I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
3
7
,
No
.
3
,
Ma
r
ch
20
2
5
,
p
p
.
1
7
5
8
~
1
77
1
I
SS
N:
2
502
-
4
7
52
,
DOI
:
1
0
.
1
1
5
9
1
/ijee
cs
.v
3
7
.
i
3
.
pp
1
7
5
8
-
1
77
1
1758
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
A nov
el hybrid m
o
del f
o
r sen
timen
t
a
na
ly
sis
in
M
O
O
C
forum
s
with
hy
bri
d word
and cha
ra
cter
-
le
v
el neural ne
twor
ks
M
o
ha
m
m
ed
J
ebba
ri
1
,
M
o
ha
m
ed
Am
ine O
ua
s
s
il
1
,
M
o
ua
a
d E
rr
a
m
i
1
,
Ra
bia
Ra
chidi
2
,
So
ufia
ne
H
a
m
ida
1,
3
,
5
,
B
o
ucha
ib Cher
ra
di
1,
2,
3
,4
,
Abdelh
a
di
Ra
iha
ni
1,
3
1
EEI
S
La
b
o
r
a
t
o
r
y
,
EN
S
ET
o
f
M
o
h
a
m
med
i
a
,
H
a
ss
a
n
I
I
U
n
i
v
e
r
si
t
y
o
f
C
a
s
a
b
l
a
n
c
a
,
M
o
h
a
mm
e
d
i
a
,
M
o
r
o
c
c
o
2
La
R
O
S
E
R
I
La
b
o
r
a
t
o
r
y
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
,
C
h
o
u
a
i
b
D
o
u
k
k
a
l
i
U
n
i
v
e
r
si
t
y
,
El
J
a
d
i
d
a
,
M
o
r
o
c
c
o
3
2
I
A
C
S
La
b
o
r
a
t
o
r
y
,
EN
S
ET
o
f
M
o
h
a
mm
e
d
i
a
,
H
a
ss
a
n
I
I
U
n
i
v
e
r
si
t
y
o
f
C
a
s
a
b
l
a
n
c
a
,
M
o
h
a
mm
e
d
i
a
,
M
o
r
o
c
c
o
4
S
TI
E
Te
a
m,
C
R
M
EF C
a
sa
b
l
a
n
c
a
-
S
e
t
t
a
t
,
p
r
o
v
i
n
c
i
a
l
s
e
c
t
i
o
n
o
f
El
J
a
d
i
d
a
,
E
l
Jad
i
d
a
,
M
o
r
o
c
c
o
5
G
EN
I
U
S
La
b
o
r
a
t
o
r
y
,
S
U
P
M
TI
o
f
R
a
b
a
t
,
R
a
b
a
t
,
M
o
r
o
c
c
o
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
1
,
2
0
2
4
R
ev
is
ed
Oct
6
,
2
0
2
4
Acc
ep
ted
Oct
9
,
2
0
2
4
S
e
n
ti
m
e
n
t
a
n
a
l
y
sis
is
c
ru
c
ial,
i
n
th
e
field
o
f
n
a
tu
ra
l
la
n
g
u
a
g
e
p
ro
c
e
ss
in
g
(NLP
).
Ha
s
a
p
p
li
c
a
ti
o
n
s
in
d
iffer
e
n
t
a
re
a
s.
Th
is
stu
d
y
fo
c
u
se
s
o
n
a
n
a
ly
z
in
g
se
n
ti
m
e
n
ts
in
m
a
ss
iv
e
o
p
e
n
o
n
l
i
n
e
c
o
u
rse
(M
OO
C)
f
o
ru
m
s
h
i
g
h
l
ig
h
ti
n
g
it
s
imp
o
rtan
c
e
i
n
u
n
d
e
rsta
n
d
i
n
g
h
o
w
u
se
rs
in
tera
c
t
a
n
d
sh
a
p
i
n
g
e
d
u
c
a
ti
o
n
a
l
stra
teg
ies
.
Th
e
st
u
d
y
p
re
se
n
ts
a
n
o
v
e
l
h
y
b
ri
d
n
e
u
ra
l
n
e
two
rk
m
o
d
e
l
sp
e
c
ifi
c
a
ll
y
tailo
re
d
f
o
r
se
n
ti
m
e
n
t
a
n
a
ly
sis
in
M
OO
C
fo
r
u
m
s.
Th
is
in
n
o
v
a
ti
v
e
m
o
d
e
l
c
o
m
b
i
n
e
s
wo
r
d
lev
e
l
a
n
d
c
h
a
ra
c
ter
lev
e
l
e
m
b
e
d
d
in
g
s
t
o
h
a
n
d
le t
h
e
li
n
g
u
isti
c
e
x
p
re
ss
io
n
s
c
o
m
m
o
n
ly
f
o
u
n
d
in
t
h
is co
n
tex
t.
Th
e
m
o
d
e
l
a
rc
h
it
e
c
tu
re
in
teg
ra
tes
b
id
irec
ti
o
n
a
l
l
o
n
g
sh
o
rt
-
term
m
e
m
o
ry
(BiL
S
TM
)
lay
e
rs
fo
r
w
o
rd
lev
e
l
e
m
b
e
d
d
in
g
s
a
n
d
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
two
r
k
s
(CNN
s)
fo
r
c
h
a
ra
c
ter
lev
e
l
e
m
b
e
d
d
in
g
s
a
imin
g
t
o
h
a
rn
e
ss
th
e
st
re
n
g
th
s
o
f
b
o
t
h
t
y
p
e
s
o
f
e
m
b
e
d
d
i
n
g
s
fo
r
a
v
iew
o
f
th
e
li
n
g
u
ist
ic
u
se
d
i
n
M
OO
C
fo
ru
m
p
o
sts.
No
tab
l
y
th
is
m
o
d
e
l
a
c
h
iev
e
s
a
n
a
c
c
u
ra
c
y
ra
te
o
f
9
3
.
1
1
%
s
h
o
wc
a
sin
g
it
s
e
ffe
c
ti
v
e
n
e
ss
,
in
se
n
ti
m
e
n
t
a
n
a
ly
sis
with
i
n
M
OO
C
fo
r
u
m
s.
T
h
i
s
re
se
a
rc
h
c
o
n
tri
b
u
tes
t
o
se
n
ti
m
e
n
t
a
n
a
ly
sis
with
in
th
e
c
o
n
tex
t
o
f
o
n
li
n
e
e
d
u
c
a
ti
o
n
.
K
ey
w
o
r
d
s
:
B
id
ir
ec
ti
o
n
al
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
C
h
ar
ac
ter
-
lev
el
em
b
ed
d
in
g
s
C
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
MO
OC
f
o
r
u
m
s
Neu
r
al
n
etwo
r
k
m
o
d
el
Sen
tim
en
t a
n
aly
s
is
W
o
r
d
-
lev
el
em
b
e
d
d
in
g
s
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
:
B
o
u
ch
aib
C
h
er
r
ad
i
E
E
I
S L
ab
o
r
ato
r
y
,
E
NSET
o
f
Mo
h
am
m
ed
ia,
Hass
an
I
I
Un
iv
er
s
ity
o
f
C
asab
lan
ca
Mo
h
am
m
ed
ia,
M
o
r
o
cc
o
E
m
ail:
b
o
u
ch
aib
.
ch
er
r
ad
i@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
I
n
r
ec
e
n
t
y
ea
r
s
,
s
en
tim
en
t
an
aly
s
is
h
as
em
er
g
ed
as
a
cr
itic
al
task
with
in
NL
P,
f
i
n
d
in
g
a
p
p
licatio
n
s
ac
r
o
s
s
d
iv
er
s
e
f
ield
s
s
u
ch
as
s
tu
d
en
t
e
n
g
ag
e
m
en
t,
p
atien
t
f
e
ed
b
ac
k
an
aly
s
is
,
cu
s
to
m
er
f
ee
d
b
ac
k
an
aly
s
is
an
d
s
o
cial
m
ed
ia
m
o
n
it
o
r
in
g
.
T
h
e
ac
cu
r
ate
class
if
ica
tio
n
o
f
s
en
tim
en
ts
ex
p
r
ess
ed
in
tex
t
u
al
d
ata
is
ess
en
tial
f
o
r
co
m
p
r
eh
e
n
d
in
g
u
s
er
o
p
in
io
n
s
an
d
e
m
o
tio
n
s
[
1
]
-
[
3
]
.
Desp
ite
ex
is
tin
g
s
o
lu
tio
n
s
th
at
e
m
p
lo
y
v
ar
io
u
s
d
ee
p
lear
n
in
g
a
n
d
tr
a
d
itio
n
al
m
ac
h
i
n
e
lear
n
in
g
m
o
d
els,
ch
allen
g
e
s
r
em
ain
i
n
e
f
f
ec
tiv
ely
a
n
aly
zi
n
g
th
e
in
cr
ea
s
in
g
ly
co
m
p
lex
a
n
d
d
i
v
er
s
e
d
ata
g
e
n
er
ated
b
y
o
n
lin
e
lear
n
in
g
p
latf
o
r
m
s
.
T
h
e
ad
v
en
t
o
f
o
n
lin
e
lear
n
i
n
g
p
latf
o
r
m
s
h
as
s
ig
n
if
ican
tly
en
r
ich
ed
th
e
a
v
ailab
le
tex
tu
al
d
ata,
p
ar
ticu
lar
ly
in
th
e
f
o
r
m
o
f
f
o
r
u
m
d
is
cu
s
s
io
n
s
,
th
er
eb
y
p
r
esen
tin
g
an
o
p
p
o
r
tu
n
ity
f
o
r
s
en
ti
m
en
t
an
aly
s
is
[
4
]
to
p
lay
a
p
iv
o
tal
r
o
le
in
u
n
d
er
s
tan
d
in
g
s
tu
d
en
t
e
n
g
ag
e
m
en
t,
s
atis
f
ac
tio
n
,
an
d
p
o
ten
tial
is
s
u
es
with
in
o
n
lin
e
lear
n
in
g
en
v
ir
o
n
m
e
n
ts
[
5
]
.
Ho
wev
er
,
c
u
r
r
en
t
s
en
tim
en
t
a
n
al
y
s
is
m
o
d
els
o
f
te
n
s
tr
u
g
g
le
t
o
ac
cu
r
ately
in
te
r
p
r
et
th
e
n
u
an
ce
d
an
d
c
o
n
tex
t
-
s
p
ec
if
ic
lan
g
u
ag
e
u
s
ed
in
th
e
s
e
f
o
r
u
m
s
,
lead
in
g
to
s
u
b
o
p
tim
al
class
i
f
icat
io
n
p
er
f
o
r
m
an
ce
.
T
h
e
u
n
d
e
r
s
tan
d
i
n
g
an
d
an
al
y
s
is
o
f
th
ese
s
en
ti
m
en
ts
ca
r
r
y
im
p
licatio
n
s
f
o
r
e
d
u
ca
tio
n
al
s
tr
ateg
ies
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l h
y
b
r
id
mo
d
el
fo
r
s
en
ti
men
t a
n
a
lysi
s
in
MOOC
fo
r
u
ms w
ith
h
yb
r
id
w
o
r
d
…
(
Mo
h
a
mme
d
Je
b
b
a
r
i
)
1759
an
d
f
ac
ilit
atin
g
d
ec
is
io
n
-
m
ak
i
n
g
in
th
e
o
f
o
n
lin
e
ed
u
ca
tio
n
[
6
]
.
E
v
alu
ate
s
tu
d
en
t
s
en
tim
en
t
s
in
o
n
lin
e
co
u
r
s
es,
id
en
tify
in
g
a
r
ea
s
wh
er
e
s
tu
d
en
ts
m
ig
h
t n
ee
d
ad
d
itio
n
al
s
u
p
p
o
r
t o
r
wh
er
e
co
u
r
s
e
c
o
n
ten
t c
an
b
e
im
p
r
o
v
ed
[
7
]
.
T
o
ad
d
r
ess
th
ese
is
s
u
es,
th
is
r
esear
ch
in
tr
o
d
u
ce
s
a
n
o
v
el
h
y
b
r
id
d
ee
p
lear
n
in
g
m
o
d
el
t
ailo
r
ed
to
ex
p
lo
it
d
iv
er
s
e
d
ata
r
ep
r
esen
t
atio
n
s
an
d
u
tili
ze
v
ar
io
u
s
d
ee
p
lear
n
in
g
m
o
d
els
f
o
r
an
aly
zi
n
g
a
d
ataset
d
er
iv
e
d
f
r
o
m
f
o
r
u
m
d
is
cu
s
s
io
n
s
o
n
th
e
Stan
f
o
r
d
m
ass
iv
e
o
p
e
n
o
n
li
n
e
co
u
r
s
e
(
MO
OC
)
p
latf
o
r
m
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
aim
s
to
en
h
an
ce
class
if
icatio
n
p
er
f
o
r
m
an
ce
b
y
in
te
g
r
atin
g
wo
r
d
-
le
v
el
a
n
d
ch
a
r
ac
ter
-
lev
el
f
ea
tu
r
es,
th
er
eb
y
im
p
r
o
v
in
g
th
e
ac
cu
r
ac
y
,
p
r
ec
i
s
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
o
f
s
en
tim
en
t
class
if
icatio
n
.
Ad
d
iti
o
n
ally
,
th
e
m
o
d
el
is
d
esig
n
ed
to
g
en
e
r
alize
ef
f
ec
ti
v
ely
to
u
n
s
ee
n
MO
OC
p
o
s
ts
,
en
s
u
r
in
g
r
o
b
u
s
t
p
er
f
o
r
m
an
ce
ac
r
o
s
s
d
iv
er
s
e
an
d
d
y
n
am
ic
te
x
t
in
p
u
ts
[
8
]
,
[
9
]
.
