I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
15
,
No
.
1
,
Feb
r
u
ar
y
20
26
,
p
p
.
1
0
0
9
~
1
0
1
8
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
1
.
p
p
1
0
0
9
-
1
0
1
8
1009
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Ins
tag
ra
m influe
ncer c
la
ss
ificat
io
n using
f
ine
-
tu
ned
BE
RT
mo
del
Ni
P
utu
Su
t
ra
m
ia
ni,
Ni
M
a
d
e
Dit
a
Dwik
a
s
a
ri,
I
Ny
o
m
a
n
P
ra
y
a
na
T
risn
a
,
I
Wa
y
a
n Ag
us
Su
ry
a
Da
rm
a
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
m
a
t
i
o
n
Te
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
,
U
d
a
y
a
n
a
U
n
i
v
e
r
si
t
y
,
B
a
l
i
,
I
n
d
o
n
e
si
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
May
17
,
2
0
2
5
R
ev
is
ed
J
an
7
,
2
0
2
6
Acc
ep
ted
J
an
22
,
2
0
2
6
In
flu
e
n
c
e
r
m
a
rk
e
ti
n
g
h
a
s
e
m
e
rg
e
d
a
s
a
p
o
we
rfu
l
stra
teg
y
in
t
o
d
a
y
’s
d
ig
i
tal
wo
rld
,
wh
e
re
so
c
ial
m
e
d
ia
sta
rs
c
a
n
in
flu
e
n
c
e
h
o
w
p
e
o
p
le
t
h
in
k
a
b
o
u
t
p
ro
d
u
c
ts.
Ho
we
v
e
r,
t
h
e
ra
p
id
g
r
o
wth
o
f
i
n
fl
u
e
n
c
e
rs
a
n
d
so
c
ial
m
e
d
ia
u
se
rs
p
re
se
n
ts
n
o
v
e
l
c
h
a
ll
e
n
g
e
s
fo
r
b
r
a
n
d
s
in
id
e
n
ti
fy
i
n
g
su
it
a
b
le
in
fl
u
e
n
c
e
rs
fo
r
th
e
ir
m
a
rk
e
ti
n
g
g
o
a
ls.
Trad
it
io
n
a
l
a
p
p
ro
a
c
h
e
s
th
a
t
re
ly
o
n
p
o
p
u
larity
a
n
d
fo
ll
o
we
r
c
o
u
n
t
a
re
n
o
lo
n
g
e
r
th
e
p
rima
ry
m
e
tri
c
s
fo
r
d
e
ter
m
in
in
g
a
n
in
flu
e
n
c
e
r’s
a
b
il
it
y
to
a
ffe
c
t
c
o
n
su
m
e
r
b
e
h
a
v
i
o
r.
T
o
a
d
d
re
ss
t
h
is
g
a
p
,
t
h
is
stu
d
y
p
r
o
p
o
se
d
a
n
in
f
lu
e
n
c
e
r
c
l
a
ss
ifi
c
a
ti
o
n
to
e
n
h
a
n
c
e
a
u
d
ie
n
c
e
targ
e
ti
n
g
a
n
d
m
a
rk
e
ti
n
g
e
ffe
c
ti
v
e
n
e
ss
.
By
u
ti
li
z
i
n
g
d
e
e
p
lea
rn
in
g
,
s
p
e
c
ifi
c
a
ll
y
fin
e
-
tu
n
e
d
b
id
irec
ti
o
n
a
l
e
n
c
o
d
e
r
re
p
re
se
n
tatio
n
s
fr
o
m
tran
sfo
rm
e
rs
(BERT
),
in
flu
e
n
c
e
r
c
las
sifica
ti
o
n
wa
s
c
a
r
ried
o
u
t
fo
r
I
n
sta
g
ra
m
u
se
rs
in
In
d
o
n
e
sia
b
a
se
d
o
n
th
e
ir
p
o
st
c
a
p
ti
o
n
s.
T
h
e
m
u
lt
il
i
n
g
u
a
l
BERT
m
o
d
e
l
is
o
p
t
imiz
e
d
th
ro
u
g
h
h
y
p
e
rp
a
ra
m
e
ter
tu
n
i
n
g
,
i
n
c
lu
d
in
g
lea
rn
i
n
g
ra
te,
b
a
tch
siz
e
,
a
n
d
sto
p
wo
rd
re
m
o
v
a
l
v
a
riatio
n
.
Wi
th
a
n
o
u
tsta
n
d
i
n
g
8
0
%
a
c
c
u
ra
c
y
,
th
e
m
o
d
e
l
p
e
rfo
rm
s
b
e
st
i
n
sit
u
a
ti
o
n
s
wh
e
re
sto
p
wo
rd
s
a
re
n
o
t
re
m
o
v
e
d
.
T
h
i
s
stu
d
y
o
n
in
flu
e
n
c
e
r
c
las
sifica
ti
o
n
u
si
n
g
a
fin
e
-
tu
n
e
d
BERT
m
o
d
e
l
h
a
s
d
e
m
o
n
stra
ted
th
e
e
ffe
c
ti
v
e
n
e
ss
o
f
BERT
in
e
n
h
a
n
c
in
g
in
fl
u
e
n
c
e
r
se
lec
ti
o
n
.
It
c
o
n
tri
b
u
tes
to
t
h
e
d
i
g
it
a
l
m
a
rk
e
ti
n
g
d
o
m
a
in
b
y
sh
o
wc
a
sin
g
t
h
e
p
o
ten
ti
a
l
o
f
d
e
e
p
lea
rn
i
n
g
fo
r
so
c
ial
m
e
d
ia
a
n
a
l
y
sis
a
n
d
c
o
n
ten
t
c
las
sifica
ti
o
n
,
p
a
v
i
n
g
th
e
wa
y
fo
r
fu
tu
re
d
a
ta
-
d
riv
e
n
m
a
rk
e
ti
n
g
stra
t
e
g
ies
.
K
ey
w
o
r
d
s
:
B
E
R
T
Fin
e
-
tu
n
in
g
I
n
s
tag
r
am
Pre
-
tr
ain
ed
m
o
d
el
So
cial
m
ed
ia
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
I
Ny
o
m
an
Pra
y
an
a
T
r
is
n
a
Dep
ar
tm
en
t o
f
I
n
f
o
r
m
atio
n
T
e
ch
n
o
lo
g
y
,
Facu
lty
o
f
E
n
g
in
ee
r
in
g
,
Ud
ay
a
n
a
Un
iv
er
s
ity
St.
Kam
p
u
s
B
u
k
it Ji
m
b
ar
an
,
K
u
ta
Selatan
,
B
ad
u
n
g
,
B
ali,
I
n
d
o
n
esia
E
m
ail:
p
r
ay
an
a.
t
r
is
n
a@
u
n
u
d
.
a
c.
id
1.
I
NT
RO
D
UCT
I
O
N
T
ec
h
n
o
lo
g
ical
ad
v
an
ce
m
e
n
ts
an
d
s
h
if
ts
in
co
n
s
u
m
er
b
eh
a
v
io
r
h
av
e
led
to
s
ig
n
if
ican
t tr
an
s
f
o
r
m
atio
n
s
in
m
o
d
e
r
n
m
ar
k
etin
g
s
tr
ateg
ie
s
.
On
e
n
o
ta
b
le
in
n
o
v
atio
n
is
t
h
e
em
er
g
en
ce
o
f
i
n
f
lu
en
ce
r
m
ar
k
etin
g
,
wh
ich
h
as
co
m
p
letely
ch
a
n
g
ed
t
h
e
asp
e
ct
o
f
d
ig
ital
m
ar
k
etin
g
b
y
c
o
m
b
in
in
g
th
e
p
o
wer
o
f
wo
r
d
-
of
-
m
o
u
th
(
W
o
M)
m
ar
k
etin
g
an
d
th
e
ex
p
an
s
iv
e
r
ea
ch
o
f
s
o
cial
m
ed
ia.
At
t
h
e
p
r
esen
t
tim
e,
in
f
lu
e
n
ce
r
m
ar
k
etin
g
is
th
e
to
p
ch
o
ice
o
f
m
an
y
b
r
a
n
d
s
in
e
n
g
ag
in
g
t
h
eir
tar
g
et
au
d
ien
ce
s
.
I
n
f
lu
e
n
ce
r
s
ar
e
o
p
in
io
n
lea
d
er
s
wh
o
h
av
e
th
e
ab
ilit
y
an
d
willin
g
n
ess
to
s
w
ay
co
n
s
u
m
er
d
ec
is
io
n
-
m
ak
in
g
ac
r
o
s
s
s
o
cial
m
ed
ia
p
latf
o
r
m
s
lik
e
Yo
u
T
u
b
e,
T
ik
T
o
k
,
an
d
I
n
s
tag
r
a
m
[
1
]
.
I
n
s
tag
r
am
is
th
e
m
o
s
t
p
o
p
u
lar
p
l
atf
o
r
m
f
o
r
in
f
lu
e
n
ce
r
s
am
o
n
g
th
ese
p
latf
o
r
m
s
[
2
]
,
p
r
o
v
id
i
n
g
an
e
x
ce
llen
t
o
p
p
o
r
tu
n
ity
f
o
r
b
r
an
d
s
to
in
cr
ea
s
e
awa
r
en
ess
an
d
v
is
ib
ilit
y
o
f
th
eir
p
r
o
d
u
cts
a
n
d
s
er
v
ices.
As
a
r
esu
lt,
in
f
lu
en
c
er
m
ar
k
etin
g
is
cu
r
r
e
n
tly
o
n
e
o
f
th
e
b
est
m
ar
k
etin
g
s
tr
ateg
i
es,
co
n
s
id
er
in
g
th
at
m
o
d
er
n
co
n
s
u
m
e
r
s
tr
u
s
t
r
ec
o
m
m
en
d
atio
n
s
f
r
o
m
p
eo
p
le
t
h
ey
f
o
llo
w
an
d
ad
m
ir
e
o
n
s
o
cial
m
ed
ia
m
o
r
e
th
an
co
n
v
en
tio
n
al
ad
v
e
r
tis
in
g
[
3
]
.
T
h
e
r
is
e
o
f
in
f
lu
en
ce
r
s
an
d
s
o
cial
m
ed
ia
u
s
er
s
cr
ea
tes
a
m
u
ltifa
ce
ted
s
itu
atio
n
f
o
r
b
r
an
d
s
.
W
h
ile
it
o
f
f
er
s
ex
citin
g
o
p
p
o
r
t
u
n
ities
,
th
is
h
as lef
t m
o
s
t b
r
an
d
s
f
ac
in
g
n
ew
ch
allen
g
es in
ch
o
o
s
in
g
t
h
e
r
ig
h
t in
f
lu
e
n
ce
r
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
1
,
Feb
r
u
ar
y
20
26
:
1
0
0
9
-
1
0
1
8
1010
wh
o
alig
n
with
th
eir
m
ar
k
etin
g
o
b
jectiv
es.
Mo
s
t
tr
ad
itio
n
al
in
f
lu
en
ce
r
s
elec
tio
n
m
eth
o
d
s
ar
e
u
s
u
ally
b
ased
o
n
th
e
n
u
m
b
er
o
f
f
o
llo
wer
s
o
r
p
o
p
u
la
r
ity
[
4
]
.
Ho
wev
er
,
g
iv
en
to
d
a
y
’
s
au
d
ien
ce
s
an
d
co
n
s
u
m
er
b
e
h
av
io
r
d
iv
er
s
ity
,
it tu
r
n
s
o
u
t th
at
r
ely
i
n
g
s
o
lely
o
n
th
e
n
u
m
b
er
o
f
f
o
l
lo
wer
s
d
o
es n
o
t a
cc
u
r
ately
r
ef
l
ec
t a
n
in
f
lu
en
ce
r
’
s
r
ea
l
p
o
ten
tial
to
war
d
s
th
e
tar
g
et
m
ar
k
et
[
5
]
.
Mo
r
e
o
v
er
,
th
e
c
o
n
v
en
tio
n
al
m
eth
o
d
s
f
o
r
in
f
lu
en
ce
r
class
if
icatio
n
o
f
ten
r
ely
o
n
s
o
m
e
s
im
p
le
m
etr
ics o
r
m
an
u
al
an
aly
s
is
.
All
o
f
th
ese
tak
e
tim
e
an
d
m
ay
n
o
t b
e
ac
cu
r
ate.
I
n
lig
h
t
o
f
th
ese
cir
cu
m
s
tan
ce
s
,
b
r
an
d
s
an
d
m
ar
k
eter
s
ar
e
d
e
m
an
d
in
g
m
o
r
e
tar
g
eted
an
d
s
o
p
h
is
ticated
ap
p
r
o
ac
h
es
to
id
en
tify
th
o
s
e
in
f
lu
en
ce
r
s
wh
o
r
ea
lly
d
r
iv
i
n
g
co
n
s
u
m
er
s
.
Mo
s
t
in
f
lu
en
ce
r
s
ten
d
to
h
a
v
e
a
s
p
ec
if
ic
n
ich
e
o
r
m
a
r
k
et
s
eg
m
en
t
in
wh
ich
t
h
ey
ar
e
in
ter
ested
o
r
ex
p
er
ts
,
s
u
ch
as
f
ash
io
n
,
b
ea
u
t
y
,
g
am
in
g
,
tr
a
v
el,
an
d
o
th
er
n
i
ch
es
th
at
r
ep
r
esen
t
th
e
s
o
cial
m
ed
ia
co
n
ten
t
th
ey
s
h
ar
e.
