ra
n
In
g
k
ar
Su
n
n
a
h
)
)
,
”
A
s
a
s
,
v
o
l
.
6
,
n
o
.
2
,
p
p
.
1
3
2
-
1
4
8
,
2
0
1
4
.
[6
]
D
.
S.
May
l
aw
at
i
an
d
G
.
A
.
P.
Sap
t
aw
at
i
,
“Set
o
f
Freq
u
en
t
W
o
rd
It
em
s
et
s
as
Feat
u
re
Re
p
res
e
n
t
a
t
i
o
n
fo
r
T
ex
t
w
i
t
h
In
d
o
n
e
s
i
a
n
Sl
an
g
,
”
Jo
u
r
n
a
l
o
f
P
h
ys
i
cs
:
C
o
n
f
er
e
n
ce
S
er
i
es
,
v
o
l
.
8
0
1
,
n
o
.
1
,
p
p
.
1
–
6
,
2
0
1
6
.
[7
]
H
.
J
i
aw
e
i
,
M.
K
amb
er,
J
.
H
an
,
M
.
K
amb
er,
an
d
J
.
P
ei
,
"
D
at
a
Mi
n
i
n
g
:
Co
n
ce
p
t
s
an
d
T
ech
n
i
q
u
e
s
,
"
3
rd
E
d
i
t
i
o
n
,
E
l
s
ev
i
er,
2
0
1
2
.
[8
]
J
u
ma
d
i
,
D
.
S.
Ma
y
l
a
w
at
i
,
B.
Su
b
aek
i
,
an
d
T
.
R
i
d
w
an
,
“O
p
i
n
i
o
n
mi
n
i
n
g
o
n
T
w
i
t
t
er
mi
cr
o
b
l
o
g
g
i
n
g
u
s
i
n
g
S
u
p
p
o
r
t
V
ect
o
r
Mach
i
n
e:
Pu
b
l
i
c
o
p
i
n
i
o
n
ab
o
u
t
St
at
e
Is
l
ami
c
U
n
i
v
er
s
i
t
y
o
f
Ban
d
u
n
g
,
”
P
r
o
ceed
i
n
g
s
o
f
2
0
1
6
4
t
h
In
t
er
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
Cy
b
er
a
n
d
IT
S
e
r
vi
ce
M
a
n
a
g
em
e
n
t
,
CITS
M
2
0
1
6
,
2
0
1
6
.
[9
]
D
.
S.
A
.
May
l
a
w
at
i
,
M.
A
.
Ramd
h
an
i
,
A
.
Ra
h
man
,
an
d
W
.
D
armal
a
k
s
a
n
a,
“In
creme
n
t
a
l
t
ech
n
i
q
u
e
w
i
t
h
s
et
o
f
fre
q
u
e
n
t
w
o
r
d
i
t
em
s
et
s
fo
r
mi
n
i
n
g
l
arg
e
In
d
o
n
es
i
an
t
e
x
t
d
at
a,
”
2
0
1
7
5
th
In
t
er
n
a
t
i
o
n
a
l
Co
n
f
er
en
ce
o
n
Cyb
e
r
a
n
d
IT
S
er
vi
ce
M
a
n
a
g
em
e
n
t
,
CITS
M
2
0
1
7
,
2
0
1
7
.
[1
0
]
A
.
A
.
O
k
fa
n
Ri
za
l
Ferd
i
a
n
s
y
ah
,
E
ma
U
t
am
i
,
“
Imp
l
eme
n
t
a
t
i
o
n
o
f
Pr
i
n
c
i
p
a
l
Co
m
p
o
n
en
t
A
n
a
l
y
s
i
s
fo
r
D
i
g
i
t
al
I
mag
e
Ret
ri
e
v
al
Sy
s
t
ems
(i
n
Bah
as
a
:
Imp
l
emen
t
as
i
Pri
n
c
i
p
a
l
Co
mp
o
n
e
n
t
A
n
a
l
y
s
i
s
U
n
t
u
k
Si
s
t
em
T
em
u
Bal
i
k
Ci
t
ra
D
i
g
i
t
al
)
,
”
Cr
ea
t
i
ve
In
f
o
r
m
a
t
i
o
n
Tech
n
o
l
o
g
y
Jo
u
r
n
a
l
,
v
o
l
.
2
,
n
o
.
3
,
2
0
1
5
.
[1
1
]
G
.
K
ary
o
n
o
,
F.
S.
U
t
o
mo
,
A
.
Si
s
t
em,
an
d
T
.