Fu
r
th
er
m
o
r
e,
t
h
e
r
es
ea
r
ch
ai
m
s
to
p
r
o
v
id
e
ac
tio
n
ab
le
in
s
ig
h
ts
b
y
eq
u
ip
p
in
g
ed
u
ca
to
r
s
,
ad
m
in
is
tr
ato
r
s
,
an
d
p
latf
o
r
m
d
ev
elo
p
er
s
with
a
p
o
wer
f
u
l
to
o
l
f
o
r
u
n
d
er
s
tan
d
in
g
s
tu
d
en
t
s
en
tim
en
t
,
en
ab
lin
g
d
ata
-
d
r
iv
en
d
ec
is
io
n
s
to
en
h
an
ce
th
e
lea
r
n
in
g
ex
p
er
ien
ce
[
1
0
]
-
[
1
2
]
.
Nav
ig
atin
g
th
e
d
y
n
am
ics
o
f
ed
u
ca
tio
n
tech
n
o
l
o
g
y
an
d
em
p
lo
y
in
g
s
en
tim
en
t
a
n
aly
s
is
with
in
th
e
co
n
tex
t
o
f
MO
OC
s
s
t
an
d
s
as
a
p
o
wer
f
u
l
to
o
l
in
ad
v
an
cin
g
th
e
f
u
tu
r
e
o
f
ac
ce
s
s
ib
le
an
d
ef
f
ec
tiv
e
o
n
lin
e
lear
n
in
g
[
8
]
.
As
MO
OC
s
co
n
tin
u
e
to
d
r
aw
p
ar
ticip
an
ts
f
r
o
m
d
iv
er
s
e
cu
lt
u
r
al
an
d
lin
g
u
is
tic
b
ac
k
g
r
o
u
n
d
s
,
s
en
tim
en
t
an
aly
s
is
b
ec
o
m
es
an
in
d
is
p
en
s
ab
le
to
o
l
f
o
r
tr
a
n
s
ce
n
d
in
g
lan
g
u
a
g
e
b
ar
r
ier
s
an
d
u
n
d
er
s
tan
d
in
g
th
e
n
u
a
n
ce
s
o
f
cr
o
s
s
-
cu
ltu
r
al
co
m
m
u
n
icatio
n
with
in
th
ese
d
ig
ital sp
ac
es
[
1
3
]
,
[
1
4
]
.
Fu
r
th
er
m
o
r
e
,
th
e
d
y
n
am
ic
e
n
v
ir
o
n
m
e
n
t
o
f
o
n
lin
e
e
d
u
ca
tio
n
d
em
an
d
s
co
n
tin
u
o
u
s
ad
a
p
tatio
n
an
d
in
n
o
v
atio
n
.
Sen
tim
en
t
an
aly
s
is
in
MO
O
C
f
o
r
u
m
s
ca
n
s
er
v
e
as
an
ea
r
ly
-
war
n
in
g
s
y
s
tem
,
aler
tin
g
ed
u
ca
to
r
s
an
d
p
latf
o
r
m
d
e
v
elo
p
e
r
s
to
last
ed
d
ev
elo
p
m
en
ts
,
ch
a
n
g
in
g
p
r
ef
er
en
ce
s
,
an
d
p
o
ten
tial
is
s
u
es
th
at
m
ig
h
t
im
p
ac
t
th
e
o
v
er
al
l le
ar
n
in
g
ex
p
er
ien
c
e
[
1
5
]
.
T
h
is
p
r
o
ac
tiv
e
a
p
p
r
o
ac
h
allo
ws f
o
r
tim
ely
ad
ju
s
tm
en
ts
to
co
u
r
s
e
co
n
ten
t,
in
s
tr
u
ctio
n
al
s
tr
ateg
ies,
an
d
p
latf
o
r
m
f
ea
tu
r
es,
u
ltima
tely
lead
in
g
to
a
m
o
r
e
r
esp
o
n
s
iv
e
a
n
d
ad
ap
tiv
e
o
n
lin
e
lear
n
in
g
e
n
v
ir
o
n
m
en
t
[
1
6
]
.
T
h
e
p
r
im
a
r
y
b
en
ef
it
o
f
s
en
tim
en
t
an
aly
s
is
lies
in
its
ca
p
ac
i
ty
to
e
v
alu
ate
u
s
er
in
ter
ac
tio
n
s
with
th
e
s
y
s
tem
.
T
h
e
p
r
o
ce
s
s
es
in
Fig
u
r
e
1
i
n
v
o
lv
e
an
a
ly
zi
n
g
u
s
er
g
en
e
r
a
ted
co
n
ten
t
ac
r
o
s
s
v
ar
io
u
s
o
n
li
n
e
p
latf
o
r
m
s
.
Fig
u
r
e
1
.
E
s
s
en
tial step
s
in
an
aly
zin
g
s
en
tim
en
ts
o
n
o
n
lin
e
p
latf
o
r
m
s
T
h
r
o
u
g
h
o
u
t
th
is
s
tu
d
y
,
a
n
o
v
e
l
h
y
b
r
id
d
ee
p
lear
n
in
g
m
o
d
el
is
in
tr
o
d
u
ce
d
,
th
is
ap
p
r
o
ac
h
in
co
r
p
o
r
ates
two
d
is
tin
ct
s
tr
ateg
ies
f
o
r
tex
t
d
ata
r
e
p
r
esen
tatio
n
an
d
em
p
l
o
y
s
two
d
if
f
er
e
n
t
d
ee
p
lear
n
in
g
m
et
h
o
d
s
d
u
r
in
g
th
e
class
if
icatio
n
p
h
ase.
T
h
e
ex
p
er
im
en
tal
r
esu
l
ts
h
ig
h
lig
h
t
th
e
s
u
p
er
io
r
class
if
icatio
n
p
er
f
o
r
m
a
n
ce
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
T
h
e
m
ain
c
o
n
tr
ib
u
tio
n
s
o
f
th
e
s
tu
d
y
in
clu
d
e:
−
A
co
m
p
r
eh
e
n
s
iv
e
r
o
ad
m
a
p
is
p
r
esen
ted
f
o
r
e
m
b
ed
d
in
g
f
o
r
u
m
d
is
cu
s
s
io
n
d
atasets
f
r
o
m
th
e
Stan
f
o
r
d
MO
OC
p
latf
o
r
m
.
−
T
ex
t
class
if
icatio
n
b
en
e
f
its
f
r
o
m
th
e
u
tili
za
tio
n
o
f
d
iv
e
r
s
e
d
ee
p
lear
n
in
g
m
eth
o
d
s
with
d
is
tin
ct
ch
ar
ac
ter
is
tics
,
r
esu
ltin
g
in
en
h
an
ce
d
class
if
icatio
n
p
er
f
o
r
m
a
n
ce
.
−
Pro
p
o
s
in
g
a
n
ew
h
y
b
r
id
m
o
d
el
th
at
in
teg
r
ates
d
if
f
er
e
n
t
tex
t
r
e
p
r
esen
tatio
n
s
a
n
d
d
ee
p
lear
n
in
g
tech
n
iq
u
es,
ac
h
iev
i
n
g
h
ig
h
cla
s
s
if
icatio
n
s
u
cc
ess
b
y
ex
tr
ac
ti
n
g
s
u
p
er
io
r
f
ea
tu
r
es
an
d
em
p
l
o
y
in
g
d
iv
er
s
e
ap
p
r
o
ac
h
es f
r
o
m
th
e
d
ataset.
−
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
is
v
alid
ated
th
r
o
u
g
h
v
ar
io
u
s
ex
p
er
im
e
n
ts
,
co
m
p
a
r
in
g
f
av
o
r
ab
ly
with
ex
is
tin
g
m
eth
o
d
s
d
o
cu
m
e
n
ted
in
th
e
lite
r
atu
r
e.
T
h
e
s
u
b
s
eq
u
en
t
s
ec
tio
n
s
o
f
t
h
is
wo
r
k
ar
e
o
r
g
a
n
ized
to
p
r
o
v
id
e
a
s
y
s
tem
atic
ex
p
lo
r
ati
o
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el.
I
n
s
ec
tio
n
2
,
a
co
m
p
r
eh
en
s
iv
e
r
e
v
iew
o
f
r
elate
d
wo
r
k
s
in
s
en
tim
en
t
a
n
aly
s
is
liter
atu
r
e
is
p
r
esen
ted
.
Sectio
n
3
d
elv
es
in
to
th
e
d
etails
o
f
th
e
ar
ch
itectu
r
e
o
f
th
e
p
r
o
p
o
s
ed
h
y
b
r
id
n
eu
r
al
n
etwo
r
k
m
o
d
el
f
o
r
s
en
tim
en
t
a
n
aly
s
is
,
em
p
h
a
s
izin
g
th
e
in
teg
r
atio
n
o
f
w
o
r
d
-
lev
el
an
d
c
h
ar
ac
ter
-
le
v
el
em
b
ed
d
in
g
s
f
o
llo
win
g
th
is
,
s
ec
tio
n
4
o
u
tlin
es
th
e
p
r
o
ce
s
s
o
f
d
ata
co
llectio
n
an
d
p
r
e
-
p
r
o
ce
s
s
in
g
,
em
p
h
asizin
g
th
e
s
ig
n
if
ican
ce
o
f
th
e
d
iv
er
s
e
Stan
f
o
r
d
MO
OC
f
o
r
u
m
d
ataset
an
d
t
h
e
m
eth
o
d
s
e
m
p
lo
y
ed
to
p
r
ep
ar
e
th
e
d
ata
f
o
r
m
o
d
el
tr
ain
i
n
g
.
f
o
llo
wed
b
y
a
n
ex
p
lo
r
atio
n
o
f
wo
r
d
r
e
p
r
esen
tatio
n
tech
n
iq
u
es,
d
is
cu
s
s
in
g
v
ar
io
u
s
en
c
o
d
in
g
m
eth
o
d
s
an
d
ju
s
tify
in
g
o
u
r
s
elec
tio
n
f
o
r
th
e
p
r
o
p
o
s
ed
h
y
b
r
id
m
o
d
el.
T
h
e
s
ec
tio
n
th
en
p
r
o
v
id
es
a
d
etailed
ex
p
o
s
itio
n
o
f
th
e
d
ee
p
lear
n
in
g
m
eth
o
d
o
lo
g
ies
em
p
lo
y
ed
.
I
n
s
ec
tio
n
5
,
th
e
s
tu
d
y
'
s
r
esu
lts
an
d
co
r
r
esp
o
n
d
i
n
g
d
is
cu
s
s
io
n
s
ar
e
p
r
esen
ted
,
s
h
o
wca
s
in
g
th
e
m
o
d
el'
s
p
r
o
f
icien
cy
th
r
o
u
g
h
v
ar
io
u
s
m
etr
ic
s
an
d
co
m
p
ar
is
o
n
s
with
o
th
er
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
1
758
-
1
77
1
1760
class
if
icatio
n
alg
o
r
ith
m
s
.
Fin
ally
,
Sectio
n
6
co
n
clu
d
es
th
e
ar
ticle
b
y
s
u
m
m
ar
izin
g
k
ey
f
in
d
in
g
s
an
d
p
r
o
p
o
s
in
g
p
o
ten
tial a
v
en
u
es f
o
r
f
u
tu
r
e
r
e
s
ea
r
ch
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
I
n
Sen
tim
en
t
an
aly
s
is
h
as
b
ee
n
ex
ten
s
iv
ely
ex
p
lo
r
e
d
in
r
ec
en
t
liter
atu
r
e,
with
r
esear
ch
er
s
in
v
esti
g
atin
g
v
ar
io
u
s
ap
p
r
o
ac
h
es
to
ex
tr
ac
t
m
ea
n
in
g
f
u
l
in
s
ig
h
ts
f
r
o
m
tex
tu
al
d
ata.
I
n
th
e
co
n
tex
t
o
f
o
n
lin
e
f
o
r
u
m
s
an
d
ed
u
ca
tio
n
al
p
latf
o
r
m
s
,
a
m
u
ltit
u
d
e
o
f
s
tu
d
ies
h
a
v
e
d
elv
ed
in
t
o
s
en
tim
en
t
an
al
y
s
is
to
u
n
d
er
s
tan
d
u
s
er
s
en
tim
en
ts
an
d
ex
p
er
ien
ce
s
.
On
e
s
ig
n
if
ican
t
ar
ea
o
f
r
esear
ch
r
ev
o
lv
es
ar
o
u
n
d
t
h
e
a
p
p
licatio
n
o
f
n
eu
r
al
n
etwo
r
k
a
r
ch
itectu
r
es
f
o
r
s
en
t
im
en
t
an
aly
s
is
.
R
ec
en
t
s
tu
d
ies,
in
clu
d
in
g
o
n
e
f
o
cu
s
ed
o
n
m
u
ltimo
d
al
s
en
tim
en
t
an
aly
s
is
,
in
tr
o
d
u
ce
d
f
o
u
r
r
ec
u
r
r
en
t n
eu
r
al
n
e
two
r
k
s
(
R
NNs)
v
ar
ian
ts
(
GR
NN,
L
R
NN,
GL
R
NN,
an
d
UGRNN)
to
an
aly
ze
v
id
e
o
c
o
n
ten
t.
I
m
p
r
ess
iv
ely
,
th
e
s
tu
d
y
ac
h
ie
v
ed
8
0
.
8
5
%
ac
cu
r
ac
y
in
s
en
tim
en
t
class
if
icatio
n
f
r
o
m
tex
t a
lo
n
e
[
1
7
]
.