T
h
ese
n
ich
es
g
iv
e
b
r
an
d
s
th
e
o
p
p
o
r
tu
n
ity
to
r
ea
c
h
th
eir
tar
g
et
a
u
d
ien
ce
th
r
o
u
g
h
th
ese
r
elev
a
n
t
in
f
lu
en
ce
r
s
[
6
]
.
I
d
en
tif
y
in
g
i
n
f
lu
en
ce
r
ca
teg
o
r
ies
will
h
elp
b
o
th
b
r
a
n
d
s
an
d
m
ar
k
eter
s
in
ta
r
g
etin
g
a
p
p
r
o
p
r
iate
au
d
ien
ce
s
s
in
ce
th
e
r
elev
a
n
ce
b
etwe
en
in
f
lu
e
n
ce
r
s
an
d
th
e
ad
v
er
tis
ed
p
r
o
d
u
cts
also
af
f
ec
ts
th
e
ef
f
ec
tiv
en
ess
o
f
m
ar
k
etin
g
ca
m
p
aig
n
s
[
7
]
.
T
h
e
r
ig
h
t
in
f
lu
e
n
ce
r
ca
n
b
o
o
s
t
th
e
ef
f
ec
tiv
en
ess
o
f
a
m
ar
k
etin
g
ca
m
p
aig
n
,
wh
ile
th
e
wr
o
n
g
o
n
e
co
u
l
d
lea
d
to
m
in
im
al
im
p
ac
t.
I
t
f
o
llo
ws
th
at
id
en
tify
in
g
i
n
f
lu
en
ce
r
ca
teg
o
r
ies
is
ess
en
tial
f
o
r
m
ar
k
eter
s
to
en
s
u
r
e
th
at
th
ey
a
r
e
wo
r
k
in
g
with
th
e
r
i
g
h
t in
f
l
u
en
ce
r
s
.
Pre
v
io
u
s
s
tu
d
ies
h
av
e
ex
p
lo
r
e
d
th
e
o
p
p
o
r
tu
n
ities
f
o
r
en
h
an
c
in
g
in
f
lu
e
n
ce
r
m
ar
k
etin
g
b
y
e
m
p
lo
y
in
g
m
ac
h
in
e
lear
n
in
g
an
d
d
ee
p
lear
n
in
g
tech
n
i
q
u
es
[
8
]
–
[
1
1
]
.
A
m
o
r
e
co
m
m
o
n
ap
p
r
o
ac
h
a
m
o
n
g
th
e
d
if
f
er
en
t
s
tr
ateg
ies
p
r
o
p
o
s
ed
in
th
e
liter
atu
r
e
is
tex
t
class
if
icatio
n
th
r
o
u
g
h
s
o
cial
m
ed
ia,
wh
er
e
tex
t
i
n
f
o
r
m
atio
n
with
in
u
s
er
s
’
p
o
s
ts
ca
n
co
n
tr
ib
u
te
as
a
v
alu
ab
le
s
o
u
r
ce
.
T
h
e
co
n
ten
t
o
f
in
f
lu
e
n
ce
r
p
o
s
ts
,
in
clu
d
i
n
g
th
e
ca
p
tio
n
s
th
e
y
wr
ite,
ca
n
b
e
b
etter
an
al
y
ze
d
to
p
r
o
v
i
d
e
m
a
r
k
eter
s
in
s
ig
h
t
s
in
to
th
e
in
ter
ests
o
f
th
e
in
f
l
u
en
ce
r
’
s
a
u
d
ien
ce
.
C
o
n
s
eq
u
en
tly
,
f
in
d
i
n
g
s
r
elev
a
n
t
in
f
lu
en
ce
r
s
will
p
o
s
itiv
ely
im
p
ac
t
b
r
an
d
awa
r
e
n
ess
an
d
p
u
r
ch
ase
in
ten
tio
n
s
,
as
th
e
co
n
ten
t
s
h
ar
ed
b
y
in
f
l
u
en
ce
r
s
is
lik
ely
to
d
r
aw
atten
tio
n
an
d
ef
f
ec
tiv
ely
in
f
lu
e
n
ce
a
u
d
ien
ce
s
d
u
e
to
its
s
ig
n
if
ican
t in
f
o
r
m
ativ
e
v
alu
e
[
1
2
]
.
T
ex
t
class
if
icatio
n
is
th
e
p
r
o
ce
s
s
o
f
ca
teg
o
r
izin
g
tex
t
in
t
o
o
n
e
o
r
m
o
r
e
p
r
e
d
ef
in
e
d
class
es
ac
co
r
d
in
g
to
th
eir
r
esp
ec
tiv
e
s
u
b
jects.
T
e
x
t
class
if
icatio
n
th
r
o
u
g
h
s
o
cial
m
ed
ia
ca
n
b
e
r
ea
lized
with
d
if
f
er
en
t
m
o
d
alities
,
o
n
e
o
f
wh
ich
is
n
atu
r
al
lan
g
u
a
g
e
p
r
o
ce
s
s
in
g
(
NL
P).
I
n
th
e
la
s
t
co
u
p
le
o
f
y
ea
r
s
,
th
e
m
o
s
t
r
e
ce
n
t
o
n
e
in
th
e
lin
e
o
f
NL
P
is
u
s
in
g
b
id
ir
ec
tio
n
a
l
en
co
d
er
r
ep
r
esen
tatio
n
s
f
r
o
m
tr
an
s
f
o
r
m
er
s
(
B
E
R
T
)
,
a
m
o
d
el
p
r
o
p
o
s
ed
b
y
Go
o
g
le
in
2
0
1
8
[
1
3
]
.
W
ith
t
h
e
ab
ilit
y
to
u
n
d
e
r
s
tan
d
co
n
tex
tu
al
r
elatio
n
s
h
ip
s
b
etwe
en
wo
r
d
s
[
1
4
]
,
B
E
R
T
r
ep
r
esen
ts
a
tr
an
s
f
o
r
m
er
-
b
ased
ar
ch
itectu
r
e
th
at
h
as a
ch
iev
ed
s
tate
-
of
-
th
e
-
ar
t r
esu
lts
in
a
v
ar
iety
o
f
NL
P task
s
,
in
clu
d
in
g
tex
t
class
if
icatio
n
.
B
E
R
T
’
s
ef
f
ec
tiv
en
ess
in
tex
t
class
if
icatio
n
h
as
b
ee
n
well
-
estab
lis
h
ed
ac
r
o
s
s
v
ar
io
u
s
d
o
m
ain
s
[
1
5
]
–
[
1
7
]
.
Acc
o
r
d
in
g
to
Me
r
ch
an
et
a
l.
[
1
8
]
,
B
E
R
T
also
o
u
tp
er
f
o
r
m
ed
o
t
h
er
tr
a
d
itio
n
al
m
ac
h
in
e
lear
n
in
g
m
eth
o
d
s
in
tex
t c
lass
if
icatio
n
,
in
clu
d
in
g
lo
g
is
tic
r
eg
r
ess
io
n
,
s
u
p
p
o
r
t v
ec
t
o
r
class
if
ier
(
SVC
)
,
m
u
ltin
o
m
ial
NB
,
H2
OAu
to
ML
,
an
d
a
f
ew
o
th
er
m
o
d
els.
I
n
r
eg
a
r
d
t
o
th
e
s
o
cial
m
e
d
i
a
in
f
lu
en
ce
r
class
if
icatio
n
,
s
tu
d
ies
b
y
Kim
et
a
l.
[
1
1
]
h
a
v
e
p
r
o
v
en
B
E
R
T
’
s
ex
ce
llen
t
p
er
f
o
r
m
an
c
e.
T
h
e
s
tu
d
y
em
p
lo
y
ed
a
m
u
l
tim
o
d
al
ap
p
r
o
ac
h
th
at
co
m
b
in
ed
tex
t
an
d
im
ag
e
d
ata
u
s
in
g
p
r
e
-
tr
ain
e
d
B
E
R
T
an
d
I
n
ce
p
tio
n
-
v
3
m
o
d
els,
with
a
p
o
s
t
-
atten
tio
n
lay
er
to
ca
lcu
late
th
e
s
co
r
e
o
f
ea
ch
p
o
s
t
in
d
escr
ib
i
n
g
th
e
ca
teg
o
r
y
o
f
th
e
i
n
f
lu
en
ce
r
.
T
h
is
ap
p
r
o
ac
h
h
as
a
v
e
r
y
g
o
o
d
p
e
r
f
o
r
m
an
ce
in
in
f
lu
en
ce
r
p
r
o
f
ilin
g
.
An
o
th
e
r
s
tu
d
y
co
n
tr
asts
wo
r
d
em
b
ed
d
i
n
g
s
,
Glo
Ve
an
d
Fas
tTe
x
t,
with
f
in
e
-
tu
n
ed
B
E
R
T
,
h
ig
h
lig
h
tin
g
th
e
s
u
p
er
io
r
ity
o
f
B
E
R
T
in
clas
s
if
y
in
g
I
n
s
tag
r
am
u
s
er
in
ter
ests
b
ased
o
n
h
as
h
tag
s
an
d
ca
p
tio
n
s
,
ac
h
iev
in
g
ac
c
u
r
ac
ies o
f
9
6
% a
n
d
9
1
%,
r
esp
ec
tiv
ely
[
1
9
]
.
B
u
ild
in
g
u
p
o
n
th
e
p
r
o
m
is
in
g
r
esu
lts
in
[
1
1
]
,
[
1
9
]
,
t
h
is
s
tu
d
y
p
r
o
p
o
s
es
a
f
in
e
-
t
u
n
ed
B
E
R
T
m
o
d
el
f
o
r
class
if
y
in
g
I
n
s
tag
r
am
in
f
lu
en
c
er
s
in
I
n
d
o
n
esia
b
ased
o
n
th
eir
p
o
s
t
ca
p
tio
n
s
.
W
e
p
r
o
p
o
s
e
to
u
s
e
m
u
ltil
in
g
u
al
B
E
R
T
,
tr
ain
ed
u
n
d
er
v
ar
i
o
u
s
h
y
p
er
p
ar
a
m
eter
tu
n
in
g
s
ce
n
ar
io
s
,
to
f
in
d
th
e
b
est
p
er
f
o
r
m
an
ce
m
o
d
els.
T
h
e
class
if
icatio
n
p
r
o
ce
s
s
in
v
o
lv
e
s
p
r
o
ce
s
s
in
g
th
e
ca
p
tio
n
tex
t
o
f
th
e
in
f
lu
en
ce
r
’
s
p
o
s
t
u
s
in
g
th
e
B
E
R
T
m
o
d
el,
wh
ich
th
en
ca
teg
o
r
izes
in
f
lu
e
n
ce
r
s
in
to
r
elev
an
t
g
r
o
u
p
s
b
a
s
ed
o
n
th
e
r
esu
lt.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
aim
s
to
s
im
p
lify
th
e
id
en
tific
atio
n
an
d
in
f
lu
en
ce
r
s
ea
r
ch
th
at
ar
e
alig
n
ed
with
th
e
b
r
an
d
m
ar
k
eti
n
g
o
b
jectiv
es.
T
h
is
r
esear
ch
is
ex
p
ec
ted
t
o
s
er
v
e
as
a
r
ef
er
en
ce
f
o
r
im
p
lem
en
tin
g
m
a
r
k
etin
g
s
tr
ateg
ies
u
s
in
g
d
ee
p
lear
n
i
n
g
tech
n
o
lo
g
y
.
Fro
m
a
n
ac
ad
em
i
c
p
er
s
p
ec
tiv
e,
th
e
B
E
R
T
m
o
d
e
l is ex
p
ec
ted
to
class
if
y
in
f
lu
e
n
ce
r
s
o
p
tim
ally
an
d
ca
n
b
e
d
e
v
elo
p
e
d
f
o
r
o
th
er
d
ig
ital m
ar
k
etin
g
s
tr
ateg
ies.
2.
M
E
T
H
O
D
T
h
e
s
tu
d
y
is
d
iv
id
ed
in
to
f
iv
e
s
tag
es
.
D
ata
ac
q
u
is
itio
n
,
d
ata
lab
elin
g
,
p
r
e
-
p
r
o
ce
s
s
in
g
,
f
in
e
-
tu
n
in
g
t
h
e
B
E
R
T
m
o
d
el,
an
d
in
f
l
u
en
ce
r
c
lass
if
icatio
n
.
T
h
e
en
tire
wo
r
k
f
lo
w
o
f
th
is
s
tu
d
y
is
s
h
o
wn
i
n
Fig
u
r
e
1
.
2
.
1
.
Da
t
a
a
cquis
it
io
n
T
h
e
d
ataset
we
u
s
ed
in
t
h
is
s
tu
d
y
was
o
b
tain
ed
th
r
o
u
g
h
s
cr
ap
in
g
o
n
s
ev
e
r
al
I
n
s
tag
r
a
m
ac
co
u
n
ts
o
f
I
n
d
o
n
esian
in
f
lu
en
ce
r
s
.
First
o
f
all,
we
g
ath
e
r
ed
a
lis
t
o
f
in
f
lu
en
ce
r
u
s
er
n
am
es,
wh
ich
was
s
o
u
r
ce
d
f
r
o
m
s
ea
r
ch
es
o
n
v
ar
io
u
s
p
latf
o
r
m
s
,
in
clu
d
in
g
I
n
s
tag
r
am
a
n
d
Go
o
g
le.
T
h
ese
u
s
er
n
a
m
es
s
er
v
ed
as
p
ar
am
eter
s
f
o
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
n
s
ta
g
r
a
m
in
flu
en
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
fin
e
-
tu
n
e
d
B
E
R
T
mo
d
el
(
N
i P
u
tu
S
u
tr
a
mia
n
i
)
1011
co
llectin
g
in
f
lu
e
n
ce
r
p
o
s
t
d
ata
.