Bal
i
k
,
“
In
f
o
rmat
i
o
n
Ret
ri
e
v
al
i
n
In
d
o
n
e
s
i
a
n
L
an
g
u
ag
e
T
e
x
t
D
o
cu
men
t
s
U
s
i
n
g
t
h
e
V
ect
o
r
Sp
ace
Ret
r
i
ev
al
Mo
d
el
Me
t
h
o
d
(
i
n
Ba
h
as
a
:
T
emu
Ba
l
i
k
In
f
o
rmas
i
Pad
a
D
o
k
u
men
T
ek
s
Berb
a
h
as
a
In
d
o
n
e
s
i
a
D
en
g
an
Me
t
o
d
e
V
ect
o
r
Sp
ace
Re
t
ri
e
v
al
Mo
d
el
)
,
”
S
em
i
n
a
r
Na
s
i
o
n
a
l
Tek
n
o
l
o
g
i
I
n
f
o
r
m
a
s
i
&
K
o
m
u
n
i
ka
s
i
Ter
a
p
a
n
2
0
1
2
(
S
em
a
n
t
i
k
2
0
1
2
)
,
p
p
.
2
8
2
–
2
8
9
,
2
0
1
2
.
[1
2
]
F.
A
mi
n
,
“
In
fo
rmat
i
o
n
Ret
ri
e
v
al
Sy
s
t
em
w
i
t
h
V
ec
t
o
r
Sp
a
ce
Mo
d
el
Ran
k
i
n
g
Met
h
o
d
(i
n
Bah
a
s
a:
Si
s
t
em
T
em
u
K
emb
al
i
In
fo
rma
s
i
d
e
n
g
a
n
Pemeri
n
g
k
at
a
n
Met
o
d
e
V
ec
t
o
r
S
p
ace
Mo
d
e
l
)
,
”
D
i
n
a
m
i
k
,
v
o
l
.
1
8
,
n
o
.
2
,
p
p
.
1
2
2
–
1
2
9
,
2
0
1
3
.
[1
3
]
M.
G
.
O
zs
o
y
,
F.
N
.
A
l
p
a
s
l
a
n
,
an
d
I.
Ci
cek
l
i
,
“T
e
x
t
s
u
mmari
za
t
i
o
n
u
s
i
n
g
L
at
en
t
Seman
t
i
c
A
n
a
l
y
s
i
s
,
”
Jo
u
r
n
a
l
o
f
In
f
o
r
m
a
t
i
o
n
S
c
i
en
ce,
v
o
l
.
3
7
,
n
o
.
4
,
p
p
.
4
0
5
–
4
1
7
,
2
0
1
1
.
[1
4
]
P.
W
.
Fo
l
t
z,
“L
at
e
n
t
s
eman
t
i
c
an
a
l
y
s
i
s
fo
r
t
e
x
t
-
b
as
e
d
res
ea
rc
h
,
”
B
e
h
a
v
i
o
r
R
es
e
a
r
c
h
M
et
h
o
d
s
,
v
o
l
.
2
8
,
n
o
.
2
,
p
p
.
1
9
7
–
2
0
2
,
1
9
9
6
.
[1
5
]
G
.
Co
s
ma
an
d
M.
J
o
y
,
“A
n
A
p
p
r
o
ach
t
o
So
u
rce
-
Co
d
e
Pl
ag
i
ar
i
s
m
D
e
t
ect
i
o
n
an
d
In
v
e
s
t
i
g
a
t
i
o
n
U
s
i
n
g
L
at
en
t
Seman
t
i
c
A
n
a
l
y
s
i
s
,
”
I
E
E
E
Tr
a
n
s
a
ct
i
o
n
s
o
n
Co
m
p
u
t
e
r
s
,
v
o
l
.
6
1
,
n
o
.
3
,
p
p
.
3
7
9
–
3
9
4
,
2
0
1
2
.
[1
6
]
M.
Mo
n
j
u
r
u
l
I
s
l
am
a
n
d
A
.
S.
M.
L
at
i
fu
l
H
o
q
u
e,
“A
u
t
o
ma
t
ed
es
s
ay
s
co
r
i
n
g
u
s
i
n
g
G
en
eral
i
zed
L
at
e
n
t
Sema
n
t
i
c
A
n
a
l
y
s
i
s
,
”
2
0
1
0
1
3
th
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
er
e
n
ce
o
n
Co
m
p
u
t
e
r
a
n
d
In
f
o
r
m
a
t
i
o
n
Tech
n
o
l
o
g
y
(ICCIT)
,
2
0
1
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.