Vim
ali
an
d
Mu
r
u
g
a
n
[
1
8
]
th
e
au
th
o
r
s
p
r
esen
t
a
m
o
d
el
th
at
in
teg
r
ates
C
NN
with
B
iLST
M,
an
d
th
e
ex
p
er
im
en
tal
f
in
d
in
g
s
d
em
o
n
s
tr
ate
its
s
u
p
er
io
r
ity
o
v
er
ex
i
s
tin
g
m
eth
o
d
s
,
ac
h
iev
in
g
an
im
p
r
ess
iv
e
9
1
.
0
5
%
ac
cu
r
ac
y
.
Ad
d
itio
n
ally
,
Mr
h
ar
et
a
l.
[
1
9
]
,
th
e
ch
allen
g
e
o
f
s
en
tim
en
t a
n
aly
s
is
in
MO
O
C
s
f
o
r
u
m
s
is
ad
d
r
ess
ed
.
T
h
e
au
th
o
r
s
p
r
o
p
o
s
e
a
B
ay
esian
C
NN
-
L
STM
m
o
d
el,
co
m
b
i
n
in
g
C
NN
an
d
L
STM
to
im
p
r
o
v
e
th
e
an
aly
s
is
o
f
s
en
tim
en
ts
an
d
m
ea
n
in
g
s
in
f
o
r
u
m
p
o
s
ts
.
T
h
e
s
tu
d
y
in
d
icate
s
th
at
th
is
co
m
b
in
ed
m
o
d
el
o
u
tp
er
f
o
r
m
s
in
d
iv
id
u
al
L
STM
a
n
d
C
NN
m
o
d
els,
ac
h
iev
in
g
a
n
o
tab
le
9
1
%
ac
cu
r
ac
y
an
d
s
u
r
p
ass
in
g
o
th
er
d
ee
p
lear
n
in
g
ar
ch
itectu
r
es e
x
am
in
e
d
in
th
e
r
esear
ch
.
Ma
r
f
an
i
et
a
l.
[
2
0
]
th
e
au
t
h
o
r
s
ad
d
r
ess
es
th
e
n
ee
d
f
o
r
ev
al
u
atin
g
lear
n
e
r
f
ee
d
b
ac
k
o
n
M
OOCs
b
y
cr
ea
tin
g
a
s
en
tim
en
t
an
aly
s
is
s
y
s
tem
.
I
t
co
m
p
ar
es
lex
ico
n
-
b
ased
(
VADE
R
)
an
d
tr
an
s
f
o
r
m
er
-
b
ased
(
B
E
R
T
)
m
o
d
els
to
aid
tr
ai
n
er
s
in
p
in
p
o
in
tin
g
c
o
u
r
s
e
im
p
r
o
v
e
m
en
t
s
.
Fin
d
in
g
s
s
h
o
w
B
E
R
T
's
ex
ce
llen
t
p
er
f
o
r
m
an
ce
,
r
ea
ch
in
g
8
4
% a
cc
u
r
ac
y
.
Z
y
o
u
t
a
n
d
Z
y
o
ut
[
2
1
]
,
r
esear
c
h
er
s
in
tr
o
d
u
ce
d
a
s
en
tim
en
t
a
n
aly
s
is
s
y
s
tem
u
tili
zin
g
th
r
ee
em
b
ed
d
in
g
tech
n
iq
u
es:
au
to
m
atic,
Glo
V
e,
an
d
B
E
R
T
,
with
an
atten
tio
n
lay
er
to
en
h
an
ce
p
e
r
f
o
r
m
an
ce
.
T
h
e
m
o
d
el
co
m
p
ar
is
o
n
s
r
ev
ea
led
th
at
th
e
B
E
R
T
-
b
ased
ap
p
r
o
ac
h
,
p
ar
ticu
lar
ly
with
B
i
-
L
STM
an
d
an
atten
tio
n
lay
er
,
ac
h
iev
ed
s
u
p
er
i
o
r
p
er
f
o
r
m
a
n
c
e
with
F
-
s
co
r
es
o
f
8
9
%
a
n
d
8
8
%,
s
ig
n
if
ican
tly
o
u
tp
e
r
f
o
r
m
in
g
o
th
er
tech
n
iq
u
es
ex
p
lo
r
ed
,
wh
ich
r
an
g
e
d
f
r
o
m
6
7
% to
6
9
%.
Mo
r
eo
v
er
,
th
e
ex
p
lo
r
atio
n
o
f
m
u
ltimo
d
al
s
en
tim
en
t
an
aly
s
is
h
as
b
ee
n
a
g
r
o
win
g
ar
ea
o
f
in
ter
est.
Stu
d
ies
co
m
b
in
in
g
in
f
o
r
m
at
io
n
f
r
o
m
d
if
f
er
e
n
t
m
o
d
alit
ies,
s
u
ch
as
wo
rd
-
lev
el
an
d
ch
ar
ac
ter
-
lev
el
em
b
ed
d
in
g
s
,
aim
to
a
d
d
r
ess
th
e
ch
allen
g
es
p
o
s
ed
b
y
d
iv
er
s
e
lin
g
u
is
tic
ex
p
r
ess
io
n
s
[
2
2
]
.
B
y
lev
er
ag
in
g
th
e
s
tr
en
g
th
s
o
f
m
u
ltip
le
m
o
d
alities
,
r
esear
ch
er
s
aim
to
c
r
ea
te
m
o
r
e
r
o
b
u
s
t
s
en
tim
en
t
a
n
aly
s
is
m
o
d
els
ca
p
ab
le
o
f
h
an
d
lin
g
n
u
a
n
ce
d
lan
g
u
ag
e
u
s
e
in
v
ar
io
u
s
co
n
tex
ts
.
3.
H
YB
RID
M
O
DE
L
ARCH
I
T
E
CT
URE F
O
R
S
E
N
T
I
M
E
N
T
ANA
L
YS
I
S
3
.
1
.
P
r
o
po
s
ed
m
o
del
I
n
th
is
s
tu
d
y
,
a
n
o
v
el
h
y
b
r
id
n
eu
r
al
n
etwo
r
k
m
o
d
el
is
p
r
esen
ted
,
tailo
r
ed
f
o
r
s
en
tim
en
t
a
n
aly
s
is
an
d
s
p
ec
if
ically
d
esig
n
ed
to
ca
p
t
u
r
e
n
u
an
ce
d
p
a
tter
n
s
with
in
MO
OC
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
T
h
e
m
o
d
el
in
te
g
r
ates
b
o
th
wo
r
d
-
lev
el
an
d
ch
a
r
ac
ter
-
lev
el
em
b
ed
d
in
g
s
,
r
ec
o
g
n
izi
n
g
th
e
im
p
o
r
ta
n
ce
o
f
h
a
n
d
lin
g
d
iv
er
s
e
lin
g
u
is
tic
ex
p
r
ess
io
n
s
p
r
e
v
alen
t
in
th
is
c
o
n
tex
t.
T
h
e
a
r
ch
itectu
r
e
b
eg
i
n
s
with
two
in
p
u
t
la
y
er
s
:
wo
r
d
in
p
u
t
a
n
d
ch
ar
ac
ter
in
p
u
t.
th
e
wo
r
d
in
p
u
t
lay
er
p
r
o
ce
s
s
es
wo
r
d
-
lev
el
e
m
b
ed
d
in
g
s
,
u
tili
zin
g
B
iLST
M
lay
er
s
.
T
h
is
allo
ws
th
e
m
o
d
el
to
ca
p
tu
r
e
s
eq
u
e
n
tial
d
ep
en
d
e
n
cies
an
d
co
n
tex
tu
al
in
f
o
r
m
atio
n
,
en
h
a
n
cin
g
its
u
n
d
er
s
tan
d
in
g
o
f
th
e
s
tr
u
ctu
r
e
an
d
m
ea
n
in
g
o
f
t
h
e
t
ex
t.
Simu
ltan
eo
u
s
ly
,
th
e
ch
ar
ac
ter
in
p
u
t
lay
er
h
an
d
les
ch
ar
ac
te
r
-
lev
el
em
b
ed
d
in
g
s
u
s
in
g
C
NNs.
T
h
ese
n
etwo
r
k
s
f
o
cu
s
o
n
m
o
r
p
h
o
lo
g
ical
an
d
s
tr
u
ctu
r
al
f
ea
tu
r
es,
p
r
o
v
id
i
n
g
a
m
o
r
e
g
r
a
n
u
lar
u
n
d
er
s
tan
d
in
g
o
f
t
h
e
lin
g
u
is
tic
n
u
an
ce
s
p
r
esen
t
i
n
t
h
e
tex
tu
al
d
ata.
T
h
e
f
ea
tu
r
es
e
x
tr
ac
ted
f
r
o
m
b
o
th
th
e
B
iLST
M
(
wo
r
d
-
le
v
el)
an
d
C
NN
(
ch
ar
ac
ter
-
lev
el)
lay
er
s
ar
e
th
en
co
n
ca
ten
ate
d
,
cr
ea
tin
g
a
m
er
g
ed
r
e
p
r
esen
tatio
n
o
f
th
e
in
f
o
r
m
atio
n
.
T
h
is
co
m
b
in
atio
n
aim
s
to
lev
er
a
g
e
th
e
co
m
p
lem
e
n
tar
y
s
tr
en
g
th
s
o
f
wo
r
d
an
d
ch
ar
ac
ter
-
lev
el
em
b
ed
d
i
n
g
s
,
p
r
o
v
id
i
n
g
a
m
o
r
e
c
o
m
p
r
e
h
en
s
iv
e
v
iew
o
f
t
h
e
lin
g
u
is
tic
co
n
te
n
t in
MO
OC
f
o
r
u
m
p
o
s
ts
.
A
co
n
n
ec
ted
lay
er
p
r
o
ce
s
s
es
th
e
co
n
ca
ten
ated
f
e
atu
r
es,
co
n
t
r
ib
u
tin
g
to
th
e
m
o
d
el'
s
ab
ilit
y
to
d
is
ce
r
n
s
en
tim
en
t
p
atter
n
s
.
T
h
e
f
in
al
lay
er
,
eq
u
ip
p
ed
with
a
s
ig
m
o
id
a
ctiv
atio
n
f
u
n
ctio
n
,
p
r
o
d
u
ce
s
b
in
ar
y
s
en
tim
en
t
p
r
ed
ictio
n
s
–
eith
er
p
o
s
itiv
e
o
r
n
eg
ativ
e.
T
h
is
ar
ch
itectu
r
e
is
en
v
is
io
n
ed
to
b
e
ad
ap
tab
l
e
an
d
ef
f
ec
tiv
e
in
h
an
d
lin
g
th
e
ch
allen
g
es p
o
s
ed
b
y
v
ar
ie
d
lan
g
u
ag
e
ex
p
r
ess
io
n
s
with
in
MO
OC
f
o
r
u
m
s
.
Fo
r
a
v
is
u
al
r
ep
r
esen
tatio
n
o
f
th
e
p
r
o
p
o
s
ed
h
y
b
r
i
d
m
o
d
el,
r
e
f
er
to
Fig
u
r
e
2
.
T
h
is
g
r
ap
h
ical
o
v
er
v
iew
illu
s
tr
ates
th
e
f
lo
w
o
f
in
f
o
r
m
atio
n
th
r
o
u
g
h
th
e
d
if
f
er
e
n
t
lay
er
s
,
em
p
h
asizin
g
th
e
in
teg
r
a
tio
n
o
f
wo
r
d
an
d
ch
ar
ac
ter
-
lev
el
em
b
ed
d
i
n
g
s
in
th
e
s
en
tim
en
t
an
aly
s
is
p
r
o
ce
s
s
.
T
h
e
m
o
d
e
l'
s
d
esig
n
aim
s
to
o
f
f
er
a
r
o
b
u
s
t
s
o
lu
tio
n
f
o
r
s
en
tim
en
t c
lass
if
icatio
n
in
MO
OC
s
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l h
y
b
r
id
mo
d
el
fo
r
s
en
ti
men
t a
n
a
lysi
s
in
MOOC
fo
r
u
ms w
ith
h
yb
r
id
w
o
r
d
…
(
Mo
h
a
mme
d
Je
b
b
a
r
i
)
1761
3
.
2
.
Sentim
ent
a
na
ly
s
is
m
o
d
el
a
rc
hite
ct
ure
T
h
e
h
y
b
r
id
n
e
u
r
al
n
etwo
r
k
ar
c
h
itectu
r
e
p
r
o
p
o
s
ed
in
th
is
r
ese
ar
ch
is
m
eticu
lo
u
s
ly
d
esig
n
e
d
to
ad
d
r
ess
th
e
ch
allen
g
es
ass
o
ciate
d
wit
h
s
en
tim
en
t
an
al
y
s
is
in
th
e
u
n
iq
u
e
c
o
n
tex
t
o
f
MO
OC
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
T
h
is
n
o
v
el
m
o
d
el
in
tr
o
d
u
ce
s
a
co
m
b
in
atio
n
o
f
wo
r
d
-
le
v
el
an
d
ch
ar
ac
ter
-
lev
el
em
b
e
d
d
in
g
s
,
ca
p
italizin
g
o
n
th
eir
co
m
p
lem
en
tar
y
s
tr
en
g
th
s
to
e
n
h
an
ce
th
e
s
y
s
tem
'
s
u
n
d
er
s
tan
d
in
g
o
f
th
e
in
tr
icate
lin
g
u
is
tic
n
u
an
ce
s
p
r
esen
t
in
th
e
tex
tu
al
d
ata.
T
h
ese
em
b
ed
d
in
g
s
u
n
d
e
r
g
o
B
iLST
M
lay
er
s
,
r
en
o
wn
ed
f
o
r
ca
p
tu
r
i
n
g
s
eq
u
en
tial
d
e
p
e
n
d
en
cies,
en
ab
lin
g
th
e
m
o
d
el
to
d
ev
elo
p
a
co
m
p
r
eh
e
n
s
iv
e
u
n
d
er
s
tan
d
in
g
o
f
co
n
tex
tu
al
in
f
o
r
m
atio
n
with
in
th
e
w
o
r
d
s
.