Fu
r
th
er
,
th
e
I
n
s
talo
ad
er
lib
r
a
r
y
was
u
s
ed
to
s
cr
ap
e
I
n
s
tag
r
am
p
o
s
ts
to
co
llect
d
ata
o
n
in
f
l
u
en
ce
r
p
o
s
ts
.
T
h
e
d
atasets
ac
q
u
ir
ed
co
m
p
r
is
ed
7
,
5
0
0
p
o
s
ts
f
o
r
m
o
d
el
b
u
ild
in
g
an
d
1
9
,
8
4
2
p
o
s
ts
f
o
r
class
if
icatio
n
task
s
,
co
llec
ted
f
r
o
m
th
e
latest
5
0
p
o
s
ts
o
f
1
5
0
an
d
4
0
0
in
f
lu
e
n
ce
r
s
,
r
esp
ec
tiv
ely
.
T
h
e
d
ata
o
b
tain
ed
in
cl
u
d
ed
ess
en
tial te
x
t in
f
o
r
m
atio
n
,
s
u
ch
as u
s
er
n
a
m
e,
p
o
s
t d
ate,
p
o
s
t U
R
L
,
an
d
ca
p
tio
n
.
Fig
u
r
e
1
.
R
esear
ch
o
v
er
v
iew
2
.
2
.
Da
t
a
la
belin
g
T
h
e
class
if
icatio
n
task
in
th
is
s
tu
d
y
is
a
f
o
r
m
o
f
s
u
p
e
r
v
is
ed
lear
n
in
g
,
w
h
er
e
th
e
m
o
d
el
is
tr
ain
ed
t
o
p
r
ed
ict
th
e
co
r
r
ec
t
lab
el
b
ase
d
o
n
in
p
u
t
f
ea
tu
r
es.
T
h
is
p
r
o
ce
s
s
u
tili
ze
s
a
d
ataset
co
m
p
o
s
ed
o
f
f
ea
tu
r
e
-
lab
el
p
air
s
[
2
0
]
.
L
ab
elin
g
is
a
cr
itic
al
p
h
ase
in
th
is
s
tu
d
y
,
as
it
p
r
o
v
id
es
th
e
n
ec
ess
ar
y
lab
els
f
o
r
tr
ain
in
g
th
e
m
o
d
el
an
d
p
er
f
o
r
m
in
g
class
if
icatio
n
.
W
e
m
an
u
ally
lab
eled
7
,
5
0
0
p
o
s
ts
f
r
o
m
3
0
0
in
f
lu
e
n
ce
r
s
,
b
y
co
n
s
id
er
in
g
th
e
co
n
tex
t
o
f
ea
ch
p
o
s
t
ca
p
tio
n
.
T
h
e
lab
els
wer
e
d
eter
m
i
n
ed
b
ased
o
n
th
e
t
o
p
f
i
v
e
ca
teg
o
r
ies
o
f
th
e
m
o
s
t
p
o
p
u
lar
ty
p
es
o
f
c
o
n
ten
t
p
r
o
d
u
ce
d
b
y
I
n
d
o
n
esian
in
f
lu
e
n
ce
r
s
ac
co
r
d
in
g
to
A
n
y
Min
d
:
f
ash
io
n
an
d
b
ea
u
ty
,
en
ter
tain
m
en
t,
f
o
o
d
,
tr
av
el,
an
d
g
am
es
an
d
g
ad
g
ets
[
2
1
]
,
wit
h
o
n
e
o
th
er
ca
teg
o
r
y
f
o
r
p
o
s
ts
th
at
d
id
n
o
t f
it
in
to
th
ese
f
iv
e
ca
teg
o
r
ies.
2
.
3
.
P
re
-
pro
ce
s
s
ing
Scr
ap
in
g
r
esu
lts
ty
p
ically
y
ield
r
aw
d
ata
th
at
r
eq
u
ir
es
f
u
r
th
er
p
r
o
ce
s
s
in
g
f
o
r
ef
f
icien
t
an
aly
s
is
.
T
h
er
ef
o
r
e,
th
e
d
atasets
wer
e
p
r
e
-
p
r
o
ce
s
s
ed
,
w
h
ich
in
cl
u
d
ed
d
eletin
g
p
o
s
ts
with
o
u
t
ca
p
tio
n
s
,
s
p
litt
in
g
h
ash
tag
s
,
ca
s
e
f
o
ld
in
g
,
clea
n
in
g
,
an
d
s
to
p
wo
r
d
s
r
em
o
v
al.
Fo
llo
win
g
th
e
r
em
o
v
al
o
f
em
p
ty
en
tr
ies,
E
n
g
lis
h
h
ash
tag
s
co
n
tain
in
g
m
o
r
e
t
h
an
o
n
e
wo
r
d
ar
e
s
eg
m
en
ted
u
s
in
g
wo
r
d
s
eg
m
en
tatio
n
f
r
o
m
th
e
E
k
p
h
r
asis
lib
r
ar
y
[
2
2
]
to
en
h
an
ce
s
em
a
n
tic
ac
cu
r
ac
y
.
L
o
wer
ca
s
in
g
i
s
th
en
d
o
n
e
to
co
n
v
er
t
all
u
p
p
er
ca
s
e
letter
s
to
lo
wer
ca
s
e,
an
d
clea
n
in
g
is
d
o
n
e
to
r
em
o
v
e
UR
L
s
,
n
u
m
b
er
s
,
s
p
ec
ial
ch
ar
ac
ter
s
,
o
r
o
th
er
i
r
r
elev
an
t
elem
en
ts
.
T
h
e
last
s
tep
is
s
to
p
wo
r
d
r
em
o
v
al.
I
n
th
is
s
tu
d
y
,
we
p
e
r
f
o
r
m
e
d
two
p
r
e
-
p
r
o
ce
s
s
in
g
s
ce
n
ar
io
s
:
with
an
d
with
o
u
t
s
to
p
wo
r
d
r
em
o
v
al.
T
h
e
r
em
o
v
al
o
f
s
to
p
wo
r
d
s
d
ep
en
d
s
o
n
lan
g
u
ag
e
d
etec
tio
n
to
s
p
ec
if
y
an
ap
p
r
o
p
r
iate
s
to
p
wo
r
d
d
ictio
n
ar
y
,
eith
er
in
I
n
d
o
n
esian
o
r
E
n
g
lis
h
,
co
n
s
id
er
in
g
th
e
d
ataset’
s
lin
g
u
is
tic
d
iv
er
s
ity
.
T
ab
le
1
illu
s
tr
ates e
x
am
p
les o
f
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
d
ata
ca
r
r
ied
o
u
t in
th
is
s
tu
d
y
.
2
.
4
.
F
ine t
un
ing
B
E
R
T
h
as
b
ec
o
m
e
th
e
s
tate
-
of
-
th
e
-
ar
t
f
o
r
v
a
r
iety
task
s
s
u
ch
as
q
u
esti
o
n
an
s
wer
in
g
,
lan
g
u
ag
e
u
n
d
er
s
tan
d
i
n
g
,
class
if
icatio
n
,
an
d
o
t
h
er
NL
P
task
s
,
with
o
u
t
s
u
b
s
tan
tial
ch
an
g
es
in
th
e
m
o
d
el
ar
c
h
itectu
r
e.
B
E
R
T
i
s
e
s
s
en
tially
a
s
tack
o
f
tr
an
s
f
o
r
m
er
en
c
o
d
er
lay
er
s
[
2
3
]
with
m
u
ltip
le
s
elf
-
atten
tio
n
“h
ea
d
s
”
.
T
h
ese
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
1
,
Feb
r
u
ar
y
20
26
:
1
0
0
9
-
1
0
1
8
1012
h
ea
d
s
wo
r
k
to
g
eth
er
to
g
en
er
a
te
d
ee
p
b
id
ir
ec
tio
n
al
r
ep
r
esen
tatio
n
s
o
f
u
n
la
b
eled
tex
t
b
y
o
b
s
er
v
in
g
th
e
co
n
tex
t
o
f
ea
ch
lay
e
r
o
n
b
o
t
h
s
id
es (
lef
t a
n
d
r
i
g
h
t)
[
1
3
]
.
T
h
is
p
ap
er
u
s
es
th
e
m
u
ltil
in
g
u
al
B
E
R
T
m
o
d
el,
s
p
ec
if
icall
y
th
e
b
er
t
-
b
ase
-
m
u
ltil
in
g
u
al
-
c
ased
f
r
o
m
Hu
g
g
in
g
f
ac
e
tr
an
s
f
o
r
m
er
s
[
1
3
]
.
Mu
ltil
in
g
u
al
B
E
R
T
was
ch
o
s
en
s
in
ce
it
h
as
b
e
en
p
r
e
-
tr
ai
n
ed
o
n
1
0
4
lan
g
u
ag
es u
s
in
g
th
e
W
ik
ip
ed
ia
d
atasets
,
wh
ich
in
clu
d
e
I
n
d
o
n
esian
an
d
E
n
g
lis
h
.
I
t h
as
p
o
r
tr
ay
ed
e
x
ce
llen
t
ca
p
ab
ilit
y
o
n
cr
o
s
s
-
lan
g
u
ag
e
g
en
er
aliza
tio
n
,
in
cl
u
d
in
g
co
d
e
-
s
witch
in
g
[
2
4
]
.
As
a
r
esu
lt,
th
is
m
o
d
el
ca
n
b
e
ap
p
lied
to
d
atasets
with
a
wid
e
r
an
g
e
o
f
lan
g
u
ag
e
v
a
r
iatio
n
s
,
as
is
o
f
ten
f
o
u
n
d
in
I
n
s
tag
r
am
ca
p
tio
n
s
wh
er
e
m
o
r
e
th
an
o
n
e
lan
g
u
ag
e
is
u
s
ed
.
T
h
e
m
o
d
el
was
th
en
f
in
e
-
t
u
n
ed
with
a
n
ew
d
ataset
to
u
n
d
er
s
tan
d
b
etter
th
e
co
n
tex
t
o
f
wo
r
d
s
a
p
p
ea
r
i
n
g
in
I
n
s
tag
r
am
p
o
s
ts
,
wh
ich
e
n
ab
l
ed
it
to
id
e
n
tify
an
d
g
r
o
u
p
in
f
lu
en
ce
r
p
o
s
ts
in
to
th
eir
ap
p
r
o
p
r
iate
ca
teg
o
r
ies.
Fin
e
-
tu
n
in
g
o
f
th
e
B
E
R
T
m
o
d
el
in
itiates
b
y
s
p
litt
in
g
th
e
lab
eled
d
atasets
in
to
th
r
ee
s
u
b
s
ets:
tr
ain
in
g
s
et,
v
alid
atio
n
s
et,
a
n
d
test
s
et.
T
h
is
s
h
all
b
e
d
o
n
e
in
an
8
0
:
1
0
:1
0
r
atio
,
th
at
is
,
8
0
%
f
o
r
tr
ai
n
in
g
d
ata,
1
0
%
f
o
r
th
e
v
alid
atio
n
d
ata,
an
d
th
e
r
e
m
ain
in
g
1
0
%
f
o
r
test
d
ata.
T
h
e
s
tu
d
y
also
ex
p
er
im
en
ted
w
ith
h
y
p
e
r
p
ar
a
m
eter
ad
ju
s
tm
en
ts
,
in
clu
d
in
g
lear
n
i
n
g
r
ate
an
d
b
atch
s
ize,
an
d
d
i
f
f
er
en
t
s
to
p
wo
r
d
r
em
o
v
al
s
tr
ateg
ies
ap
p
lied
to
th
e
d
ata
to
f
in
d
th
e
b
est
m
o
d
els.
Acc
o
r
d
in
g
to
T
ab
le
2
,
h
y
p
er
p
ar
am
eter
tu
n
in
g
was
test
ed
u
s
in
g
th
ese
ad
ju
s
tm
en
ts
in
6
s
ce
n
ar
io
s
,
in
clu
d
in
g
v
ar
iatio
n
s
in
b
atch
s
izes:
1
6
an
d
3
2
,
a
n
d
lear
n
in
g
r
ate:
2
e
-
5
,
3
e
-
5
,
an
d
5
e
-
5
ac
co
r
d
in
g
to
th
e
o
r
ig
in
al
B
E
R
T
p
ap
er
[
1
3
]
.
Fu
r
th
er
m
o
r
e,
a
n
ad
d
itio
n
al
ex
p
er
im
e
n
t
is
co
n
d
u
cted
b
y
u
tili
zin
g
s
to
p
wo
r
d
r
em
o
v
al,
r
esu
ltin
g
i
n
a
to
tal
o
f
1
2
ex
p
e
r
im
en
ts
f
r
o
m
6
s
ce
n
ar
io
s
.
T
h
is
aim
s
to
f
in
d
th
e
m
o
s
t o
p
tim
al
h
y
p
er
p
ar
am
eter
co
m
b
i
n
atio
n
wh
ile
n
o
ticin
g
th
e
im
p
ac
t
o
f
u
s
in
g
s
to
p
w
o
r
d
r
em
o
v
al
o
n
p
er
f
o
r
m
an
ce
f
o
r
th
e
B
E
R
T
m
o
d
el.
Fin
ally
,
a
lin
ea
r
class
if
icatio
n
lay
er
is
ad
d
ed
o
n
to
p
o
f
th
e
B
E
R
T
o
u
tp
u
t
to
p
r
o
d
u
ce
th
e
n
u
m
b
e
r
o
f
o
u
t
p
u
ts
ac
co
r
d
i
n
g
to
t
h
e
lab
els u
s
ed
.