T
h
is
f
ac
ilit
ates
th
e
d
is
ce
r
n
m
e
n
t
o
f
n
u
a
n
ce
d
s
en
tim
en
t
p
atte
r
n
s
in
h
er
e
n
t
in
th
e
s
eq
u
en
tial
s
tr
u
ctu
r
e
o
f
MO
OC
f
o
r
u
m
p
o
s
ts
.
Simu
ltan
eo
u
s
ly
,
t
h
e
m
o
d
el
in
clu
d
es
a
C
h
ar
ac
ter
I
n
p
u
t
lay
er
th
at
p
r
o
ce
s
s
es
ch
ar
ac
ter
-
lev
el
em
b
ed
d
in
g
s
u
s
in
g
C
NNs.
L
ev
er
ag
in
g
th
e
s
tr
en
g
th
s
o
f
C
NNs,
th
is
lay
er
e
x
ce
ls
in
ca
p
tu
r
in
g
m
o
r
p
h
o
lo
g
ical
an
d
s
tr
u
ctu
r
al
f
ea
tu
r
es
with
in
th
e
ch
ar
ac
ter
s
o
f
th
e
te
x
t.
T
h
e
g
r
an
u
lar
a
n
aly
s
is
at
th
e
ch
ar
a
cter
lev
el
en
h
an
ce
s
th
e
m
o
d
el'
s
ca
p
ac
ity
to
r
ec
o
g
n
ize
s
u
b
tle
v
ar
iatio
n
s
in
lan
g
u
ag
e
ex
p
r
ess
io
n
s
,
co
n
tr
ib
u
tin
g
s
ig
n
if
ican
tly
to
o
v
er
all
s
en
tim
en
t a
n
aly
s
is
.
T
h
e
in
teg
r
atio
n
o
f
f
ea
tu
r
es
p
l
ay
s
a
p
i
v
o
tal
r
o
le
in
th
e
ar
ch
i
tectu
r
e.
Ou
tp
u
ts
f
r
o
m
b
o
th
t
h
e
B
iLST
M
(
wo
r
d
-
lev
el
)
an
d
C
NN
(
ch
ar
a
cter
-
lev
el)
lay
er
s
a
r
e
co
n
ca
ten
ated
,
cr
ea
tin
g
a
u
n
if
ied
r
ep
r
es
en
tatio
n
o
f
f
ea
tu
r
es
d
er
iv
ed
f
r
o
m
b
o
th
m
o
d
alities
.
T
h
is
in
teg
r
atio
n
aim
s
to
ca
p
tu
r
e
a
h
o
lis
tic
v
iew
o
f
lin
g
u
is
tic
co
n
ten
t
in
MO
O
C
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
B
y
f
u
s
in
g
f
ea
tu
r
es
f
r
o
m
wo
r
d
a
n
d
ch
ar
a
cter
-
lev
el
em
b
e
d
d
in
g
s
,
th
e
m
o
d
el
g
ain
s
th
e
ab
ilit
y
to
co
n
s
id
er
b
o
t
h
co
n
tex
tu
al
in
f
o
r
m
atio
n
an
d
m
o
r
p
h
o
lo
g
ic
al
in
tr
icac
ies,
co
n
tr
ib
u
tin
g
to
a
m
o
r
e
n
u
an
ce
d
s
en
tim
en
t a
n
aly
s
is
.
T
h
e
m
o
d
el
f
u
r
t
h
er
in
c
o
r
p
o
r
at
es
a
f
u
lly
co
n
n
ec
ted
lay
e
r
,
w
h
er
e
th
e
c
o
n
ca
ten
ated
f
ea
tu
r
e
s
u
n
d
er
g
o
p
r
o
ce
s
s
in
g
.
T
h
is
lay
e
r
in
tr
o
d
u
ce
s
a
lev
el
o
f
ab
s
tr
ac
tio
n
a
n
d
co
m
p
lex
ity
to
r
ef
in
e
f
ea
t
u
r
e
r
ep
r
esen
tatio
n
s
,
p
r
ep
ar
in
g
th
em
f
o
r
th
e
f
in
al
s
tag
es
o
f
s
en
tim
en
t
p
r
ed
ictio
n
.
th
e
f
u
lly
co
n
n
ec
ted
ar
c
h
itectu
r
e
en
s
u
r
es
th
at
t
h
e
m
o
d
el
ca
n
ex
tr
ac
t
h
ig
h
e
r
-
lev
el
p
atter
n
s
an
d
r
elatio
n
s
h
ip
s
f
r
o
m
th
e
co
m
b
in
e
d
wo
r
d
a
n
d
ch
ar
ac
ter
-
le
v
el
f
ea
tu
r
es,
en
h
a
n
cin
g
th
e
o
v
e
r
all
s
en
tim
en
t a
n
aly
s
is
ca
p
ab
ilit
ies o
f
th
e
m
o
d
el.
T
h
e
m
o
d
el
cu
lm
in
ates
in
an
o
u
tp
u
t
lay
e
r
an
d
p
r
ed
ictio
n
p
h
a
s
e,
wh
er
e
a
f
i
n
al
lay
e
r
eq
u
ip
p
ed
with
a
s
ig
m
o
id
ac
tiv
atio
n
f
u
n
ctio
n
p
r
o
d
u
ce
s
b
in
a
r
y
s
en
tim
en
t
p
r
e
d
ictio
n
s
eith
er
p
o
s
itiv
e
o
r
n
e
g
ativ
e.
th
is
d
ec
is
iv
e
lay
er
s
y
n
th
esizes
th
e
in
s
ig
h
ts
g
ain
ed
f
r
o
m
th
e
co
m
b
in
ed
f
e
atu
r
es,
m
ar
k
in
g
t
h
e
co
m
p
letio
n
o
f
th
e
s
en
tim
en
t
an
aly
s
is
p
r
o
ce
s
s
.
T
h
e
m
o
d
el'
s
p
r
o
f
icien
cy
in
p
r
o
v
id
in
g
ac
c
u
r
ate
s
en
tim
en
t
class
if
icatio
n
s
f
o
r
MO
OC
f
o
r
u
m
p
o
s
ts
is
attr
ib
u
ted
to
th
is
f
in
al
lay
er
,
wh
ic
h
en
ca
p
s
u
lates
th
e
co
llectiv
e
u
n
d
er
s
tan
d
in
g
d
er
iv
e
d
f
r
o
m
t
h
e
in
teg
r
ated
wo
r
d
an
d
ch
ar
ac
ter
-
lev
el
f
ea
tu
r
es.
T
h
is
co
n
clu
d
in
g
p
h
ase
u
n
d
er
s
co
r
es
th
e
s
o
p
h
is
ticatio
n
an
d
a
d
ap
tab
ilit
y
o
f
t
h
e
h
y
b
r
id
ar
c
h
i
tectu
r
e
in
ad
d
r
ess
in
g
th
e
in
tr
icac
ies o
f
lan
g
u
ag
e
u
s
e
with
in
o
n
lin
e
e
d
u
c
atio
n
al
f
o
r
u
m
s
.
T
h
e
in
te
g
r
atio
n
o
f
b
o
th
wo
r
d
an
d
ch
ar
ac
ter
-
lev
el
em
b
e
d
d
in
g
s
,
f
o
llo
wed
b
y
a
r
ef
in
ed
p
r
ed
ictio
n
m
ec
h
an
is
m
,
co
n
tr
ib
u
tes
to
t
h
e
m
o
d
el'
s
ef
f
icac
y
in
ca
p
tu
r
in
g
n
u
an
ce
d
s
en
tim
en
t
p
atter
n
s
.
T
h
e
h
y
b
r
id
ar
ch
it
ec
tu
r
e'
s
ab
ilit
y
to
d
is
till
m
ea
n
in
g
f
u
l
in
s
ig
h
ts
f
r
o
m
d
iv
er
s
e
lin
g
u
is
tic
ex
p
r
ess
io
n
s
p
o
s
itio
n
s
it
as
a
r
o
b
u
s
t
s
o
lu
tio
n
f
o
r
s
en
tim
en
t
an
aly
s
is
,
tailo
r
ed
to
t
h
e
u
n
i
q
u
e
ch
ar
ac
ter
is
tics
o
f
MO
OC
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
Fig
u
r
e
2
.
Ov
e
r
v
iew
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
1
758
-
1
77
1
1762
4.
M
E
T
H
O
DS
T
h
is
s
ec
tio
n
d
etails
th
e
m
eth
o
d
s
u
s
ed
in
o
u
r
s
tu
d
y
co
llectio
n
an
d
p
r
e
-
p
r
o
ce
s
s
in
g
o
f
a
d
a
taset.
T
h
is
s
ec
tio
n
o
u
tlin
es
th
e
m
eth
o
d
s
u
s
ed
in
th
is
s
tu
d
y
,
b
eg
in
n
i
n
g
with
th
e
d
ata
co
llectio
n
an
d
p
r
e
-
p
r
o
ce
s
s
in
g
o
f
MO
OC
f
o
r
u
m
p
o
s
ts
.
W
e
em
p
lo
y
ed
v
ar
i
o
u
s
wo
r
d
r
ep
r
ese
n
tatio
n
tech
n
iq
u
es
a
n
d
d
ee
p
lear
n
in
g
m
eth
o
d
s
,
in
clu
d
in
g
L
STM
,
B
iLST
M,
an
d
C
NNs.
Fin
ally
,
th
e
m
o
d
el
was
tr
ain
ed
an
d
e
v
alu
ated
u
s
in
g
ap
p
r
o
p
r
iat
e
m
etr
ics to
en
s
u
r
e
ac
cu
r
ate
s
en
tim
en
t a
n
aly
s
is
.
4
.
1
.
Da
t
a
s
et
s
T
h
e
d
ataset
u
tili
ze
d
in
th
is
s
t
u
d
y
,
a
cr
u
cial
co
m
p
o
n
en
t
f
o
r
d
ev
elo
p
in
g
an
d
tr
ai
n
in
g
th
e
s
en
tim
en
t
an
aly
s
is
m
o
d
el,
was
g
en
er
o
u
s
ly
p
r
o
v
id
ed
b
y
An
d
r
ea
s
Paep
ck
e
(
p
ae
p
c
k
e@
cs.st
an
f
o
r
d
.
ed
u
)
f
r
o
m
th
e
Stan
f
o
r
d
MO
OC
f
o
r
u
m
.
I
t
co
m
p
r
is
es
a
co
llectio
n
o
f
p
o
s
ts
ex
tr
ac
ted
f
r
o
m
th
e
Stan
f
o
r
d
MO
OC
f
o
r
u
m
(
2
4
6
8
4
an
o
n
y
m
ized
f
o
r
u
m
p
o
s
ts
)
.
E
a
ch
p
o
s
t
in
th
e
d
ataset
is
ass
o
ciate
d
with
s
en
tim
en
t
lab
els,
wh
ich
ar
e
cr
u
cial
f
o
r
s
u
p
er
v
is
in
g
th
e
tr
ai
n
in
g
o
f
th
e
m
o
d
el.
T
h
e
d
ataset
is
d
iv
er
s
e,
en
co
m
p
ass
in
g
a
r
an
g
e
o
f
to
p
ics
a
n
d
d
is
cu
s
s
io
n
s
r
ef
lectiv
e
o
f
th
e
d
y
n
am
ic
n
atu
r
e
o
f
MO
OC
f
o
r
u
m
s
.
I
t
in
clu
d
es
b
o
t
h
p
o
s
itiv
e
a
n
d
n
e
g
ativ
e
s
en
tim
en
ts
,
p
r
o
v
id
in
g
a
b
alan
ce
d
r
ep
r
esen
tatio
n
o
f
t
h
e
p
a
r
ticip
a
n
ts
'
o
p
in
io
n
s
an
d
ex
p
r
ess
io
n
s
.
T
h
e
v
ar
iety
o
f
lin
g
u
is
tic
s
ty
les
an
d
ex
p
r
ess
io
n
s
with
in
th
e
d
ataset
is
es
s
en
tial
f
o
r
test
in
g
th
e
m
o
d
el'
s
ab
ilit
y
to
h
an
d
le
d
iv
er
s
e
lan
g
u
ag
e
p
atter
n
s
co
m
m
o
n
ly
f
o
u
n
d
in
MO
OC
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
I
n
Fig
u
r
e
3
,
wo
r
d
clo
u
d
v
is
u
a
lizatio
n
s
ar
e
p
r
esen
ted
f
o
r
b
o
t
h
p
o
s
itiv
e
an
d
n
eg
ativ
e
s
en
ti
m
en
ts
.
T
h
e
p
o
s
itiv
e
s
en
tim
en
t
wo
r
d
clo
u
d
(
a)
illu
s
tr
ates
th
e
m
o
s
t
f
r
e
q
u
en
tly
o
cc
u
r
r
in
g
w
o
r
d
s
in
p
o
s
itiv
ely
class
if
ied
p
o
s
ts
,
o
f
f
er
i
n
g
i
n
s
ig
h
ts
in
to
p
r
ev
alen
t
th
em
es.
C
o
n
v
er
s
ely
,
th
e
n
e
g
ativ
e
s
en
tim
en
t
wo
r
d
cl
o
u
d
(
b
)
s
h
o
wca
s
es
wo
r
d
s
co
m
m
o
n
ly
f
o
u
n
d
in
n
eg
ativ
ely
class
if
ied
p
o
s
ts
,
p
r
o
v
id
in
g
a
s
n
a
p
s
h
o
t
o
f
lan
g
u
a
g
e
ass
o
ciate
d
with
n
eg
ativ
ity
.