T
h
e
tr
ain
in
g
s
ess
io
n
s
ar
e
d
ef
in
ed
to
r
u
n
f
o
r
2
5
ep
o
c
h
s
p
e
r
s
ce
n
ar
io
,
with
ea
r
l
y
s
to
p
p
i
n
g
u
s
ed
to
p
r
ev
en
t
o
v
er
f
itti
n
g
.
T
h
en
,
th
e
s
e
m
o
d
els
will
b
e
ev
alu
ated
u
s
in
g
s
ev
er
al
m
etr
ics
s
u
ch
as
a
cc
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
to
d
eter
m
in
e
th
e
b
est
m
o
d
el.
T
h
e
f
in
al
p
er
f
o
r
m
an
ce
ev
al
u
atio
n
is
co
n
d
u
cted
o
n
th
e
test
d
ata,
an
d
t
h
e
m
o
d
el
with
th
e
b
est p
er
f
o
r
m
a
n
ce
will th
en
b
e
a
p
p
lied
to
th
e
u
n
lab
ele
d
d
ata.
T
ab
le
1
.
E
x
am
p
le
o
f
p
r
e
-
p
r
o
ce
s
s
in
g
d
ata
La
n
g
u
a
g
e
C
a
p
t
i
o
n
Pre
-
p
r
o
c
e
ssi
n
g
p
h
a
se
C
l
e
a
n
c
a
p
t
i
o
n
En
g
l
i
sh
W
h
a
t
h
a
p
p
e
n
e
d
i
n
N
e
p
a
l
?
F
u
l
l
e
p
i
s
o
d
e
a
v
a
i
l
a
b
l
e
i
n
Y
o
u
T
u
b
e
.
Li
n
k
i
n
b
i
o
l
i
k
e
,
sh
a
r
e
a
n
d
c
o
mm
e
n
t
p
l
e
a
se
#
n
e
p
a
l
#
j
u
n
g
l
e
#
h
i
k
i
n
g
a
d
v
e
n
t
u
r
e
s
#
t
r
a
v
e
l
st
o
r
i
e
s
#
i
n
d
o
n
e
s
i
a
n
b
a
c
k
p
a
c
k
e
r
#
b
a
c
k
p
a
c
k
e
r
#
i
n
d
o
n
e
si
a
S
p
l
i
t
h
a
s
h
t
a
g
W
h
a
t
h
a
p
p
e
n
e
d
i
n
N
e
p
a
l
?
f
u
l
l
e
p
i
s
o
d
e
a
v
a
i
l
a
b
l
e
i
n
Y
o
u
T
u
b
e
.
L
i
n
k
i
n
b
i
o
l
i
k
e
,
s
h
a
r
e
a
n
d
c
o
mm
e
n
t
p
l
e
a
se
n
e
p
a
l
j
u
n
g
l
e
h
i
k
i
n
g
a
d
v
e
n
t
u
r
e
s
t
r
a
v
e
l
st
o
r
i
e
s
i
n
d
o
n
e
si
a
n
b
a
c
k
p
a
c
k
e
r
b
a
c
k
p
a
c
k
e
r
i
n
d
o
n
e
si
a
C
a
se
f
o
l
d
i
n
g
w
h
a
t
h
a
p
p
e
n
e
d
i
n
n
e
p
a
l
?
f
u
l
l
e
p
i
so
d
e
a
v
a
i
l
a
b
l
e
i
n
y
o
u
t
u
b
e
.
l
i
n
k
i
n
b
i
o
l
i
k
e
,
sh
a
r
e
a
n
d
c
o
mm
e
n
t
p
l
e
a
se
n
e
p
a
l
j
u
n
g
l
e
h
i
k
i
n
g
a
d
v
e
n
t
u
r
e
s
t
r
a
v
e
l
st
o
r
i
e
s
i
n
d
o
n
e
si
a
n
b
a
c
k
p
a
c
k
e
r
b
a
c
k
p
a
c
k
e
r
i
n
d
o
n
e
si
a
Te
x
t
c
l
e
a
n
i
n
g
w
h
a
t
h
a
p
p
e
n
e
d
i
n
n
e
p
a
l
f
u
l
l
e
p
i
s
o
d
e
a
v
a
i
l
a
b
l
e
i
n
y
o
u
t
u
b
e
l
i
n
k
i
n
b
i
o
l
i
k
e
s
h
a
r
e
a
n
d
c
o
mm
e
n
t
p
l
e
a
s
e
n
e
p
a
l
j
u
n
g
l
e
h
i
k
i
n
g
a
d
v
e
n
t
u
r
e
s
t
r
a
v
e
l
st
o
r
i
e
s
i
n
d
o
n
e
si
a
n
b
a
c
k
p
a
c
k
e
r
b
a
c
k
p
a
c
k
e
r
i
n
d
o
n
e
s
i
a
S
t
o
p
w
o
r
d
r
e
mo
v
a
l
h
a
p
p
e
n
e
d
n
e
p
a
l
f
u
l
l
e
p
i
s
o
d
e
a
v
a
i
l
a
b
l
e
y
o
u
t
u
b
e
l
i
n
k
b
i
o
l
i
k
e
sh
a
r
e
c
o
m
me
n
t
p
l
e
a
s
e
n
e
p
a
l
j
u
n
g
l
e
h
i
k
i
n
g
a
d
v
e
n
t
u
r
e
s
t
r
a
v
e
l
s
t
o
r
i
e
s
i
n
d
o
n
e
si
a
n
b
a
c
k
p
a
c
k
e
r
b
a
c
k
p
a
c
k
e
r
i
n
d
o
n
e
s
i
a
I
n
d
o
n
e
si
a
n
A
v
a
i
l
a
b
l
e
n
o
w
:
2
0
.
0
8
.
2
3
K
o
l
e
k
si
j
a
k
e
t
#
K
e
m
b
a
l
i
M
e
r
d
e
k
a
su
d
a
h
ri
l
i
s se
c
a
r
a
o
f
f
i
c
i
a
l
di
w
e
b
si
t
e
c
a
r
d
i
n
a
l
.
c
o
.
i
d
C
u
m
a
a
d
a
5
0
p
c
s
y
a
n
g
b
i
s
a
k
a
m
u
d
a
p
a
t
k
a
n
d
i
l
i
n
k
i
n
i
:
h
t
t
p
s
:
/
/
sh
o
r
t
u
r
l
.
a
t
/
d
x
B
D
S
G
e
t
y
o
u
r
s!
S
p
l
i
t
h
a
s
h
t
a
g
A
v
a
i
l
a
b
l
e
n
o
w
:
2
0
.
0
8
.
2
3
K
o
l
e
k
s
i
j
a
k
e
t
k
e
m
b
a
l
i
m
e
rd
e
k
a
s
u
d
a
h
ri
l
i
s
se
c
a
r
a
o
f
f
i
c
i
a
l
di
w
e
b
si
t
e
c
a
r
d
i
n
a
l
.
c
o
.
i
d
C
u
m
a
a
d
a
5
0
p
c
s
y
a
n
g
b
i
s
a
k
a
m
u
d
a
p
a
t
k
a
n
d
i
l
i
n
k
i
n
i
:
h
t
t
p
s
:
/
/
sh
o
r
t
u
r
l
.
a
t
/
d
x
B
D
S
G
e
t
y
o
u
r
s!
C
a
se
f
o
l
d
i
n
g
a
v
a
i
l
a
b
l
e
n
o
w
:
2
0
.
0
8
.
2
3
k
o
l
e
k
si
j
a
k
e
t
k
e
m
b
a
l
i
m
e
rd
e
k
a
s
u
d
a
h
ri
l
i
s
se
c
a
r
a
o
f
f
i
c
i
a
l
di
w
e
b
si
t
e
c
a
r
d
i
n
a
l
.
c
o
.
i
d
c
u
m
a
a
d
a
5
0
p
c
s
y
a
n
g
b
i
s
a
k
a
m
u
d
a
p
a
t
k
a
n
d
i
l
i
n
k
i
n
i
:
h
t
t
p
s
:
/
/
sh
o
r
t
u
r
l
.
a
t
/
d
x
b
d
s
g
e
t
y
o
u
r
s!
Te
x
t
c
l
e
a
n
i
n
g
a
v
a
i
l
a
b
l
e
n
o
w
k
o
l
e
k
si
j
a
k
e
t
k
e
m
b
a
l
i
m
e
rd
e
k
a
s
u
d
a
h
ri
l
i
s se
c
a
r
a
o
f
f
i
c
i
a
l
di
w
e
b
si
t
e
c
a
r
d
i
n
a
l
c
o
i
d
c
u
m
a
a
d
a
p
c
s
y
a
n
g
b
i
sa
k
a
m
u
d
a
p
a
t
k
a
n
d
i
l
i
n
k
i
n
i
g
e
t
y
o
u
r
s
S
t
o
p
w
o
r
d
r
e
mo
v
a
l
a
v
a
i
l
a
b
l
e
n
o
w
k
o
l
e
k
si
j
a
k
e
t
m
e
rd
e
k
a
r
i
l
i
s
o
f
f
i
c
i
a
l
w
e
b
si
t
e
c
a
r
d
i
n
a
l
c
o
i
d
c
u
m
a
p
c
s
k
a
m
u
d
a
p
a
t
k
a
n
l
i
n
k
g
e
t
y
o
u
r
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
n
s
ta
g
r
a
m
in
flu
en
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
fin
e
-
tu
n
e
d
B
E
R
T
mo
d
el
(
N
i P
u
tu
S
u
tr
a
mia
n
i
)
1013
T
ab
le
2
.
B
E
R
T
h
y
p
e
r
p
ar
am
ete
r
tu
n
in
g
s
ce
n
ar
io
S
c
e
n
a
r
i
o
B
a
t
c
h
si
z
e
Le
a
r
n
i
n
g
r
a
t
e
1
16
2e
-
5
2
16
3e
-
5
3
16
5e
-
5
4
32
2e
-
5
5
32
3e
-
5
6
32
5e
-
5
2
.
5
.
I
nfluence
r
cla
s
s
if
ica
t
io
n
T
o
class
if
y
in
f
l
u
en
ce
r
s
b
ased
o
n
th
ei
r
co
n
ten
t,
a
f
i
n
e
-
tu
n
e
d
m
u
ltil
in
g
u
al
B
E
R
T
m
o
d
el
wa
s
f
ir
s
t
u
s
ed
to
ca
teg
o
r
ize
u
n
lab
eled
p
o
s
ts
f
r
o
m
in
f
lu
en
ce
r
s
.
T
h
en
,
a
s
im
p
le
alg
o
r
ith
m
is
u
s
ed
t
o
a
s
s
ig
n
a
d
o
m
in
an
t
ca
teg
o
r
y
f
o
r
ea
ch
in
f
lu
en
ce
r
.
T
h
is
alg
o
r
ith
m
ca
lcu
lates
th
e
f
r
eq
u
e
n
cy
o
f
th
e
p
r
e
-
d
ef
in
ed
ca
teg
o
r
ies
b
ased
o
n
th
e
B
E
R
T
m
o
d
el’
s
p
o
s
t
clas
s
i
f
icatio
n
s
,
an
d
th
e
ca
teg
o
r
y
with
th
e
h
ig
h
est
f
r
eq
u
en
c
y
will
b
e
co
n
s
id
er
ed
as
th
e
in
f
lu
en
ce
r
’
s
p
r
im
ar
y
n
ic
h
e.
I
t
is
im
p
o
r
tan
t
to
n
o
te
th
at
th
e
o
t
h
er
ca
teg
o
r
y
was
ex
clu
d
ed
as
it
lack
ed
a
s
p
ec
if
ic
f
o
cu
s
an
d
lim
ited
in
s
ig
h
ts
in
t
o
in
f
l
u
en
ce
r
an
d
au
d
ien
ce
p
r
e
f
er
en
ce
s
.
T
h
e
f
o
cu
s
r
em
ain
s
o
n
f
iv
e
well
-
d
ef
i
n
ed
ca
teg
o
r
ies:
f
ash
io
n
an
d
b
ea
u
ty
,
en
ter
tain
m
en
t,
f
o
o
d
,
tr
av
el,
a
n
d
g
am
es a
n
d
g
a
d
g
ets
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
will
ex
p
lain
t
h
e
r
esu
lts
o
f
m
o
d
el
s
elec
tio
n
,
m
o
d
el
e
v
alu
atio
n
,
an
d
in
f
lu
en
ce
r
class
if
icatio
n
th
at
h
av
e
b
ee
n
c
ar
r
ied
o
u
t
in
th
is
s
tu
d
y
.
M
o
d
el
s
elec
tio
n
is
th
e
f
ir
s
t
s
tep
in
d
eter
m
in
in
g
th
e
b
est
m
o
d
el
o
n
tr
ain
in
g
an
d
v
alid
atio
n
d
ata,
wh
ile
m
o
d
el
ev
a
lu
atio
n
ex
am
in
es
th
e
p
er
f
o
r
m
an
ce
o
f
t
h
e
b
est
p
r
ev
io
u
s
ly
s
elec
ted
m
o
d
el
o
n
test
d
ata.
Fin
ally
,
th
e
b
est
m
o
d
els
ar
e
ap
p
lied
to
u
n
lab
eled
i
n
f
lu
en
ce
r
d
ata.
T
h
e
d
is
cu
s
s
io
n
s
o
f
ea
ch
ar
e
p
r
esen
ted
as f
o
llo
ws
.
3
.
1
.
M
o
del selec
t
io
n
Sev
er
al
m
u
ltil
in
g
u
al
B
E
R
T
tr
ain
in
g
s
ce
n
ar
i
o
s
wer
e
r
u
n
wi
th
d
if
f
e
r
en
t
p
ar
am
eter
co
m
b
i
n
atio
n
s
to
f
in
d
t
h
e
b
est
m
o
d
el
with
r
e
g
ar
d
to
a
n
in
f
lu
en
ce
r
p
o
s
t
c
lass
if
icatio
n
task
.
T
h
is
h
as
b
ee
n
p
e
r
f
o
r
m
ed
f
o
r
1
2
e
x
p
er
im
e
n
ts
d
ep
e
n
d
in
g
o
n
th
e
p
ar
a
m
eter
s
d
escr
ib
e
d
in
T
ab
le
2
.
T
ab
le
3
s
h
o
ws
th
e
ex
p
er
im
en
tal
r
esu
lts
o
f
B
E
R
T
p
er
f
o
r
m
an
ce
o
n
v
alid
atio
n
d
ata
f
o
r
ea
c
h
f
in
e
-
t
u
n
i
n
g
s
ce
n
ar
io
.
Acc
o
r
d
in
g
t
o
th
e
f
in
d
in
g
s
,
m
o
d
el
s
ce
n
ar
io
5
with
b
atch
s
ize
o
f
3
2
a
n
d
a
lear
n
in
g
r
ate
o
f
3
e
-
5
p
er
f
o
r
m
s
th
e
b
est
o
n
d
ata
with
o
u
t
s
to
p
wo
r
d
r
em
o
v
al.
T
h
is
m
o
d
el
r
ea
ch
e
d
th
e
m
o
s
t
o
u
ts
tan
d
in
g
p
er
f
o
r
m
an
ce
c
o
m
p
ar
e
d
to
o
th
er
s
ce
n
ar
io
s
,
with
an
ac
cu
r
ac
y
o
f
8
0
% a
n
d
a
lo
s
s
o
f
0
.
0
2
9
.
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
o
f
B
E
R
T
p
er
f
o
r
m
an
ce
o
n
v
alid
atio
n
d
ata
S
c
e
n
a
r
i
o
B
a
t
c
h
si
z
e
Le
a
r
n
i
n
g
r
a
t
e
S
t
o
p
w
o
r
d
s r
e
mo
v
a
l
W
i
t
h
o
u
t
st
o
p
w
o
r
d
s r
e
mo
v
a
l
Lo
ss
A
c
c
u
r
a
c
y
(
%)
Lo
ss
A
c
c
u
r
a
c
y
(
%)
1
16
2e
-
5
0
.
0
6
5
7
7
.
0
2
0
.
0
6
0
7
6
.
6
1
2
16
3e
-
5
0
.
0
6
6
7
7
.
0
2
0
.
0
6
2
7
7
.
8
4
3
16
5e
-
5
0
.
0
6
8
7
5
.
1
0
0
.
0
6
3
7
6
.
7
4
4
32
2e
-
5
0
.
0
3
4
7
5
.
3
8
0
.
0
3
1
7
6
.
8
8
5
32
3e
-
5
0
.
0
3
5
7
6
.
0
6
0
.
0
2
9
8
0
.
0
3
6
32
5e
-
5
0
.
0
3
1
7
4
.
9
7
0
.
0
2
9
7
9
.
0
7
Fu
r
th
e
r
m
o
r
e
,
t
h
e
r
e
was
a
g
e
n
er
al
t
r
e
n
d
t
h
at
m
o
d
els
t
r
a
in
e
d
o
n
a
b
at
ch
s
iz
e
o
f
3
2
h
a
d
s
i
g
n
if
ic
an
tl
y
d
i
f
f
er
e
n
t
l
o
s
s
v
a
lu
es
c
o
m
p
a
r
e
d
t
o
t
h
o
s
e
tr
ai
n
e
d
wit
h
a
b
atc
h
s
i
ze
o
f
1
6
.
A
n
o
t
h
e
r
t
h
i
n
g
t
o
b
e
o
b
s
e
r
v
e
d
is
t
h
e
r
e
wer
e
s
ig
n
i
f
i
ca
n
t
d
if
f
er
en
ce
s
i
n
p
er
f
o
r
m
a
n
c
e
b
etw
ee
n
m
o
d
els
tr
ai
n
e
d
o
n
d
ata
w
it
h
an
d
wit
h
o
u
t
s
t
o
p
w
o
r
d
r
e
m
o
v
al.
T
h
e
f
in
d
i
n
g
s
s
h
o
w
t
h
at
m
o
d
e
ls
t
r
a
in
ed
o
n
d
ata
wit
h
o
u
t
s
t
o
p
w
o
r
d
r
em
o
v
al
c
o
n
s
is
t
en
t
ly
o
u
t
p
er
f
o
r
m
e
d
m
o
d
els
t
r
ai
n
ed
o
n
d
at
a
wit
h
s
t
o
p
wo
r
d
r
em
o
v
a
l.
Us
u
all
y
,
in
class
i
f
i
ca
t
io
n
t
as
k
s
,
s
t
o
p
w
o
r
d
s
wi
ll
b
e
r
e
m
o
v
ed
in
o
r
d
er
t
o
all
o
w
f
o
c
u
s
o
n
th
o
s
e
w
o
r
d
s
t
h
a
t
ar
e
m
o
r
e
im
p
o
r
ta
n
t
t
o
t
h
e
te
x
t
,
t
o
i
m
p
r
o
v
e
te
x
t
a
n
al
y
s
is
wit
h
co
m
p
u
tati
o
n
a
l
e
f
f
ici
e
n
c
y
[
2
5
]
.
H
o
w
ev
er
,
i
n
t
h
is
ca
s
e,
i
t
was
f
o
u
n
d
t
h
at
r
e
m
o
v
i
n
g
s
to
p
w
o
r
d
s
s
li
g
h
tl
y
d
r
o
p
p
e
d
th
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
B
E
R
T
m
o
d
el
c
o
m
p
a
r
e
d
t
o
t
r
ai
n
i
n
g
f
r
o
m
c
o
m
p
let
el
y
o
r
i
g
i
n
al
t
ex
t
w
it
h
o
u
t
s
to
p
w
o
r
d
r
e
m
o
v
al.
T
h
e
r
ea
s
o
n
m
a
y
b
e
r
elat
e
d
t
o
t
h
e
m
ec
h
an
is
m
o
f
t
h
e
B
E
R
T
,
wh
ic
h
r
e
q
u
i
r
es
t
h
e
c
o
n
te
x
t
o
f
t
h
e
w
h
o
le
s
en
t
e
n
ce
[
1
4
]
in
o
r
d
er
to
u
n
d
er
s
t
a
n
d
th
e
p
r
o
p
e
r
s
e
n
s
e
o
f
it
.
I
t
m
ea
n
s
t
h
at
r
em
o
v
i
n
g
s
t
o
p
w
o
r
d
s
r
e
d
u
c
es
t
h
e
am
o
u
n
t
o
f
co
n
t
ex
t
an
d
,
c
o
r
r
es
p
o
n
d
in
g
l
y
,
t
h
e
m
o
d
el
’
s
ac
c
u
r
a
c
y
b
ec
o
m
es
less
i
n
t
e
x
t
class
if
ic
ati
o
n
.
3
.
2
.
M
o
del e
v
a
lua
t
i
o
n
E
v
a
l
u
a
t
i
o
n
o
n
t
h
e
b
es
t
m
o
d
e
l
w
a
s
c
o
n
d
u
c
t
e
d
t
o
a
s
s
es
s
t
h
e
m
o
d
e
l
’
s
p
e
r
f
o
r
m
a
n
c
e
i
n
cl
a
s
s
i
f
y
in
g
t
e
x
t
i
n
t
o
t
h
e
a
p
p
r
o
p
r
i
a
t
e
c
a
t
e
g
o
r
i
e
s
.
T
h
e
e
v
a
l
u
a
t
i
o
n
p
r
o
c
e
s
s
w
as
p
e
r
f
o
r
m
e
d
o
n
7
3
1
t
e
s
t
d
a
t
a
,
u
s
i
n
g
c
o
n
f
u
s
i
o
n
m
a
t
r
i
x
m
e
t
r
i
c
s
s
u
c
h
as
a
c
c
u
r
a
c
y
,
p
r
e
c
is
i
o
n
,
r
e
c
a
ll
,
a
n
d
F
1
-
s
c
o
r
e
.
T
h
e
r
e
s
u
l
ts
o
f
t
h
is
e
v
a
l
u
a
ti
o
n
a
r
e
s
h
o
w
n
i
n
T
a
b
l
e
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
1
,
Feb
r
u
ar
y
20
26
:
1
0
0
9
-
1
0
1
8
1014
T
ab
le
4
.
Mo
d
el
p
er
f
o
r
m
a
n
ce
e
v
alu
atio
n
La
b
e
l
s
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
sc
o
r
e
F
a
sh
i
o
n
a
n
d
b
e
a
u
t
y
8
0
%
0
.
8
2
0
.
8
5
0
.
8
4
En
t
e
r
t
a
i
n
me
n
t
0
.
6
3
0
.
4
5
0
.
5
3
F
o
o
d
0
.
8
3
0
.
9
3
0
.
8
8
Tr
a
v
e
l
0
.
7
8
0
.
8
8
0
.
8
2
G
a
mes
a
n
d
g
a
d
g
e
t
0
.
9
2
0
.
8
1
0
.
8
6
O
t
h
e
r
s
0
.
7
7
0
.
7
2
0
.
7
4
T
h
e
ev
alu
atio
n
r
esu
lts
s
h
o
wed
th
at
th
e
m
o
d
el
m
ain
tain
ed
g
o
o
d
p
er
f
o
r
m
an
ce
with
a
n
ac
cu
r
ac
y
o
f
8
0
%
o
n
th
e
test
d
ata,
s
u
g
g
esti
n
g
th
at
th
e
m
o
d
el
is
q
u
ite
ef
f
ec
tiv
e
at
class
if
y
in
g
th
e
m
ajo
r
ity
o
f
tex
t
ca
p
tio
n
s
in
to
th
e
ap
p
r
o
p
r
iate
ca
teg
o
r
ie
s
.
Fu
r
th
er
m
o
r
e
,
b
ased
o
n
th
e
p
r
ec
is
io
n
,
r
ec
all,
a
n
d
F1
-
s
co
r
e
m
etr
ics
u
s
ed
t
o
ev
alu
ate
th
e
m
o
d
el
’
s
p
er
f
o
r
m
a
n
ce
f
o
r
ea
c
h
ca
teg
o
r
y
,
v
a
r
iatio
n
s
in
th
e
m
o
d
el
p
e
r
f
o
r
m
an
ce
wer
e
f
o
u
n
d
in
ea
c
h
ca
teg
o
r
y
.
B
ased
o
n
th
ese
f
in
d
in
g
s
,
we
co
n
d
u
cted
a
q
u
alitativ
e
an
aly
s
is
o
f
th
e
m
o
d
el’
s
p
er
f
o
r
m
a
n
ce
f
o
r
ea
ch
ca
teg
o
r
y
.
Fig
u
r
e
2
s
h
o
ws
th
e
co
n
f
u
s
io
n
m
atr
i
x
,
th
at
co
m
p
ar
es
th
e
m
o
d
el'
s
p
r
ed
icted
v
al
u
es
with
th
e
ac
tu
a
l
test
d
ata
v
alu
es.
Fig
u
r
e
2
.
C
o
n
f
u
s
io
n
m
atr
i
x
o
f
m
o
d
el
class
if
icatio
n
r
esu
lts
W
h
en
ex
am
in
in
g
ea
c
h
ca
teg
o
r
y
,
“
f
o
o
d
”
a
n
d
“g
am
es
an
d
g
ad
g
et
”
p
er
f
o
r
m
e
d
th
e
b
est,
with
th
e
m
ajo
r
ity
o
f
ac
cu
r
ate
p
r
ed
ictio
n
s
,
r
esu
ltin
g
in
h
ig
h
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
es.
T
h
is
in
d
icate
s
th
at
th
e
m
o
d
el
is
v
er
y
ef
f
ec
ti
v
e
at
r
ec
o
g
n
izin
g
an
d
class
if
y
in
g
p
o
s
ts
r
elate
d
t
o
f
o
o
d
a
n
d
g
ad
g
ets.
Ho
wev
er
,
s
o
m
e
ca
teg
o
r
ies
s
h
o
we
d
lo
w
er
p
er
f
o
r
m
a
n
ce
an
d
in
ac
c
u
r
a
te
lab
els,
af
f
ec
tin
g
th
e
o
v
er
a
ll
m
o
d
el
ac
c
u
r
ac
y
.
E
x
am
p
les o
f
in
c
o
r
r
ec
t m
o
d
el
p
r
ed
ictio
n
s
o
n
test
d
ata
ca
n
b
e
s
ee
n
in
T
ab
le
5
.
T
h
e
co
n
f
u
s
io
n
m
atr
i
x
in
Fig
u
r
e
2
h
ig
h
lig
h
ts
s
ig
n
if
ican
t
c
h
allen
g
es
in
th
e
class
if
icatio
n
p
r
o
ce
s
s
.
I
t
s
h
o
ws
th
at
th
e
“
en
te
r
tain
m
en
t”
ca
teg
o
r
y
ex
p
er
ien
ce
s
a
h
ig
h
r
ate
o
f
f
alse
n
eg
ativ
es
an
d
s
ig
n
if
ican
t
co
n
f
u
s
io
n
with
th
e
“
o
th
er
s
”
ca
teg
o
r
y
,
in
d
icatin
g
th
at
t
h
e
m
o
d
el
s
tr
u
g
g
le
d
to
ac
c
u
r
ately
id
en
tify
an
d
d
is
tin
g
u
is
h
with
in
th
ese
ty
p
es
o
f
co
n
ten
t.