T
h
ese
v
is
u
aliza
tio
n
s
o
f
f
e
r
a
s
u
cc
in
ct
y
et
in
f
o
r
m
ativ
e
o
v
e
r
v
iew
o
f
s
en
tim
e
n
t
p
atter
n
s
,
h
ig
h
li
g
h
tin
g
k
ey
wo
r
d
s
th
at
co
n
tr
ib
u
te
s
ig
n
if
ican
tly
to
s
en
tim
en
t
class
if
ic
atio
n
.
T
h
e
F
ig
u
r
e
3
(
a)
s
h
o
ws
th
e
wo
r
d
clo
u
d
f
o
r
p
o
s
itiv
e
s
en
tim
en
ts
,
an
d
th
e
F
ig
u
r
e
3
(
b
)
s
h
o
ws
th
e
wo
r
d
cl
o
u
d
f
o
r
n
e
g
ativ
e
s
en
tim
en
ts
.
T
h
ese
v
is
u
aliza
tio
n
s
h
elp
u
s
s
ee
th
e
m
o
s
t c
o
m
m
o
n
wo
r
d
s
ass
o
ciate
d
with
p
o
s
itiv
e
a
n
d
n
e
g
ativ
e
f
ee
lin
g
s
in
th
e
f
o
r
u
m
d
is
cu
s
s
io
n
s
.
(
a)
(
b
)
Fig
u
r
e
3
.
Sen
tim
en
t a
n
aly
s
is
wo
r
d
clo
u
d
s
p
o
s
itiv
e
(
a)
a
n
d
n
eg
ativ
e
an
d
(
b
)
s
en
tim
en
ts
4
.
2
.
Da
t
a
prepro
ce
s
s
ing
Data
p
r
ep
r
o
ce
s
s
in
g
is
a
cr
u
cia
l
s
tep
in
ex
tr
ac
tin
g
v
alu
ab
le
in
f
o
r
m
atio
n
f
r
o
m
tex
t
u
al
d
ata,
p
ar
ticu
lar
ly
in
th
e
co
n
tex
t
o
f
s
en
tim
en
t
a
n
aly
s
is
.
I
n
th
is
s
tu
d
y
f
o
cu
s
ed
o
n
s
en
tim
en
t
an
aly
s
is
with
in
MO
OC
f
o
r
u
m
s
,
ef
f
ec
tiv
e
d
ata
p
r
ep
r
o
ce
s
s
in
g
p
lay
s
a
v
ital
r
o
le
in
u
n
d
er
s
tan
d
in
g
u
s
er
in
ter
ac
tio
n
s
an
d
im
p
r
o
v
in
g
e
d
u
ca
tio
n
al
s
tr
ateg
ies.
T
h
e
p
r
ep
r
o
ce
s
s
in
g
p
h
ase
in
v
o
lv
es
co
n
v
er
tin
g
r
aw
tex
t
d
ata
in
to
a
f
o
r
m
at
s
u
itab
le
f
o
r
m
o
d
el
tr
ain
in
g
.
A
s
er
ies
o
f
ess
en
tial
s
tep
s
in
d
ata
clea
n
in
g
,
in
clu
d
i
n
g
tex
t
n
o
r
m
aliza
tio
n
,
is
th
e
i
n
itial
p
h
ase,
aim
in
g
to
s
tan
d
ar
d
ize
tex
t
b
y
co
n
v
e
r
tin
g
it
to
a
co
m
m
o
n
f
o
r
m
,
in
clu
d
in
g
lo
wer
ca
s
in
g
an
d
h
an
d
lin
g
v
ar
iatio
n
s
.
s
to
p
wo
r
d
r
em
o
v
al
co
m
es
n
ex
t
,
elim
in
atin
g
co
m
m
o
n
,
n
o
n
-
in
f
o
r
m
ativ
e
wo
r
d
s
to
f
o
cu
s
o
n
m
ea
n
in
g
f
u
l
co
n
ten
t.
L
em
m
atiza
tio
n
f
o
ll
o
ws,
r
ed
u
c
in
g
wo
r
d
s
to
th
eir
b
ase
f
o
r
m
f
o
r
c
o
n
s
is
ten
cy
.
s
p
ec
ial
ch
a
r
a
cter
h
an
d
li
n
g
th
e
n
d
ea
ls
with
n
o
n
-
alp
h
a
n
u
m
er
ic
ch
ar
ac
ter
s
,
r
ef
i
n
in
g
th
e
d
ata
a
n
d
r
e
d
u
cin
g
n
o
is
e.
T
h
e
s
tep
s
in
Fig
u
r
e
4
cr
ea
te
a
s
o
lid
f
o
u
n
d
atio
n
f
o
r
m
ea
n
i
n
g
f
u
l
an
aly
s
is
o
f
tex
tu
al
d
ata.
Af
ter
th
ese
s
tep
s
th
e
tex
t
s
eq
u
en
ce
s
ar
e
to
k
en
ized
u
s
in
g
b
o
th
wo
r
d
-
lev
el
an
d
ch
ar
ac
ter
-
lev
el
t
o
k
en
izer
s
,
a
n
d
s
u
b
s
eq
u
en
t
p
ad
d
in
g
e
n
s
u
r
es
u
n
if
o
r
m
in
p
u
t
d
im
en
s
io
n
s
f
o
r
b
o
th
m
o
d
alitie
s
.
4
.
3
.
Wo
rd
re
presenta
t
io
n
T
h
e
s
t
u
d
y
i
n
[
2
3
]
f
o
u
n
d
t
h
a
t
e
m
p
l
o
y
i
n
g
w
o
r
d
e
m
b
e
d
d
i
n
g
i
n
c
o
n
j
u
n
c
t
i
o
n
w
i
t
h
t
h
e
L
S
T
M
a
l
g
o
r
i
t
h
m
p
r
o
v
i
d
e
a
n
a
c
c
u
r
a
c
y
r
a
t
e
o
f
8
7
.
6
%
.
I
n
t
h
e
i
n
v
e
s
ti
g
a
t
i
o
n
,
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
t
h
r
e
e
w
id
e
l
y
u
t
i
l
i
ze
d
w
o
r
d
e
m
b
e
d
d
i
n
g
a
l
g
o
r
i
t
h
m
s
w
as
e
v
al
u
a
t
e
d
,
n
a
m
e
l
y
W
o
r
d
2
V
e
c
[
3
]
,
G
l
o
b
a
l
V
e
ct
o
r
s
(
Gl
o
V
e
)
[
2
3
]
,
an
d
F
a
s
t
T
e
x
t
[
2
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l h
y
b
r
id
mo
d
el
fo
r
s
en
ti
men
t a
n
a
lysi
s
in
MOOC
fo
r
u
ms w
ith
h
yb
r
id
w
o
r
d
…
(
Mo
h
a
mme
d
Je
b
b
a
r
i
)
1763
Fig
u
r
e
4
.
Data
p
r
ep
r
o
ce
s
s
in
g
s
tep
s
4
.
3
.
1
.
Wo
rd2
Vec
,
G
lo
Ve
a
nd
F
a
s
t
T
ex
t
W
o
r
d
2
Vec
p
r
o
d
u
ce
s
wo
r
d
v
e
cto
r
s
th
r
o
u
g
h
a
co
m
b
i
n
atio
n
o
f
s
u
p
er
v
is
ed
an
d
u
n
s
u
p
e
r
v
is
ed
lear
n
in
g
tech
n
iq
u
es,
r
e
n
o
wn
ed
f
o
r
its
ca
p
ac
ity
to
ca
p
t
u
r
e
th
e
s
em
an
tics
o
f
wo
r
d
s
a
n
d
th
eir
s
y
n
tactic
co
n
n
ec
tio
n
s
.
T
h
e
tr
ain
in
g
m
eth
o
d
s
em
p
lo
y
ed
b
y
W
o
r
d
2
Vec
in
clu
d
e
co
n
tin
u
o
u
s
b
ag
-
of
-
wo
r
d
s
(
C
B
O
W
)
,
wh
er
e
th
e
m
o
d
el
en
d
ea
v
o
r
s
to
p
r
ed
ict
a
ta
r
g
et
wo
r
d
b
ased
o
n
a
p
r
o
v
i
d
ed
c
o
n
tex
t,
an
d
Sk
ip
Gr
am
,
wh
er
e
th
e
m
o
d
el
aim
s
t
o
p
r
ed
ict
th
e
co
n
tex
t w
o
r
d
g
iv
e
n
a
tar
g
et
wo
r
d
.
Glo
Ve
u
tili
ze
s
wo
r
d
co
-
o
cc
u
r
r
en
ce
f
o
r
th
e
g
en
e
r
atio
n
o
f
w
o
r
d
v
ec
to
r
s
.
B
y
in
te
g
r
atin
g
th
e
s
tr
en
g
th
s
o
f
co
-
o
cc
u
r
r
en
ce
m
atr
ices
an
d
m
atr
ix
f
ac
to
r
izatio
n
,
Glo
Ve
a
ch
iev
es
m
o
r
e
p
r
ec
is
e
wo
r
d
v
e
cto
r
s
[
2
4
]
.
Fas
tTe
x
t
ad
o
p
ts
a
d
is
tin
ctiv
e
s
tr
ateg
y
b
y
s
eg
m
en
tin
g
ea
c
h
wo
r
d
in
t
o
s
u
b
wo
r
d
s
,
r
ef
er
r
ed
to
as
n
-
g
r
am
s
,
an
d
u
tili
ze
s
th
ese
s
u
b
wo
r
d
s
to
f
o
r
m
wo
r
d
v
ec
to
r
s
.
T
h
is
m
eth
o
d
is
esp
ec
ially
b
en
ef
icial
f
o
r
lan
g
u
ag
es f
ea
tu
r
in
g
co
m
p
o
u
n
d
wo
r
d
s
an
d
f
le
x
ib
le
s
u
f
f
ix
es.
Fi
g
u
r
e
5
il
lu
s
tr
ates th
e
f
u
n
d
am
e
n
tal
p
r
o
ce
s
s
o
f
wo
r
d
em
b
e
d
d
i
n
g
with
in
th
e
m
o
d
el
ar
ch
itectu
r
e,
wh
ich
in
co
r
p
o
r
at
es
a
d
ed
icate
d
W
o
r
d
I
n
p
u
t
lay
er
f
o
r
p
r
o
ce
s
s
in
g
wo
r
d
-
le
v
el
em
b
ed
d
i
n
g
s
u
s
in
g
W
o
r
d
2
Vec
,
Glo
v
e
a
n
d
Fas
tTe
x
t e
m
b
ed
d
in
g
s
.
Fig
u
r
e
5
.
T
h
e
f
u
n
d
am
e
n
tal
p
r
o
ce
s
s
o
f
wo
r
d
em
b
e
d
d
in
g
4
.
3
.
2
.
Wo
rd
em
bedd
ing
s
a
nd
cha
ra
ct
er
e
m
bedd
ing
s
T
h
e
m
o
d
el
in
co
r
p
o
r
ates
a
Ke
r
as
em
b
ed
d
in
g
lay
e
r
f
o
r
wo
r
d
em
b
ed
d
in
g
s
,
e
n
ab
lin
g
th
e
l
ea
r
n
in
g
o
f
co
n
tin
u
o
u
s
v
ec
to
r
r
ep
r
esen
tatio
n
s
f
o
r
ea
ch
w
o
r
d
in
th
e
d
ataset.
T
h
e
em
b
e
d
d
i
n
g
lay
er
is
co
n
f
ig
u
r
e
d
with
ess
en
tial
p
ar
am
eter
s
,
s
u
ch
as
in
p
u
t_
d
im
,
wh
ich
c
o
r
r
esp
o
n
d
s
to
th
e
v
o
ca
b
u
lar
y
s
ize
d
e
r
iv
ed
f
r
o
m
th
e
u
n
iq
u
e
wo
r
d
s
in
t
h
e
d
ataset,
a
n
d
o
u
tp
u
t_
d
im
,
d
ef
in
in
g
th
e
d
i
m
en
s
io
n
ality
o
f
th
e
d
en
s
e
wo
r
d
em
b
ed
d
i
n
g
s
.
Ad
d
itio
n
ally
,
t
h
e
in
p
u
t_
le
n
g
t
h
p
ar
a
m
eter
e
n
s
u
r
es
co
n
s
is
te
n
cy
in
th
e
in
p
u
t
s
eq
u
en
ce
le
n
g
th
,
f
ac
ilit
atin
g
th
e
m
o
d
el'
s
ca
p
ac
ity
to
ca
p
t
u
r
e
s
em
an
tic
r
elatio
n
s
h
ip
s
b
etwe
e
n
wo
r
d
s
a
n
d
d
er
iv
e
m
ea
n
in
g
f
u
l
r
ep
r
esen
tatio
n
s
f
r
o
m
th
e
tex
t d
ata.
Ad
d
itio
n
ally
,
th
e
m
o
d
el
in
co
r
p
o
r
ates
wo
r
d
em
b
ed
d
in
g
s
o
b
t
ain
ed
f
r
o
m
p
r
e
-
tr
ain
ed
m
o
d
els,
in
clu
d
in
g
W
o
r
d
2
Vec
,
Glo
Ve,
an
d
Fas
tTe
x
t.
T
h
e
W
o
r
d
2
Vec
m
o
d
el,
p
r
etr
ain
ed
u
s
in
g
th
e
Gen
s
im
lib
r
ar
y
.
T
h
is
m
o
d
el
in
itializes
th
e
wo
r
d
em
b
ed
d
i
n
g
lay
er
with
m
ea
n
i
n
g
f
u
l
weig
h
ts
,
cr
ea
tin
g
an
em
b
ed
d
in
g
m
atr
ix
wh
er
e
ea
ch
wo
r
d
'
s
em
b
ed
d
in
g
is
p
o
p
u
late
d
if
p
r
esen
t
in
th
e
W
o
r
d
2
Vec
m
o
d
el.