E
n
te
r
tain
m
en
t
p
o
s
ts
f
r
eq
u
en
tly
lack
ex
p
licit
ca
p
tio
n
s
an
d
in
s
tead
r
ely
o
n
v
is
u
al
co
n
ten
t
with
c
ap
tio
n
s
th
at
u
s
e
co
m
m
o
n
wo
r
d
s
with
o
u
t
clea
r
p
atter
n
s
,
as
s
h
o
wn
in
s
ev
er
al
ex
am
p
les
in
T
ab
le
5
.
M
o
r
eo
v
er
,
b
ec
au
s
e
e
n
ter
tain
m
en
t
is
s
u
b
jectiv
e
an
d
d
ep
en
d
s
o
n
in
d
i
v
id
u
al
p
e
r
ce
p
tio
n
,
it
m
ay
b
e
d
if
f
icu
lt
to
f
u
lly
d
escr
ib
e
th
is
co
n
ten
t
u
s
in
g
j
u
s
t
wo
r
d
s
.
C
o
n
v
er
s
ely
,
t
h
e
“
o
th
e
r
s
”
c
ateg
o
r
y
h
ad
is
s
u
es
with
b
o
th
h
ig
h
f
alse n
eg
ativ
es
an
d
f
alse p
o
s
itiv
es,
s
h
o
win
g
th
at
th
e
m
o
d
el
also
s
tr
u
g
g
led
t
o
class
if
y
co
n
ten
t in
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
n
s
ta
g
r
a
m
in
flu
en
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
fin
e
-
tu
n
e
d
B
E
R
T
mo
d
el
(
N
i P
u
tu
S
u
tr
a
mia
n
i
)
1015
th
is
ca
teg
o
r
y
.
T
h
is
ca
teg
o
r
y
co
v
er
s
a
wid
e
r
an
g
e
o
f
to
p
ic
s
,
m
ak
in
g
it
m
o
r
e
ch
allen
g
in
g
f
o
r
th
e
m
o
d
el
to
class
if
y
th
em
co
r
r
ec
tly
.
T
h
e
o
v
er
lap
b
etwe
en
th
ese
two
ca
te
g
o
r
ies
is
also
d
u
e
to
th
eir
u
s
e
o
f
co
m
m
o
n
wo
r
d
s
,
wh
ich
f
u
r
th
er
c
o
m
p
licates
ac
cu
r
ate
class
if
icatio
n
.
T
h
ese
ch
allen
g
es
co
n
tr
ib
u
te
to
th
e
m
o
d
el’
s
o
v
er
all
ac
cu
r
ac
y
r
ate
o
f
ar
o
u
n
d
8
0
%,
r
ef
lectin
g
th
e
in
h
e
r
en
t
ch
allen
g
es
in
class
if
y
in
g
s
u
b
jectiv
e
an
d
d
iv
er
s
e
co
n
ten
t
with
tex
tu
al
cu
es a
lo
n
e.
T
ab
le
5
.
E
x
am
p
le
o
f
in
co
r
r
ec
t
m
o
d
el
p
r
e
d
ictio
n
C
a
p
t
i
o
n
Tr
u
e
l
a
b
e
l
s
P
r
e
d
i
c
t
e
d
l
a
b
e
l
s
H
i
d
u
p
l
a
g
i
c
a
p
e
k
-
c
a
p
e
k
n
y
a
m
a
l
a
h
k
e
n
a
p
r
a
n
k
g
e
m
b
e
l
!!
U
n
t
u
n
g
g
u
e
t
e
t
e
p
e
n
j
o
y
s
o
a
l
n
y
a
a
d
a
@
p
u
c
u
k
h
a
r
u
mi
d
#
S
e
g
e
ri
n
D
i
Ma
n
a
A
j
a
#
K
e
s
e
g
a
r
a
n
U
n
t
u
k
S
e
m
u
a
#
D
i
t
e
m
e
n
i
n
T
e
h
P
u
c
u
k
H
a
r
u
m
En
t
e
r
t
a
i
n
me
n
t
F
o
o
d
J
a
n
g
a
n
d
i
t
o
n
t
o
n
s
a
m
p
e
a
k
h
i
r
a
j
a
En
t
e
r
t
a
i
n
me
n
t
O
t
h
e
r
s
v
l
o
g
p
t
2
!
En
t
e
r
t
a
i
n
me
n
t
O
t
h
e
r
s
L
a
g
i
n
g
o
b
ro
l
a
p
a
h
a
y
o
s
a
m
a
p
a
k
b
o
s
S
u
r
a
b
a
y
a
.
.
@
r
o
y
a
d
i
d
h
a
r
m
a
O
t
h
e
r
s
En
t
e
r
t
a
i
n
me
n
t
Be
g
i
n
i
n
i
h
l
i
k
a
l
i
k
u
u
n
t
u
k
p
e
n
g
i
r
i
m
a
n
j
u
a
l
a
n
n
y
a
@
g
r
a
t
u
.
i
d
,
u
n
t
u
n
g
a
d
a
so
l
u
si
n
y
a
!
#
G
o
S
e
n
d
C
a
r
#
B
e
st
S
e
l
l
e
r
G
o
S
e
n
d
O
t
h
e
r
s
F
o
o
d
K
a
l
o
a
j
a
j
a
m
a
n
i
t
u
u
d
a
h
a
d
a
G
r
a
b
j
i
a
n
,
m
u
n
g
k
i
n
g
a
k
a
k
a
n
a
d
a
i
s
t
i
l
a
h
“
c
i
n
t
a
k
u
b
e
r
a
t
d
i
o
n
g
k
o
s
”.
J
a
l
a
n
-
j
a
l
a
n
k
e
m
a
n
a
p
u
n
n
a
i
k
G
r
a
b
B
i
k
e
,
d
i
s
k
o
n
5
0
%
d
e
n
g
a
n
k
o
d
e
p
ro
m
o
G
R
A
B
JI
A
N
.
@
g
r
a
b
i
d
O
t
h
e
r
s
Tr
a
v
e
l
S
p
o
n
t
a
n
e
o
u
s
n
i
g
h
t
st
r
o
l
l
i
n
Tr
a
v
e
l
O
t
h
e
r
s
3
.
3
.
I
nfluence
r
cla
s
s
if
ica
t
io
n r
esu
lt
s
Usi
n
g
o
u
r
f
in
e
-
tu
n
e
d
B
E
R
T
m
o
d
el,
we
class
if
ied
4
0
0
I
n
d
o
n
esian
in
f
lu
en
ce
r
s
with
a
to
tal
o
f
1
9
,
8
4
2
u
n
lab
eled
p
o
s
ts
to
id
e
n
tify
t
h
eir
ca
teg
o
r
ies.
T
h
is
class
if
icat
io
n
p
r
o
ce
s
s
r
esu
lted
in
ca
teg
o
r
y
p
r
ed
ictio
n
s
f
o
r
ea
ch
p
o
s
t
an
d
a
f
in
al
ca
teg
o
r
y
f
o
r
ea
ch
in
f
lu
e
n
ce
r
.
T
h
e
r
esu
l
ts
d
em
o
n
s
tr
ate
th
at
o
u
r
m
o
d
el
p
er
f
o
r
m
e
d
well
in
class
if
y
in
g
in
f
lu
en
ce
r
p
o
s
ts
in
to
r
elev
an
t c
ateg
o
r
ies.
T
h
e
p
o
s
t c
lass
if
icatio
n
r
esu
lts
ar
e
s
h
o
wn
in
Fig
u
r
e
3
.
Fig
u
r
e
3
.
I
n
f
lu
e
n
ce
r
p
o
s
t c
lass
if
icatio
n
r
esu
lts
B
ased
o
n
th
e
p
o
s
t
class
if
icat
io
n
r
esu
lts
in
Fig
u
r
e
3
,
th
e
m
ajo
r
ity
o
f
p
o
s
ts
f
all
in
to
th
e
“
o
th
er
s
”
ca
teg
o
r
y
,
f
o
llo
wed
b
y
“
f
o
o
d
”,
“tr
av
el”,
“f
ash
io
n
an
d
b
ea
u
ty
”,
“e
n
ter
tain
m
e
n
t”,
an
d
“
g
am
es
an
d
g
ad
g
et
”
ca
teg
o
r
ies.
T
h
e
d
o
m
in
an
c
e
o
f
th
e
“
o
th
er
s
”
ca
teg
o
r
y
is
d
u
e
to
th
e
v
ar
iatio
n
in
in
f
lu
en
ce
r
co
n
ten
t,
as
m
o
s
t
in
f
lu
en
ce
r
s
h
av
e
p
o
s
ts
o
u
ts
id
e
th
eir
m
ai
n
n
ic
h
e,
s
u
ch
as
a
b
o
u
t
th
ei
r
p
e
r
s
o
n
al
o
r
m
o
r
e
g
en
er
al
to
p
ics.
Fr
o
m
th
ese
p
o
s
t
class
if
icatio
n
r
esu
lt
s
,
we
ca
n
th
en
d
eter
m
i
n
e
th
e
ca
teg
o
r
y
f
o
r
ea
ch
in
f
lu
en
ce
r
.
T
h
e
d
is
tr
ib
u
tio
n
o
f
4
0
0
I
n
d
o
n
esian
in
f
lu
e
n
ce
r
s
ac
r
o
s
s
ea
ch
ca
teg
o
r
y
ca
n
b
e
s
ee
n
in
Fig
u
r
e
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
1
,
Feb
r
u
ar
y
20
26
:
1
0
0
9
-
1
0
1
8
1016
Fig
u
r
e
4
.
I
n
f
lu
e
n
ce
r
class
if
icatio
n
r
esu
lts
4.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
p
r
o
p
o
s
ed
an
i
n
f
l
u
en
ce
r
class
if
icatio
n
m
eth
o
d
u
s
in
g
B
E
R
T
.
T
h
e
m
o
d
el
d
ev
elo
p
m
en
t
p
r
o
ce
s
s
,
wh
ich
in
clu
d
e
d
h
y
p
er
p
ar
am
eter
tu
n
in
g
a
n
d
s
to
p
wo
r
d
r
em
o
v
al
ex
p
e
r
im
en
t,
a
ch
iev
ed
s
ig
n
if
ican
t
ac
cu
r
ac
y
in
ca
teg
o
r
izin
g
p
o
s
ts
in
to
s
ix
r
elev
an
t
ca
teg
o
r
ies:
f
ash
io
n
an
d
b
ea
u
ty
,
e
n
ter
tain
m
en
t,
f
o
o
d
,
tr
av
el,
g
am
es
an
d
g
ad
g
et,
an
d
o
th
er
s
.
Ou
r
b
est
-
p
er
f
o
r
m
in
g
m
o
d
el,
with
an
8
0
%
ac
cu
r
ac
y
in
p
o
s
t
class
if
icatio
n
,
was
tr
ain
ed
with
a
b
atch
s
ize
o
f
3
2
,
lear
n
in
g
r
ate
o
f
3
e
-
5
,
a
n
d
d
ata
with
o
u
t
s
to
p
wo
r
d
s
r
em
o
v
ed
.
T
h
is
s
tu
d
y
also
d
em
o
n
s
tr
ates
th
at
u
s
in
g
B
E
R
T
f
o
r
te
x
t
class
if
icatio
n
ef
f
ec
tiv
ely
ca
teg
o
r
izes
4
0
0
I
n
d
o
n
esian
in
f
lu
e
n
ce
r
s
b
ased
o
n
th
eir
p
o
s
t
ca
p
tio
n
s
.
T
h
is
alig
n
s
with
o
u
r
i
n
itial
g
o
al
o
f
c
ateg
o
r
izin
g
d
iv
er
s
e
in
f
lu
e
n
ce
r
co
n
ten
t,
o
f
f
er
i
n
g
a
s
o
p
h
is
ticated
an
d
tar
g
eted
a
p
p
r
o
ac
h
to
in
f
l
u
en
ce
r
id
en
tif
icatio
n
.
L
o
o
k
in
g
f
o
r
war
d
,
th
e
f
in
d
i
n
g
s
o
f
th
is
r
esear
ch
p
r
esen
t
s
ig
n
if
ican
t
o
p
p
o
r
tu
n
ities
f
o
r
f
u
r
th
er
d
e
v
elo
p
m
en
t
an
d
ap
p
licatio
n
.
Ho
wev
er
,
b
ec
a
u
s
e
ca
p
tio
n
s
o
n
ly
p
r
o
v
id
e
a
lim
it
ed
am
o
u
n
t
o
f
in
f
o
r
m
atio
n
,
th
ey
m
ay
n
o
t
b
e
s
u
f
f
icien
t.