I
n
p
ar
allel,
th
e
m
o
d
e
l
lev
er
ag
es
Glo
Ve
an
d
Fas
tTe
x
t e
m
b
ed
d
in
g
s
to
f
u
r
th
er
e
n
r
ich
its
u
n
d
er
s
tan
d
in
g
o
f
wo
r
d
s
em
an
tics
.
I
n
p
a
r
allel,
ch
ar
ac
ter
em
b
e
d
d
i
n
g
s
f
o
c
u
s
o
n
r
ep
r
esen
tin
g
in
d
iv
id
u
al
ch
a
r
ac
ter
s
with
i
n
ea
ch
wo
r
d
.
An
em
b
ed
d
in
g
lay
e
r
f
o
r
c
h
ar
ac
t
er
s
is
em
p
lo
y
ed
,
f
o
llo
wed
b
y
th
e
a
p
p
licatio
n
o
f
a
1
D
C
o
n
v
o
lu
tio
n
al
lay
er
(
C
o
n
v
1
D)
to
th
e
ch
a
r
ac
ter
e
m
b
ed
d
in
g
s
.
T
h
is
co
n
v
o
lu
tio
n
al
lay
er
u
tili
ze
s
f
ilter
s
,
k
er
n
el
s
ize,
an
d
R
eL
U
ac
tiv
atio
n
f
u
n
ctio
n
s
to
d
is
ce
r
n
lo
ca
l
p
atter
n
s
an
d
f
ea
tu
r
es
wi
th
in
ch
ar
ac
ter
s
eq
u
en
ce
s
.
Su
b
s
eq
u
en
tly
,
a
Glo
b
al
Max
-
Po
o
lin
g
lay
e
r
(
Glo
b
alM
a
x
Po
o
lin
g
1
D)
is
u
s
ed
to
ex
tr
ac
t th
e
m
o
s
t c
r
u
cial
f
ea
tu
r
es
[
2
5
]
,
co
n
tr
ib
u
tin
g
to
an
en
h
an
ce
d
u
n
d
er
s
tan
d
in
g
o
f
c
h
ar
ac
ter
-
lev
el
p
atter
n
s
.
T
h
e
co
m
b
in
atio
n
o
f
wo
r
d
em
b
ed
d
in
g
s
,
ch
ar
ac
ter
em
b
e
d
d
in
g
s
,
an
d
th
e
in
teg
r
ati
o
n
o
f
p
r
e
-
tr
ain
ed
W
o
r
d
2
Vec
em
b
e
d
d
in
g
s
e
n
ab
l
es
th
e
m
o
d
el
t
o
ca
p
tu
r
e
a
c
o
m
p
r
eh
en
s
iv
e
s
et
o
f
f
ea
t
u
r
es
f
r
o
m
th
e
in
p
u
t
tex
t.
T
h
is
h
y
b
r
id
ap
p
r
o
ac
h
e
n
s
u
r
es
th
at
th
e
m
o
d
el
lear
n
s
n
o
t
o
n
l
y
s
em
an
tic
r
elatio
n
s
h
ip
s
b
etwe
en
wo
r
d
s
b
u
t
also
s
tr
u
ctu
r
al
an
d
ch
ar
ac
te
r
-
lev
el
p
atter
n
s
with
in
ea
ch
wo
r
d
.
T
h
e
in
f
o
r
m
atio
n
f
r
o
m
th
ese
em
b
ed
d
in
g
b
r
a
n
ch
es
is
co
n
ca
ten
ated
an
d
p
ass
ed
th
r
o
u
g
h
a
f
u
lly
co
n
n
ec
ted
lay
er
with
a
s
ig
m
o
id
ac
tiv
atio
n
f
u
n
ctio
n
f
o
r
th
e
f
in
al
b
in
ar
y
class
if
icatio
n
.
T
h
e
o
p
tim
al
ac
cu
r
ac
y
in
th
e
w
o
r
d
em
b
ed
d
in
g
a
p
p
r
o
ac
h
is
ac
h
iev
ed
b
y
u
s
in
g
s
p
ec
if
i
c
co
n
f
ig
u
r
atio
n
p
ar
am
eter
s
,
as d
etailed
in
T
ab
le
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
1
758
-
1
77
1
1764
No
tab
ly
,
f
o
r
wo
r
d
em
b
e
d
d
i
n
g
,
t
h
e
k
e
y
p
ar
am
eter
s
in
cl
u
d
e
a
n
em
b
ed
d
in
g
d
im
en
s
io
n
o
f
3
0
0
(
m
atch
in
g
th
e
W
o
r
d
2
Vec
m
o
d
el'
s
v
ec
to
r
s
ize)
,
a
v
o
ca
b
u
lar
y
s
ize
d
eter
m
in
ed
b
y
th
e
d
ataset,
an
o
u
tp
u
t
d
im
en
s
io
n
o
f
3
0
0
,
an
d
an
in
p
u
t
len
g
th
d
eter
m
i
n
ed
b
y
th
e
p
ad
d
ed
s
eq
u
en
c
e
len
g
th
.
On
th
e
o
th
er
h
an
d
,
f
o
r
ch
ar
ac
ter
-
lev
el
em
b
e
d
d
in
g
,
a
d
is
tin
ct
s
et
o
f
p
ar
am
eter
s
is
e
m
p
lo
y
ed
,
in
clu
d
in
g
6
4
f
ilter
s
an
d
a
k
er
n
el
s
ize
o
f
3
f
o
r
th
e
ch
ar
ac
ter
-
lev
el
C
NN,
1
2
8
L
STM
u
n
its
with
a
d
r
o
p
o
u
t
r
ate
o
f
0
.
2
,
1
2
8
f
ilter
s
an
d
a
k
er
n
el
s
ize
o
f
3
f
o
r
th
e
wo
r
d
-
le
v
el
C
NN,
an
d
a
c
o
n
ca
ten
ated
d
r
o
p
o
u
t
r
ate
o
f
0
.
5
.
T
h
is
d
is
tin
ctio
n
in
p
ar
am
et
er
s
u
n
d
er
s
co
r
es
th
e
tailo
r
ed
co
n
f
ig
u
r
atio
n
s
n
ec
ess
ar
y
to
o
p
tim
ize
ac
c
u
r
ac
y
in
s
en
tim
en
t
an
aly
s
is
f
o
r
b
o
th
wo
r
d
an
d
ch
ar
ac
ter
em
b
ed
d
in
g
s
.
T
ab
le
1
.
Mo
d
el
co
n
f
ig
u
r
atio
n
p
ar
am
eter
s
f
o
r
s
en
tim
en
t a
n
al
y
s
is
P
a
r
a
me
t
e
r
N
u
meri
c
V
a
l
u
e
E
m
b
e
d
d
i
n
g
d
i
me
n
s
i
o
n
(
w
o
r
d
)
3
0
0
(
W
o
rd
2
V
e
c
m
o
d
e
l
'
s
v
ect
o
r
s
i
z
e)
V
o
ca
b
u
l
ary
s
i
z
e
(
w
o
r
d
)
D
e
t
er
mi
n
e
d
b
y
t
h
e
d
a
t
a
s
e
t
O
u
t
p
u
t
d
i
me
n
s
i
o
n
(
w
o
rd
)
3
0
0
(
W
o
rd
2
V
e
c
m
o
d
e
l
'
s
v
ect
o
r
s
i
z
e)
In
p
u
t
l
e
n
g
t
h
(
w
o
rd
)
D
e
t
er
mi
n
e
d
b
y
t
h
e
p
a
d
d
e
d
s
e
q
u
e
n
ce
l
e
n
g
t
h
Fi
l
t
e
rs
(c
h
ar
C
N
N
)
64
K
er
n
e
l
s
i
ze
(c
h
ar
C
N
N
)
3
L
S
T
M
U
n
i
t
s
1
2
8
D
r
o
p
o
u
t
r
a
t
e
(
L
S
T
M)
0
.
2
Fi
l
t
e
rs
(
w
o
rd
C
N
N
)
1
2
8
K
er
n
e
l
s
i
ze
(
w
o
r
d
C
N
N
)
3
D
r
o
p
o
u
t
r
a
t
e
(c
o
n
ca
t
e
n
a
t
e
d
)
0
.
5
Param
et
e
r
N
u
mer
i
c
V
al
u
e
E
m
b
e
d
d
i
n
g
d
i
me
n
s
i
o
n
(
w
o
r
d
)
3
0
0
(W
o
r
d
2
V
ec
mo
d
el
'
s
v
ec
t
o
r
s
i
ze)
4
.
4
.
Dee
p lea
rning
m
e
t
ho
ds
4
.
4
.
1
.
L
o
ng
s
ho
rt
-
t
er
m
m
e
mo
ry
(
L
ST
M
)
L
STM
is
a
ty
p
e
o
f
r
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
ar
ch
itectu
r
e
d
esig
n
ed
to
ad
d
r
ess
th
e
v
an
is
h
in
g
g
r
ad
ien
t
p
r
o
b
lem
th
at
o
f
ten
o
cc
u
r
s
i
n
tr
ad
itio
n
al
R
NNs.
T
h
e
d
if
f
i
cu
lty
p
o
s
ed
b
y
th
e
v
an
is
h
in
g
g
r
a
d
ien
t
p
r
o
b
lem
h
in
d
er
s
th
e
ab
ilit
y
o
f
R
NNs
t
o
ef
f
ec
tiv
ely
ca
p
tu
r
e
a
n
d
r
em
em
b
er
lo
n
g
-
ter
m
d
ep
e
n
d
en
cie
s
in
s
eq
u
en
tial
d
ata
[
2
6
]
.
L
STM
was
in
tr
o
d
u
ce
d
t
o
o
v
er
co
m
e
th
is
is
s
u
e
b
y
in
c
o
r
p
o
r
ati
n
g
a
m
em
o
r
y
ce
ll
wit
h
a
m
o
r
e
co
m
p
lex
s
tr
u
ctu
r
e.
T
h
e
k
ey
c
o
m
p
o
n
en
t
s
o
f
an
L
STM
u
n
it in
Fig
u
r
e
6
in
clu
d
e:
−
C
ell
State
: C(t)
T
h
e
lo
n
g
-
ter
m
m
em
o
r
y
th
at
ca
n
s
to
r
e
i
n
f
o
r
m
a
tio
n
o
v
er
lo
n
g
s
eq
u
en
ce
s
.
−
Hid
d
en
State: h
(
t)
T
h
e
s
h
o
r
t
-
te
r
m
m
em
o
r
y
o
r
o
u
t
p
u
t o
f
th
e
L
STM
u
n
it.
−
I
n
p
u
t G
ate:
i(
t)
C
o
n
tr
o
ls
th
e
f
l
o
w
o
f
in
f
o
r
m
atio
n
in
to
th
e
ce
ll st
ate.
−
Fo
r
g
et
Gate
: f
(
t)
Dec
id
es wh
at
in
f
o
r
m
atio
n
f
r
o
m
th
e
ce
ll st
at
e
s
h
o
u
ld
b
e
d
is
ca
r
d
ed
o
r
f
o
r
g
o
tten
.
−
Ou
tp
u
t G
ate:
o
(
t)
Dete
r
m
in
es t
h
e
n
ex
t
h
id
d
en
s
tate
b
ased
o
n
t
h
e
cu
r
r
e
n
t in
p
u
t a
n
d
t
h
e
m
em
o
r
y
ce
ll.
Fig
u
r
e
6
.
T
h
e
s
tr
u
ctu
r
e
o
f
a
s
i
n
g
u
lar
L
STM
m
e
m
o
r
y
ce
ll
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l h
y
b
r
id
mo
d
el
fo
r
s
en
ti
men
t a
n
a
lysi
s
in
MOOC
fo
r
u
ms w
ith
h
yb
r
id
w
o
r
d
…
(
Mo
h
a
mme
d
Je
b
b
a
r
i
)
1765
4
.
4
.
2
.
B
idi
re
ct
io
na
l
lo
ng
s
ho
rt
-
t
er
m m
e
m
o
ry
(
B
iL
S
T
M
)
B
iLST
M
is
an
ex
ten
s
io
n
o
f
t
h
e
tr
ad
itio
n
al
L
STM
a
r
ch
itect
u
r
e,
e
n
h
an
ci
n
g
its
ab
ilit
y
to
c
ap
tu
r
e
a
n
d
u
n
d
er
s
tan
d
s
eq
u
en
tial
p
atter
n
s
in
d
ata
[
2
7
]
.
W
h
ile
s
tan
d
ar
d
L
STM
s
p
r
o
ce
s
s
in
p
u
t
s
eq
u
en
c
es
in
o
n
e
d
ir
ec
tio
n
(
f
r
o
m
t
h
e
b
e
g
in
n
in
g
to
th
e
e
n
d
)
,
B
iLST
M
p
r
o
ce
s
s
th
e
s
eq
u
en
ce
s
in
b
o
th
f
o
r
war
d
an
d
b
ac
k
war
d
d
ir
ec
tio
n
s
s
im
u
l
tan
eo
u
s
ly
.
T
h
is
b
id
ir
ec
ti
o
n
al
p
r
o
ce
s
s
in
g
allo
ws
th
e
n
et
wo
r
k
to
co
n
s
id
er
p
ast
an
d
f
u
t
u
r
e
co
n
tex
t
f
o
r
ea
ch
tim
e
s
tep
,
en
ab
lin
g
b
etter
m
o
d
elin
g
o
f
d
ep
e
n
d
en
cies
in
b
o
t
h
d
ir
ec
tio
n
s
,
an
d
en
h
an
ce
s
th
e
n
etwo
r
k
'
s
ca
p
ab
ilit
y
to
ca
p
tu
r
e
lo
n
g
-
ter
m
d
ep
e
n
d
en
cies
an
d
co
n
tex
t
in
s
eq
u
e
n
tial
d
ata
[
2
8
]
.