Fu
t
u
r
e
s
tu
d
ies
s
h
o
u
ld
co
n
s
id
er
in
co
r
p
o
r
atin
g
ad
d
iti
o
n
al
d
ata
s
o
u
r
ce
s
,
s
u
ch
as
i
m
ag
es,
v
id
eo
c
o
n
ten
t,
o
r
en
g
ag
em
en
t
m
etr
ics,
to
cr
ea
te
a
m
o
r
e
c
o
m
p
r
eh
en
s
iv
e
i
n
f
lu
en
ce
r
class
if
icatio
n
s
y
s
te
m
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
t
r
ib
u
to
r
R
o
les
T
a
x
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
i
d
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Ni
Pu
tu
Su
tr
am
ian
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Ni
M
a
d
e
Dit
a
Dwi
k
as
ar
i
✓
✓
✓
✓
✓
✓
✓
✓
I
Ny
o
m
an
Pra
y
an
a
T
r
is
n
a
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
I
W
ay
an
Ag
u
s
Su
r
y
a
Dar
m
a
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
n
s
ta
g
r
a
m
in
flu
en
ce
r
cla
s
s
ifica
tio
n
u
s
in
g
fin
e
-
tu
n
e
d
B
E
R
T
mo
d
el
(
N
i P
u
tu
S
u
tr
a
mia
n
i
)
1017
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
i
n
g
s
o
f
th
is
s
tu
d
y
ar
e
av
ailab
le
f
r
o
m
th
e
co
r
r
esp
o
n
d
in
g
au
th
o
r
,
[
I
NPT]
,
u
p
o
n
r
ea
s
o
n
ab
le
r
eq
u
est.
RE
F
E
R
E
NC
E
S
[
1
]
H.
-
C
.
Li
n
,
P
.
F
.
B
r
u
n
i
n
g
,
a
n
d
H
.
S
w
a
r
n
a
,
“
U
s
i
n
g
o
n
l
i
n
e
o
p
i
n
i
o
n
l
e
a
d
e
r
s
t
o
p
r
o
m
o
t
e
t
h
e
h
e
d
o
n
i
c
a
n
d
u
t
i
l
i
t
a
r
i
a
n
v
a
l
u
e
o
f
p
r
o
d
u
c
t
s
a
n
d
s
e
r
v
i
c
e
s
,
”
Bu
s
i
n
e
ss H
o
ri
z
o
n
s
,
v
o
l
.
6
1
,
n
o
.
3
,
p
p
.
4
3
1
–
4
4
2
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
u
s
h
o
r
.
2
0
1
8
.
0
1
.
0
1
0
.
[
2
]
H
y
p
e
A
u
d
i
t
o
r
,
“
S
t
a
t
e
o
f
i
n
f
l
u
e
n
c
e
r
mark
e
t
i
n
g
2
0
2
4
I
n
d
o
n
e
s
i
a
,
”
H
y
p
e
A
u
d
i
t
o
r
.
A
c
c
e
sse
d
:
F
e
b
.
1
1
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
h
y
p
e
a
u
d
i
t
o
r
.
c
o
m/
r
e
so
u
r
c
e
s/
w
h
i
t
e
p
a
p
e
r
s/
st
a
t
e
-
of
-
i
n
f
l
u
e
n
c
e
-
mark
e
t
i
n
g
-
i
n
d
o
n
e
s
i
a
-
2
0
2
4
/
[
3
]
R
.
E
d
e
l
m
a
n
,
I
n
b
r
a
n
d
s
w
e
t
ru
s
t
?
2
0
1
9
e
d
e
l
m
a
n
t
r
u
s
t
b
a
r
o
m
e
t
e
r s
p
e
c
i
a
l
re
p
o
rt
.
C
h
i
c
a
g
o
,
U
n
i
t
e
d
S
t
a
t
e
s
:
E
d
e
l
ma
n
,
2
0
1
9
.
[
4
]
S
.
W
i
e
s
,
A
.
B
l
e
i
e
r
,
a
n
d
A
.
Ed
e
l
i
n
g
,
“
F
i
n
d
i
n
g
g
o
l
d
i
l
o
c
k
s
i
n
f
l
u
e
n
c
e
r
s
:
h
o
w
f
o
l
l
o
w
e
r
c
o
u
n
t
d
r
i
v
e
s
s
o
c
i
a
l
me
d
i
a
e
n
g
a
g
e
m
e
n
t
,
”
J
o
u
rn
a
l
o
f
M
a
rke
t
i
n
g
,
v
o
l
.
8
7
,
n
o
.
3
,
p
p
.
3
8
3
–
4
0
5
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
7
7
/
0
0
2
2
2
4
2
9
2
2
1
1
2
5
1
3
1
.
[
5
]
M
.
D
.
V
e
i
r
ma
n
,
V
.
C
a
u
b
e
r
g
h
e
,
a
n
d
L
.
H
u
d
d
e
r
s,
“
M
a
r
k
e
t
i
n
g
t
h
r
o
u
g
h
I
n
st
a
g
r
a
m i
n
f
l
u
e
n
c
e
r
s:
t
h
e
i
m
p
a
c
t
o
f
n
u
m
b
e
r
o
f
f
o
l
l
o
w
e
r
s a
n
d
p
r
o
d
u
c
t
d
i
v
e
r
g
e
n
c
e
o
n
b
r
a
n
d
a
t
t
i
t
u
d
e
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
d
v
e
rt
i
si
n
g
,
v
o
l
.
3
6
,
n
o
.
5
,
p
p
.
7
9
8
–
8
2
8
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
8
0
/
0
2
6
5
0
4
8
7
.
2
0
1
7
.
1
3
4
8
0
3
5
.
[
6
]
C
.
G
u
a
n
a
n
d
E.
Y
.
Li
,
“
A
n
o
t
e
o
n
i
n
f
l
u
e
n
c
e
r
mar
k
e
t
i
n
g
i
n
s
o
c
i
a
l
me
d
i
a
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
t
e
r
n
e
t
M
a
rk
e
t
i
n
g
a
n
d
Ad
v
e
rt
i
si
n
g
,
v
o
l
.
1
5
,
n
o
.
2
,
p
p
.
1
2
3
–
1
3
0
,
2
0
2
1
.
[
7
]
L.
J
a
n
s
se
n
,
A
.
P
.
S
c
h
o
u
t
e
n
,
a
n
d
E.
A
.
J.
C
r
o
e
s
,
“
I
n
f
l
u
e
n
c
e
r
a
d
v
e
r
t
i
si
n
g
o
n
I
n
st
a
g
r
a
m
:
p
r
o
d
u
c
t
-
i
n
f
l
u
e
n
c
e
r
f
i
t
a
n
d
n
u
m
b
e
r
o
f
f
o
l
l
o
w
e
r
s
a
f
f
e
c
t
a
d
v
e
r
t
i
si
n
g
o
u
t
c
o
m
e
s
a
n
d
i
n
f
l
u
e
n
c
e
r
e
v
a
l
u
a
t
i
o
n
s
v
i
a
c
r
e
d
i
b
i
l
i
t
y
a
n
d
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Ad
v
e
rt
i
si
n
g
,
v
o
l
.
4
1
,
n
o
.
1
,
p
p
.
1
0
1
–
1
2
7
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
8
0
/
0
2
6
5
0
4
8
7
.
2
0
2
1
.
1
9
9
4
2
0
5
.
[
8
]
B
.
E
.
E
l
b
a
g
h
a
z
a
o
u
i
,
M
.
A
m
n
a
i
,
a
n
d
Y
.
F
a
k
h
r
i
,
“
D
a
t
a
p
r
o
f
i
l
i
n
g
a
n
d
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
t
o
i
d
e
n
t
i
f
y
i
n
f
l
u
e
n
c
e
r
s
f
r
o
m
so
c
i
a
l
me
d
i
a
p
l
a
t
f
o
r
ms,
”
J
o
u
rn
a
l
o
f
I
C
T
S
t
a
n
d
a
r
d
i
z
a
t
i
o
n
,
v
o
l
.
1
0
,
n
o
.
2
,
p
p
.
2
0
1
–
2
1
8
,
M
a
y
2
0
2
2
,
d
o
i
:
1
0
.
1
3
0
5
2
/
j
i
c
t
s
2
2
4
5
-
8
0
0
X
.
1
0
2
6
.
[
9
]
S
.
S
.
R
a
n
i
a
n
d
S
.
S
.
B
a
b
y
,
“
R
e
a
l
-
t
i
me
i
n
f
l
u
e
n
c
e
r
d
e
t
e
c
t
i
o
n
i
n
Tw
i
t
t
e
r
u
si
n
g
a
h
y
b
r
i
d
a
p
p
r
o
a
c
h
,
”
Pro
c
e
d
i
a
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
2
1
5
,
p
p
.
4
6
1
–
4
7
0
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
r
o
c
s.
2
0
2
2
.
1
2
.
0
4
8
.
[
1
0
]
B
.
B
a
s
h
a
r
i
a
n
d
E.
F
.
-
Er
si
,
“
I
n
f
l
u
e
n
t
i
a
l
p
o
s
t
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
n
I
n
st
a
g
r
a
m
t
h
r
o
u
g
h
c
a
p
t
i
o
n
a
n
d
h
a
s
h
t
a
g
a
n
a
l
y
s
i
s,
”
Me
a
su
r
e
m
e
n
t
a
n
d
C
o
n
t
r
o
l
,
v
o
l
.
5
3
,
n
o
.
3
–
4
,
p
p
.
4
0
9
–
4
1
5
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
7
7
/
0
0
2
0
2
9
4
0
1
9
8
7
7
4
8
9
.
[
1
1
]
S
.
K
i
m,
J
.
-
Y
.
Ji
a
n
g
,
M
.
N
a
k
a
d
a
,
J.
H
a
n
,
a
n
d
W
.
W
a
n
g
,
“
M
u
l
t
i
mo
d
a
l
p
o
s
t
a
t
t
e
n
t
i
v
e
p
r
o
f
i
l
i
n
g
f
o
r
i
n
f
l
u
e
n
c
e
r
mark
e
t
i
n
g
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
T
h
e
W
e
b
C
o
n
f
e
re
n
c
e
2
0
2
0
,
2
0
2
0
,
p
p
.
2
8
7
8
–
2
8
8
4
,
d
o
i
:
1
0
.
1
1
4
5
/
3
3
6
6
4
2
3
.
3
3
8
0
0
5
2
.
[
1
2
]
C
.
L
o
u
a
n
d
S
.
Y
u
a
n
,
“
I
n
f
l
u
e
n
c
e
r
mar
k
e
t
i
n
g
:
h
o
w
m
e
ssa
g
e
v
a
l
u
e
a
n
d
c
r
e
d
i
b
i
l
i
t
y
a
f
f
e
c
t
c
o
n
s
u
mer t
r
u
s
t
o
f
b
r
a
n
d
e
d
c
o
n
t
e
n
t
o
n
s
o
c
i
a
l
med
i
a
,
”
J
o
u
r
n
a
l
o
f
I
n
t
e
r
a
c
t
i
v
e
A
d
v
e
rt
i
si
n
g
,
v
o
l
.
1
9
,
n
o
.
1
,
p
p
.
5
8
–
7
3
,
F
e
b
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
8
0
/
1
5
2
5
2
0
1
9
.
2
0
1
8
.
1
5
3
3
5
0
1
.
[
1
3
]
J.
D
e
v
l
i
n
,
M
.
-
W
.
C
h
a
n
g
,
K
.
L
e
e
,
a
n
d
K
.
T
o
u
t
a
n
o
v
a
,
“
B
ER
T
:
p
r
e
-
t
r
a
i
n
i
n
g
o
f
d
e
e
p
b
i
d
i
r
e
c
t
i
o
n
a
l
t
r
a
n
sf
o
r
m
e
r
s
f
o
r
l
a
n
g
u
a
g
e
u
n
d
e
r
s
t
a
n
d
i
n
g
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
N
A
AC
L
-
H
L
T
2
0
1
9
,
2
0
1
9
,
p
p
.
4
1
7
1
–
4
1
8
6
.
[
1
4
]
R
.
A
.
R
a
j
a
g
e
d
e
,
“
I
mp
r
o
v
i
n
g
a
u
t
o
ma
t
i
c
e
ssa
y
sc
o
r
i
n
g
f
o
r
I
n
d
o
n
e
si
a
n
l
a
n
g
u
a
g
e
u
si
n
g
s
i
mp
l
e
r
mo
d
e
l
a
n
d
r
i
c
h
e
r
f
e
a
t
u
r
e
,
”
K
i
n
e
t
i
k
:
G
a
m
e
T
e
c
h
n
o
l
o
g
y
,
I
n
f
o
rm
a
t
i
o
n
S
y
st
e
m
,
C
o
m
p
u
t
e
r
N
e
t
w
o
rk,
C
o
m
p
u
t
i
n
g
,
E
l
e
c
t
r
o
n
i
c
s,
a
n
d
C
o
n
t
ro
l
,
v
o
l
.
6
,
n
o
.
1
,
p
p
.
1
1
–
1
8
,
2
0
2
1
,
d
o
i
:
1
0
.
2
2
2
1
9
/
k
i
n
e
t
i
k
.
v
6
i
1
.
1
1
9
6
.
[
1
5
]
B
.
Ju
a
r
t
o
a
n
d
Y
u
l
i
a
n
t
o
,
“
I
n
d
o
n
e
s
i
a
n
n
e
w
s
c
l
a
ssi
f
i
c
a
t
i
o
n
u
si
n
g
I
n
d
o
B
e
r
t
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
i
n
E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
1
,
n
o
.
2
,
p
p
.
4
5
4
–
4
6
0
,
2
0
2
3
.
[
1
6
]
H
.
S
.
W
i
c
a
k
sa
n
a
,
R
.
K
u
s
u
ma
n
i
n
g
r
u
m,
a
n
d
R
.
G
e
r
n
o
w
o
,
“
D
e
t
e
r
mi
n
i
n
g
c
o
mm
u
n
i
t
y
h
a
p
p
i
n
e
ss
i
n
d
e
x
w
i
t
h
t
r
a
n
sf
o
r
m
e
r
s
a
n
d
a
t
t
e
n
t
i
o
n
-
b
a
se
d
d
e
e
p
l
e
a
r
n
i
n
g
,
”
I
AE
S
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
3
,
n
o
.
2
,
p
p
.