T
h
e
ar
ch
itec
tu
r
e
o
f
a
B
iLST
M
in
v
o
lv
es
two
s
ep
ar
ate
L
STM
lay
er
s
,
o
n
e
p
r
o
ce
s
s
in
g
th
e
s
eq
u
en
ce
f
o
r
war
d
a
n
d
th
e
o
th
er
b
ac
k
war
d
.
T
h
e
o
u
tp
u
ts
f
r
o
m
b
o
th
d
ir
ec
tio
n
s
ar
e
co
n
ca
ten
ate
d
b
ef
o
r
e
b
ei
n
g
p
ass
ed
to
s
u
b
s
eq
u
en
t
la
y
er
s
o
r
u
s
ed
f
o
r
m
a
k
in
g
p
r
ed
ictio
n
s
[
2
6
]
,
[
2
9
]
.
4
.
4
.
3
.
Co
nv
o
lutio
na
l
neura
l
n
et
wo
rk
(
CNN)
C
NNs
ar
e
a
class
o
f
d
ee
p
n
e
u
r
al
n
etwo
r
k
s
p
r
im
a
r
ily
d
esig
n
ed
f
o
r
p
r
o
ce
s
s
in
g
a
n
d
an
aly
zin
g
v
is
u
al
d
ata.
C
NNs
h
av
e
p
r
o
v
e
n
h
ig
h
ly
ef
f
ec
tiv
e
in
v
ar
io
u
s
co
m
p
u
ter
v
is
io
n
task
s
,
s
u
ch
as
im
ag
e
r
ec
o
g
n
itio
n
,
tex
t
class
if
icatio
n
,
o
b
ject
d
etec
tio
n
,
an
d
im
ag
e
class
if
icatio
n
[
3
0
]
,
[
3
1
]
.
T
h
e
ar
c
h
itectu
r
e
o
f
C
NNs
is
in
s
p
ir
ed
b
y
th
e
v
is
u
al
p
r
o
c
ess
in
g
in
th
e
h
u
m
an
b
r
ain
.
A
n
o
tab
le
s
tr
en
g
th
o
f
C
NNs
lies
in
th
eir
ca
p
ac
i
ty
to
au
to
m
atica
lly
id
en
tify
im
p
o
r
tan
t
f
ea
tu
r
es
with
o
u
t
m
an
u
al
in
ter
v
en
tio
n
,
allo
win
g
th
em
to
tack
le
p
r
o
b
lem
s
lack
in
g
s
p
atial
o
r
tem
p
o
r
al
d
e
p
en
d
e
n
cies
in
th
ei
r
f
ea
tu
r
es
[
2
6
]
.
A
s
tan
d
a
r
d
C
NN
s
tr
u
ctu
r
e
co
m
p
r
is
es
in
p
u
t
an
d
o
u
tp
u
t
lay
e
r
s
,
alo
n
g
with
h
i
d
d
en
la
y
er
s
aim
e
d
at
ex
tr
ac
tin
g
d
is
tin
ctiv
e
f
ea
t
u
r
es
[
3
0
]
.
4
.
4
.
4
.
Sentim
ent
a
na
ly
s
is
t
hro
ug
h B
iL
ST
M
a
nd
CNN
a
rc
hite
ct
ures
T
h
e
b
i
d
ir
ec
tio
n
al
ap
p
r
o
ac
h
B
iLST
M
en
ab
les
th
e
alg
o
r
ith
m
t
o
ca
p
tu
r
e
in
tr
icate
co
n
te
x
tu
al
in
f
o
r
m
atio
n
an
d
s
eq
u
e
n
tial
d
ep
en
d
e
n
cies
with
in
wo
r
d
-
l
ev
el
em
b
e
d
d
in
g
s
.
I
n
th
is
m
o
d
el,
B
iLST
M
is
s
tr
ateg
ically
ap
p
lied
to
th
e
W
o
r
d
I
n
p
u
t
lay
e
r
,
wh
er
e
it
p
r
o
ce
s
s
es
wo
r
d
-
lev
el
em
b
ed
d
i
n
g
s
[
3
2
]
.
W
ith
in
th
e
m
o
d
el,
C
NNs
ar
e
h
ar
n
ess
ed
t
o
p
r
o
ce
s
s
ch
ar
ac
ter
-
lev
el
em
b
ed
d
in
g
s
.
C
NNs
ex
ce
l
in
ca
p
tu
r
in
g
s
p
atial
p
atter
n
s
an
d
r
elatio
n
s
h
i
p
s
with
in
d
ata,
m
ak
in
g
th
em
well
-
s
u
ited
f
o
r
i
d
en
tify
in
g
m
o
r
p
h
o
l
o
g
ical
an
d
s
tr
u
ctu
r
al
f
ea
tu
r
es
with
in
th
e
ch
ar
ac
ter
s
o
f
th
e
tex
t.
T
h
is
g
r
an
u
lar
a
n
aly
s
is
a
t
th
e
c
h
ar
ac
ter
lev
el
s
ig
n
if
ican
tly
en
h
an
ce
s
th
e
m
o
d
el'
s
ab
ilit
y
to
r
ec
o
g
n
ize
s
u
b
tle
v
ar
iatio
n
s
in
lan
g
u
ag
e
e
x
p
r
ess
io
n
s
.
4
.
5
.
T
ra
ini
ng
a
nd
ev
a
lua
t
io
n
T
h
e
m
o
d
el
ex
p
er
ie
n
ce
s
tr
ain
in
g
u
tili
zin
g
a
b
in
ar
y
cr
o
s
s
-
en
tr
o
p
y
l
o
s
s
f
u
n
ctio
n
a
n
d
th
e
Ad
am
o
p
tim
izer
.
T
h
e
tr
ain
in
g
p
r
o
ce
s
s
i
s
ex
ec
u
ted
o
n
a
d
iv
id
ed
d
ataset,
allo
ca
tin
g
8
0
%
f
o
r
tr
ain
in
g
an
d
2
0
%
f
o
r
v
alid
atio
n
.
Dr
o
p
o
u
t
la
y
er
s
ar
e
in
clu
d
e
d
to
ad
d
r
ess
o
v
e
r
f
i
ttin
g
d
u
r
in
g
th
e
tr
ain
in
g
p
h
a
s
e.
I
n
ad
d
itio
n
to
moni
to
r
in
g
th
e
l
o
s
s
an
d
v
alid
atio
n
cu
r
v
es,
p
r
ec
is
io
n
a
n
d
F
1
s
co
r
e
m
etr
ics
ar
e
em
p
lo
y
ed
to
p
r
o
v
id
e
a
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
ev
alu
atio
n
o
f
t
h
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
.
T
h
e
tr
ain
in
g
p
h
ase
in
v
o
l
v
es
iter
ativ
ely
u
p
d
atin
g
th
e
m
o
d
el
p
ar
am
eter
s
to
m
in
im
ize
th
e
b
i
n
ar
y
c
r
o
s
s
-
en
t
r
o
p
y
lo
s
s
,
o
p
tim
izin
g
its
a
b
ilit
y
to
m
ak
e
ac
c
u
r
ate
s
en
ti
m
en
t
p
r
ed
ictio
n
s
.
T
h
e
in
cl
u
s
io
n
o
f
d
r
o
p
o
u
t
lay
er
s
s
er
v
es
as
a
r
eg
u
lar
izatio
n
tech
n
iq
u
e,
en
h
an
cin
g
th
e
m
o
d
el'
s
g
en
er
aliza
tio
n
ca
p
a
b
ilit
ies
b
y
p
r
ev
en
tin
g
it
f
r
o
m
r
ely
in
g
to
o
h
ea
v
ily
o
n
s
p
ec
if
ic
f
ea
tu
r
es
p
r
esen
t in
t
h
e
tr
ain
in
g
d
ata.
Fo
r
th
e
ass
es
s
m
en
t
o
f
th
e
m
o
d
el'
s
g
en
er
aliza
tio
n
an
d
p
r
e
d
ictiv
e
ca
p
ab
ilit
ies,
th
e
lear
n
in
g
cu
r
v
es
g
en
er
ated
a
n
d
th
e
s
tan
d
ar
d
a
cc
u
r
ac
y
m
etr
ic
(
1
)
,
p
r
ec
is
io
n
(
2
)
,
r
ec
all
(
3
)
a
n
d
F1
s
co
r
e
(
4
)
ar
e
em
p
lo
y
e
d
to
ev
alu
ate
th
e
m
o
d
el'
s
p
r
ec
is
io
n
in
co
r
r
ec
tly
class
if
y
in
g
p
o
s
itiv
e
s
en
tim
en
ts
an
d
its
o
v
er
al
l
ab
ilit
y
to
b
alan
ce
p
r
ec
is
io
n
an
d
r
ec
all.
A
cc
u
r
a
cy
=
+
T
P
+F
P
+F
N
+T
N
(
1
)
P
r
ec
is
io
n
=
TP
T
P
+F
P
(
2
)
R
ec
a
ll =
TP
T
P
+F
N
(
3
)
F1
-
s
co
r
e
= 2
*
P
r
e
c
is
ion*R
ac
all
P
r
e
c
is
ion+
R
e
c
all
(
4
)
Vis
u
aliza
tio
n
s
o
f
th
e
lo
s
s
an
d
v
alid
atio
n
cu
r
v
es
o
f
f
e
r
a
d
y
n
am
ic
r
e
p
r
esen
tatio
n
o
f
th
e
m
o
d
el's
lear
n
in
g
p
r
o
ce
s
s
.
T
h
ese
cu
r
v
e
s
aid
in
u
n
d
e
r
s
tan
d
in
g
h
o
w
w
ell
th
e
m
o
d
el
g
e
n
er
alize
s
to
n
ew,
u
n
s
ee
n
d
ata
a
n
d
h
ig
h
lig
h
t
p
o
ten
tial a
r
ea
s
f
o
r
i
m
p
r
o
v
e
m
en
t o
r
o
p
tim
izatio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
1
758
-
1
77
1
1766
5.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
I
n
th
is
s
ec
tio
n
,
a
c
o
m
p
r
e
h
en
s
iv
e
an
aly
s
is
o
f
th
e
p
er
f
o
r
m
an
c
e
an
d
lear
n
in
g
d
y
n
am
ics
o
f
th
e
s
en
tim
en
t
an
aly
s
is
m
o
d
el
tailo
r
ed
f
o
r
MO
OC
p
o
s
ts
is
p
r
o
v
id
e
d
.
T
h
e
ev
al
u
atio
n
is
ce
n
ter
ed
ar
o
u
n
d
k
e
y
m
et
r
ics,
in
clu
d
in
g
ac
c
u
r
ac
y
,
an
d
l
o
s
s
,
an
d
v
alid
atio
n
cu
r
v
es.
5
.
1
.
E
x
perim
ent
a
l
s
et
up
Sev
er
al
ex
p
er
im
en
ts
wer
e
ca
r
r
ied
o
u
t to
af
f
ir
m
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
wo
d
is
tin
ct
ex
p
er
im
en
tal
m
eth
o
d
o
l
o
g
ies
wer
e
em
p
lo
y
ed
t
o
ass
ess
th
e
p
er
f
o
r
m
an
ce
o
f
th
is
h
y
b
r
i
d
m
o
d
el.
T
h
e
i
n
itial
ap
p
r
o
ac
h
in
v
o
lv
ed
a
c
o
m
p
ar
is
o
n
o
f
t
h
e
b
asic
d
ee
p
lear
n
in
g
m
o
d
el,
m
ea
s
u
r
e
d
b
y
t
h
e
s
u
cc
e
s
s
o
f
class
if
icatio
n
ac
h
iev
ed
b
y
th
e
p
r
o
p
o
s
ed
m
o
d
el.
I
n
th
e
s
ec
o
n
d
a
p
p
r
o
ac
h
,
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
was
co
m
p
ar
ed
with
p
r
ev
io
u
s
s
tu
d
i
es
e
m
p
h
asizin
g
th
e
im
p
o
r
ta
n
c
e
o
f
d
ee
p
lear
n
in
g
m
o
d
els
in
tex
t
class
if
icatio
n
p
r
o
b
lem
s
,
al
o
n
g
with
tr
ad
itio
n
al
m
ac
h
in
e
lear
n
in
g
.
Fu
r
t
h
er
m
o
r
e,
t
h
e
r
esu
lts
o
f
ea
ch
m
o
d
el
wer
e
ev
al
u
ated
in
d
ep
en
d
en
tly
b
e
f
o
r
e
t
h
eir
co
m
b
in
atio
n
in
th
is
h
y
b
r
id
m
o
d
e
l.
T
h
e
ex
p
er
im
e
n
tal
s
etu
p
s
ar
e
ex
ec
u
ted
o
n
a
r
o
b
u
s
t
co
m
p
u
tin
g
in
f
r
astru
ctu
r
e
p
r
o
v
id
e
d
b
y
T
h
e
Natio
n
al
C
en
ter
f
o
r
Scien
tific
an
d
T
ec
h
n
ical
R
esear
ch
(
C
NR
ST)
.
T
h
is
in
f
r
astru
ctu
r
e,
s
p
e
cif
ically
tailo
r
ed
f
o
r
h
ig
h
-
p
e
r
f
o
r
m
an
ce
c
o
m
p
u
tin
g
(
HPC
)
,
b
o
asts
an
im
p
r
ess
iv
e
a
r
r
ay
o
f
r
eso
u
r
ce
s
,
in
cl
u
d
in
g
1
6
7
2
C
PU
co
r
es,
3
9
6
T
B
o
f
s
to
r
ag
e,
1
0
.