1
7
5
3
–
1
7
6
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
a
i
.
v
1
3
.
i
2
.
p
p
1
7
5
3
-
1
7
6
1
.
[
1
7
]
N
.
K
.
N
i
ss
a
a
n
d
E.
Y
u
l
i
a
n
t
i
,
“
M
u
l
t
i
-
l
a
b
e
l
t
e
x
t
c
l
a
s
si
f
i
c
a
t
i
o
n
o
f
I
n
d
o
n
e
si
a
n
c
u
st
o
m
e
r
r
e
v
i
e
w
s
u
s
i
n
g
b
i
d
i
r
e
c
t
i
o
n
a
l
e
n
c
o
d
e
r
r
e
p
r
e
s
e
n
t
a
t
i
o
n
s fr
o
m t
r
a
n
sf
o
r
mers
l
a
n
g
u
a
g
e
m
o
d
e
l
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r E
n
g
i
n
e
e
r
i
n
g
,
v
o
l
.
1
3
,
n
o
.
5
,
p
p
.
5
6
4
1
–
5
6
5
2
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
c
e
.
v
1
3
i
5
.
p
p
5
6
4
1
-
5
6
5
2
.
[
1
8
]
E.
C
.
G
.
M
e
r
c
h
a
n
,
R
.
G
.
B
r
i
z
u
e
l
a
,
a
n
d
S
.
G
.
-
C
a
r
v
a
j
a
l
,
“
C
o
m
p
a
r
i
n
g
B
ER
T
a
g
a
i
n
s
t
t
r
a
d
i
t
i
o
n
a
l
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
m
o
d
e
l
s
i
n
t
e
x
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
a
t
i
o
n
a
l
a
n
d
C
o
g
n
i
t
i
v
e
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
2
,
n
o
.
4
,
p
p
.
3
5
2
–
3
5
6
,
2
0
2
3
,
d
o
i
:
1
0
.
4
7
8
5
2
/
b
o
n
v
i
e
w
j
c
c
e
3
2
0
2
8
3
8
.
[
1
9
]
S
.
H
a
m
d
i
,
A
.
H
a
m
d
i
,
a
n
d
S
.
B
.
Y
a
h
i
a
,
“
B
E
R
T
a
n
d
w
o
r
d
e
m
b
e
d
d
i
n
g
f
o
r
i
n
t
e
r
e
st
m
i
n
i
n
g
o
f
I
n
st
a
g
r
a
m
u
sers
,
”
i
n
A
d
v
a
n
c
e
s
i
n
C
o
m
p
u
t
a
t
i
o
n
a
l
C
o
l
l
e
c
t
i
v
e
I
n
t
e
l
l
i
g
e
n
c
e
,
2
0
2
2
,
p
p
.
1
2
3
–
136
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
1
-
1
6
2
1
0
-
7
_
1
0
.
[
2
0
]
V
.
N
a
s
t
e
s
k
i
,
“
A
n
o
v
e
r
v
i
e
w
o
f
t
h
e
s
u
p
e
r
v
i
se
d
mac
h
i
n
e
l
e
a
r
n
i
n
g
m
e
t
h
o
d
s,
”
H
O
RI
Z
O
N
S
.
B
,
v
o
l
.
4
,
p
p
.
5
1
–
6
2
,
2
0
1
7
,
d
o
i
:
1
0
.
2
0
5
4
4
/
h
o
r
i
z
o
n
s
.
b
.
0
4
.
1
.
1
7
.
p
0
5
.
[
2
1
]
A
n
y
M
i
n
d
,
“
I
n
f
l
u
e
n
c
e
r
mark
e
t
i
n
g
i
n
I
n
d
o
n
e
s
i
a
mar
k
e
t
o
v
e
r
v
i
e
w
2
0
2
1
,
”
A
n
y
Mi
n
d
.
A
c
c
e
ss
e
d
:
S
e
p
.
2
6
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
a
n
y
mi
n
d
g
r
o
u
p
.
c
o
m
/
[
2
2
]
C
.
B
a
z
i
o
t
i
s,
N
.
P
e
l
e
k
i
s,
a
n
d
C
.
D
o
u
l
k
e
r
i
d
i
s,
“
D
a
t
a
S
t
o
r
i
e
s
a
t
S
e
m
Ev
a
l
-
2
0
1
7
t
a
sk
4
:
d
e
e
p
LST
M
w
i
t
h
a
t
t
e
n
t
i
o
n
f
o
r
mess
a
g
e
-
l
e
v
e
l
a
n
d
t
o
p
i
c
-
b
a
se
d
se
n
t
i
m
e
n
t
a
n
a
l
y
si
s
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s o
f
t
h
e
1
1
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
W
o
rks
h
o
p
o
n
S
e
m
a
n
t
i
c
E
v
a
l
u
a
t
i
o
n
(
S
e
m
Ev
a
l
-
2
0
1
7
)
,
2
0
1
7
,
p
p
.
7
4
7
–
7
5
4
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
s1
7
-
2
1
2
6
.
[
2
3
]
A
.
V
a
sw
a
n
i
e
t
a
l
.
,
“
A
t
t
e
n
t
i
o
n
i
s
a
l
l
y
o
u
n
e
e
d
,
”
i
n
3
1
s
t
C
o
n
f
e
re
n
c
e
o
n
N
e
u
ra
l
I
n
f
o
rm
a
t
i
o
n
Pr
o
c
e
ssi
n
g
S
y
st
e
m
s
(
N
I
PS
2
0
1
7
,
2
0
2
5
,
p
p
.
1
–
11
,
d
o
i
:
1
0
.
6
5
2
1
5
/
2
q
5
8
a
4
2
6
.
[
2
4
]
T.
P
i
r
e
s,
E.
S
c
h
l
i
n
g
e
r
,
a
n
d
D
.
G
a
r
r
e
t
t
e
,
“
H
o
w
m
u
l
t
i
l
i
n
g
u
a
l
i
s
mu
l
t
i
l
i
n
g
u
a
l
B
ER
T?
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
5
7
t
h
An
n
u
a
l
M
e
e
t
i
n
g
o
f
t
h
e
Ass
o
c
i
a
t
i
o
n
f
o
r
C
o
m
p
u
t
a
t
i
o
n
a
l
L
i
n
g
u
i
st
i
c
s
,
2
0
1
9
,
p
p
.
4
9
9
6
–
5
0
0
1
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
p
1
9
-
1
4
9
3
.
[
2
5
]
D
.
J.
L
a
d
a
n
i
a
n
d
N
.
P
.
D
e
sa
i
,
“
S
t
o
p
w
o
r
d
i
d
e
n
t
i
f
i
c
a
t
i
o
n
a
n
d
r
e
mo
v
a
l
t
e
c
h
n
i
q
u
e
s
o
n
TC
a
n
d
I
R
a
p
p
l
i
c
a
t
i
o
n
s
:
a
s
u
r
v
e
y
,
”
i
n
2
0
2
0
6
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
Ad
v
a
n
c
e
d
C
o
m
p
u
t
i
n
g
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
S
y
st
e
m
s
(
I
C
AC
C
S
)
,
2
0
2
0
,
p
p
.
4
6
6
–
4
7
2
,
d
o
i
:
1
0
.
1
1
0
9
/
i
c
a
c
c
s4
8
7
0
5
.
2
0
2
0
.
9
0
7
4
1
6
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
1
,
Feb
r
u
ar
y
20
26
:
1
0
0
9
-
1
0
1
8
1018
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ni
Pu
tu
S
u
tr
a
m
i
a
n
i
re
c
e
iv
e
d
a
b
a
c
h
e
lo
r'
s
d
e
g
re
e
in
Co
m
p
u
t
e
r
S
y
ste
m
s
fro
m
S
TIKOM
Ba
li
i
n
2
0
1
1
,
a
m
a
ste
r'
s
d
e
g
re
e
in
In
f
o
rm
a
ti
o
n
S
y
ste
m
s
a
n
d
C
o
m
p
u
ter
M
a
n
a
g
e
m
e
n
t
fro
m
Ud
a
y
a
n
a
Un
iv
e
rsity
in
2
0
1
5
,
a
n
d
a
d
o
c
to
ra
l
d
e
g
re
e
in
Co
m
p
u
ter
S
c
ien
c
e
a
t
th
e
De
p
a
rtme
n
t
o
f
I
n
fo
rm
a
ti
c
s,
I
n
st
it
u
t
Tek
n
o
lo
g
i
S
e
p
u
lu
h
No
p
e
m
b
e
r
in
2
0
2
2
.
He
r
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
c
o
m
p
u
ter
v
isio
n
,
ima
g
e
p
ro
c
e
ss
in
g
,
a
n
d
a
rti
fici
a
l
in
telli
g
e
n
c
e
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
su
tram
ian
i@
u
n
u
d
.
a
c
.
id
.
Ni
Ma
d
e
Dita
Dw
ik
a
sa
r
i
h
o
l
d
s
a
b
a
c
h
e
lo
r'
s
d
e
g
re
e
in
I
n
fo
rm
a
ti
o
n
Tec
h
n
o
l
o
g
y
fro
m
th
e
F
a
c
u
lt
y
o
f
E
n
g
i
n
e
e
rin
g
,
Ud
a
y
a
n
a
Un
i
v
e
rsity
,
I
n
d
o
n
e
sia
,
e
a
rn
e
d
in
2
0
2
4
.
He
r
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
t
h
e
d
i
g
it
a
l
e
c
o
n
o
m
y
,
d
a
ta
sc
ien
c
e
,
a
n
d
n
a
tu
ra
l
l
a
n
g
u
a
g
e
p
ro
c
e
ss
in
g
,
wit
h
a
fo
c
u
s
o
n
u
si
n
g
d
a
ta
-
d
riv
e
n
so
l
u
ti
o
n
s
t
o
e
n
h
a
n
c
e
e
c
o
n
o
m
ic
p
ro
c
e
ss
e
s
a
n
d
d
riv
e
i
n
n
o
v
a
ti
o
n
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
at
e
m
a
il
:
d
it
a
.
d
w
ik
a
sa
ri0
8
@s
tu
d
e
n
t
.
u
n
u
d
.
a
c
.
id
.
I
Ny
o
m
a
n
Pr
a
y
a
n
a
Tr
isn
a
re
c
e
iv
e
d
t
h
e
b
a
c
h
e
l
o
r’s
a
n
d
m
a
ste
r'
s
d
e
g
re
e
in
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
El
e
c
tro
n
ics
fro
m
th
e
U
n
iv
e
rsitas
G
a
d
jah
M
a
d
a
,
Yo
g
y
a
k
a
rta,
I
n
d
o
n
e
sia
,
in
2
0
1
7
a
n
d
2
0
2
0
re
sp
e
c
ti
v
e
ly
.
He
is
c
u
rre
n
tl
y
a
n
a
ss
istan
t
p
r
o
fe
ss
o
r
in
t
h
e
In
fo
rm
a
ti
o
n
Tec
h
n
o
l
o
g
y
S
tu
d
y
P
ro
g
ra
m
,
F
a
c
u
lt
y
o
f
E
n
g
i
n
e
e
rin
g
,
U
d
a
y
a
n
a
Un
i
v
e
rsity
,
Ba
li
,
I
n
d
o
n
e
sia
.
His
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
m
a
c
h
in
e
lea
rn
in
g
,
e
v
o
l
u
ti
o
n
a
ry
c
o
m
p
u
tatio
n
,
n
a
t
u
ra
l
lan
g
u
a
g
e
m
o
d
e
ll
in
g
,
a
n
d
tex
t
m
in
in
g
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
p
ra
y
a
n
a
.
tri
s
n
a
@u
n
u
d
.
a
c
.
i
d
.
I
Wa
y
a
n
A
g
u
s
S
u
r
y
a
Da
r
m
a
h
o
l
d
s
a
Do
c
t
o
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
d
e
g
re
e
fro
m
In
stit
u
t
Te
k
n
o
lo
g
i
S
e
p
u
lu
h
N
o
p
e
m
b
e
r,
In
d
o
n
e
sia
,
wit
h
th
e
Diss
e
rtatio
n
“
G
ra
p
h
fe
a
tu
re
f
u
sio
n
fo
r
t
h
e
re
c
o
g
n
it
i
o
n
o
f
h
ig
h
ly
v
a
ry
in
g
a
n
d
c
o
m
p
le
x
Ba
li
n
e
se
c
a
rv
in
g
m
o
ti
fs
”
.
He
is
c
u
rre
n
tl
y
a
n
a
ss
i
s
tan
t
p
ro
fe
ss
o
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
t
De
p
a
rtme
n
t
o
f
In
f
o
rm
a
ti
o
n
Tec
h
n
o
l
o
g
y
,
U
d
a
y
a
n
a
Un
iv
e
rsity
,
Ba
li
,
I
n
d
o
n
e
sia
.
He
h
a
s
p
u
b
l
ish
e
d
m
o
re
t
h
a
n
1
0
sc
o
p
u
s
in
d
e
x
e
d
jo
u
rn
a
l
a
rti
c
les
a
n
d
c
o
n
fe
re
n
c
e
p
a
p
e
rs.
His res
e
a
r
c
h
in
tere
sts a
re
in
a
rti
ficia
l
in
tell
i
g
e
n
c
e
,
c
o
m
p
u
ter
v
isio
n
,
a
n
d
c
o
m
p
u
tati
o
n
a
l
i
n
telli
g
e
n
c
e
.
He
c
a
n
b
e
c
o
n
tac
te
d
a
t
e
m
a
il
:
a
g
u
ss
u
ry
a
@u
n
u
d
.
a
c
.
id
.
Evaluation Warning : The document was created with Spire.PDF for Python.