4
T
B
o
f
R
AM
,
an
d
GPU
ca
p
ab
ilit
ies
s
p
r
ea
d
ac
r
o
s
s
3
8
n
o
d
es.
T
h
e
m
o
d
els
ar
e
co
n
s
tr
u
cted
u
s
in
g
Py
th
o
n
3
.
7
an
d
v
ar
io
u
s
l
ib
r
ar
ies,
in
clu
d
in
g
T
en
s
o
r
Flo
w
an
d
Scik
it
-
lear
n
.
5.
2
.
T
esting
a
nd
t
ra
ini
ng
re
s
ults
T
h
e
le
ar
n
i
n
g
cu
r
v
es
o
f
th
e
m
o
d
el
in
Fig
u
r
e
7
d
ep
ict
tr
ain
in
g
lo
s
s
an
d
ac
cu
r
ac
y
cu
r
v
es.
T
h
e
tr
ain
in
g
lo
s
s
cu
r
v
e
d
em
o
n
s
tr
ates
a
co
n
tin
u
o
u
s
r
ed
u
ctio
n
f
r
o
m
1
.
1
2
3
3
to
0
.
2
1
4
5
.
T
h
is
d
escen
t
h
ig
h
lig
h
ts
th
e
m
o
d
el'
s
ef
f
ec
tiv
e
ad
a
p
tatio
n
to
th
e
c
o
m
p
lex
ities
o
f
th
e
tr
ain
in
g
d
at
a.
At
th
e
s
am
e
tim
e,
th
e
tr
ain
in
g
ac
cu
r
ac
y
c
u
r
v
e
d
em
o
n
s
tr
ates
a
co
n
s
is
ten
t
as
ce
n
t,
lev
elin
g
o
f
f
at
9
5
.
9
1
%.
T
h
is
s
tab
ilizat
io
n
s
u
g
g
ests
t
h
at
th
e
m
o
d
el
h
as
s
u
cc
ess
f
u
lly
r
ec
o
g
n
ized
a
n
d
le
ar
n
ed
in
tr
icate
p
atter
n
s
with
in
th
e
tr
ain
in
g
s
et,
co
n
tr
ib
u
ti
n
g
t
o
its
p
er
f
o
r
m
an
ce
.
T
h
e
v
alid
atio
n
cu
r
v
es,
also
illu
s
tr
ated
in
Fig
u
r
e
7
,
f
u
r
t
h
e
r
h
ig
h
lig
h
t
th
e
m
o
d
el'
s
r
o
b
u
s
tn
ess
an
d
g
en
er
aliza
tio
n
c
ap
ab
ilit
ies.
T
h
e
v
alid
atio
n
lo
s
s
cu
r
v
e
co
n
s
is
ten
tly
d
ec
r
ea
s
es
f
r
o
m
0
.
5
0
4
9
to
0
.
2
2
1
3
,
s
h
o
wca
s
in
g
th
e
m
o
d
el'
s
ap
titu
d
e
f
o
r
ad
ap
tin
g
to
p
r
ev
io
u
s
ly
u
n
s
ee
n
v
alid
atio
n
d
ata.
T
h
e
v
a
lid
atio
n
ac
cu
r
ac
y
,
a
cr
itical
m
etr
ic
f
o
r
ass
ess
in
g
th
e
m
o
d
el'
s
r
ea
l
-
wo
r
ld
p
er
f
o
r
m
an
ce
,
e
x
h
ib
its
im
p
r
o
v
em
en
t
ac
r
o
s
s
ep
o
ch
s
,
cu
lm
in
atin
g
in
a
p
ea
k
ac
c
u
r
ac
y
o
f
9
3
.
8
9
%.
T
h
is
s
tab
ilizatio
n
in
d
ica
tes
th
at
t
h
e
m
o
d
el
h
as
ad
ep
tly
id
en
tifie
d
an
d
ac
q
u
ir
ed
co
m
p
lex
p
atter
n
s
with
in
th
e
tr
ain
in
g
s
et,
en
h
a
n
cin
g
its
r
o
b
u
s
t
p
er
f
o
r
m
an
ce
.
Fig
u
r
e
8
illu
s
tr
ates
th
e
co
n
f
u
s
io
n
m
atr
ix
,
a
p
iv
o
ta
l
elem
en
t
d
ep
ictin
g
th
e
m
o
d
el'
s
p
r
o
f
icien
cy
in
d
is
ce
r
n
in
g
p
o
s
itiv
e
an
d
n
eg
ativ
e
s
en
tim
en
ts
.
No
tab
ly
,
th
e
m
o
d
el
ac
h
iev
ed
an
o
u
ts
tan
d
in
g
ac
cu
r
ac
y
o
f
9
3
.
1
1
%,
af
f
ir
m
in
g
its
ef
f
ec
tiv
en
ess
in
ca
p
tu
r
in
g
n
u
an
ce
d
s
en
tim
en
ts
with
in
o
n
lin
e
ed
u
ca
tio
n
al
d
is
c
u
s
s
io
n
s
.
Fig
u
r
e
7
.
T
r
ain
in
g
a
n
d
v
alid
atio
n
o
f
th
e
s
en
tim
en
t a
n
al
y
s
is
m
o
d
el
T
ab
le
2
o
f
f
er
s
a
co
m
p
ar
ativ
e
ev
alu
atio
n
o
f
th
e
tr
a
d
itio
n
al
class
if
icatio
n
alg
o
r
ith
m
s
,
in
clu
d
in
g
r
an
d
o
m
f
o
r
est
(
R
F),
d
ec
is
io
n
tr
ee
(
DT
)
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
,
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
,
n
eu
r
al
n
etwo
r
k
(
NN)
,
an
d
lo
g
is
tic
r
eg
r
ess
io
n
(
L
R
)
,
with
th
e
p
r
o
p
o
s
ed
m
o
d
el,
b
ased
o
n
v
ar
i
o
u
s
p
er
f
o
r
m
a
n
ce
m
etr
ics.
T
h
e
tab
le
p
r
esen
ts
Acc
u
r
ac
y
,
Pre
cisi
o
n
,
R
ec
all,
an
d
F
1
-
Sco
r
e
v
alu
es
f
o
r
ea
ch
m
o
d
el.
I
t
s
er
v
es
as
a
co
m
p
r
eh
e
n
s
iv
e
co
m
p
ar
is
o
n
to
o
l,
em
p
h
asizin
g
th
e
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
in
ter
m
s
o
f
ac
cu
r
ac
y
a
n
d
o
th
er
m
etr
ics.
T
h
ese
r
esu
lts
co
llectiv
ely
u
n
d
er
s
co
r
e
t
h
e
ef
f
icac
y
o
f
th
e
h
y
b
r
id
s
en
tim
en
t
an
aly
s
is
m
o
d
el
in
th
e
co
n
te
x
t
o
f
MO
OC
p
o
s
ts
.
T
h
e
im
p
r
ess
iv
e
ac
cu
r
ac
y
,
co
u
p
led
with
s
tab
le
lear
n
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2
5
0
2
-
4
7
52
A
n
o
ve
l h
y
b
r
id
mo
d
el
fo
r
s
en
ti
men
t a
n
a
lysi
s
in
MOOC
fo
r
u
ms w
ith
h
yb
r
id
w
o
r
d
…
(
Mo
h
a
mme
d
Je
b
b
a
r
i
)
1767
d
y
n
am
ics
an
d
v
alid
atio
n
p
e
r
f
o
r
m
an
ce
,
p
o
s
itio
n
s
th
e
m
o
d
e
l
as
a
v
alu
ab
le
t
o
o
l
f
o
r
d
ec
ip
h
er
in
g
s
en
tim
en
t
n
u
an
ce
s
in
th
e
o
n
lin
e
e
d
u
ca
tio
n
al
d
is
cu
s
s
io
n
s
.
T
ab
le
2
.
C
o
m
p
a
r
ativ
e
ev
alu
ati
o
n
o
f
m
ac
h
in
e
lea
r
n
in
g
alg
o
r
it
h
m
s
an
d
a
p
r
o
p
o
s
ed
m
o
d
el
M
o
d
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
-
S
c
o
r
e
NN
7
3
.
9
2
0
.
7
4
0
.
7
4
0
.
7
4
DT
7
3
.
0
1
0
.
7
4
0
.
7
3
0
.
7
4
RF
8
4
.
3
4
0
.
8
0
0
.
8
4
0
.
7
8
KNN
8
0
.
9
5
0
.
7
4
0
.
8
1
0
.
7
7
LR
8
1
.
2
7
0
.
8
4
0
.
9
5
0
.
9
0
T
h
e
p
r
o
p
o
s
e
d
m
o
d
el
9
3
.
1
1
0
.
8
3
0
.
9
7
0
.
8
4
Fig
u
r
e
8
.
Sen
tim
en
t a
n
aly
s
is
co
n
f
u
s
io
n
m
atr
ix
5.
3
.
Dis
cus
s
io
n
T
h
e
p
r
o
p
o
s
ed
h
y
b
r
id
m
o
d
el
co
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
s
b
o
th
i
n
d
iv
id
u
al
d
ee
p
lear
n
in
g
a
p
p
r
o
ac
h
es
an
d
tr
ad
itio
n
al
m
ac
h
in
e
lear
n
in
g
al
g
o
r
ith
m
s
.
T
h
is
s
u
g
g
ests
th
at
c
o
m
b
in
in
g
wo
r
d
-
lev
el
an
d
ch
ar
ac
ter
-
lev
el
f
ea
tu
r
es
p
r
o
v
id
es
a
m
o
r
e
c
o
m
p
r
e
h
en
s
iv
e
u
n
d
er
s
tan
d
i
n
g
o
f
s
en
tim
en
t
in
ed
u
ca
tio
n
al
co
n
tex
ts
.
T
h
e
s
tr
o
n
g
p
er
f
o
r
m
an
ce
o
n
v
alid
atio
n
d
ata
in
d
icate
s
th
at
o
u
r
m
o
d
el
ca
n
ef
f
ec
tiv
ely
g
en
er
alize
to
u
n
s
ee
n
MO
OC
p
o
s
ts
,
a
cr
u
cial
f
ac
to
r
f
o
r
r
ea
l
-
wo
r
ld
a
p
p
licatio
n
.
T
h
e
h
ig
h
ac
cu
r
ac
y
(
9
3
.
1
1
%)
i
n
d
is
tin
g
u
is
h
in
g
b
etwe
en
p
o
s
itiv
e
a
n
d
n
eg
ativ
e
s
en
tim
en
ts
d
em
o
n
s
tr
ates
o
u
r
m
o
d
el'
s
ab
ilit
y
to
ca
p
tu
r
e
s
u
b
t
le
em
o
tio
n
al
cu
es
i
n
e
d
u
ca
tio
n
al
d
is
cu
s
s
io
n
s
.
Ou
r
m
o
d
el'
s
p
er
f
o
r
m
an
ce
s
u
g
g
ests
it
co
u
ld
b
e
a
v
alu
ab
le
to
o
l
f
o
r
ed
u
ca
to
r
s
an
d
ad
m
in
is
tr
ato
r
s
in
u
n
d
er
s
tan
d
in
g
s
tu
d
en
t sen
tim
en
t a
t scale
,
p
o
t
en
tially
in
f
o
r
m
i
n
g
in
ter
v
en
tio
n
s
o
r
co
u
r
s
e
im
p
r
o
v
em
en
ts
.
Fig
u
r
e
9
p
r
o
v
i
d
es
a
co
m
p
r
e
h
en
s
iv
e
ev
alu
atio
n
o
f
th
e
p
er
f
o
r
m
a
n
ce
m
etr
ics
f
o
r
d
if
f
e
r
en
t
th
r
ee
s
en
tim
en
t
an
aly
s
is
m
o
d
els:
m
o
d
el
with
wo
r
d
-
lev
el,
m
o
d
el
with
ch
ar
ac
ter
-
lev
el,
an
d
th
e
h
y
b
r
id
m
o
d
el
.
E
ac
h
m
o
d
el
is
ass
ess
ed
b
ased
o
n
i
ts
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
Sco
r
e
,
ess
en
tial
m
etr
ics
in
g
a
u
g
in
g
th
e
ef
f
ec
tiv
en
ess
o
f
s
en
tim
en
t
an
a
ly
s
is
.
T
h
e
"M
o
d
el
with
W
o
r
d
-
L
ev
el"
ap
p
r
o
ac
h
ac
h
ie
v
es
a
s
o
lid
ac
cu
r
ac
y
with
b
alan
ce
d
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
Sco
r
e.
I
n
co
n
tr
ast,
th
e
m
o
d
el
with
ch
ar
ac
ter
-
lev
el
s
u
r
p
a
s
s
es
i
t
in
ac
cu
r
ac
y
,
d
em
o
n
s
tr
atin
g
n
o
tab
le
s
tr
en
g
t
h
s
in
p
r
ec
is
io
n
an
d
r
ec
all.
H
o
wev
er
,
Fig
u
r
e
9
d
em
o
n
s
tr
ates
th
at
th
e
p
r
o
p
o
s
ed
h
y
b
r
id
m
o
d
el,
s
h
o
wca
s
ed
u
n
d
er
th
e
p
r
o
p
o
s
ed
m
o
d
el
o
u
tp
er
f
o
r
m
s
b
o
th
with
th
e
h
ig
h
e
s
t
ac
cu
r
ac
y
an
d
r
ec
all,
in
d
icatin
g
its
s
u
p
er
io
r
a
b
ilit
y
to
ca
p
tu
r
e
n
u
an
ce
d
s
en
tim
en
t p
atter
n
s
in
MO
OC
f
o
r
u
m
p
o
s
ts
.
T
h
e
h
ig
h
ac
cu
r
ac
y
(
9
3
.
1
1
%)
in
d
is
tin
g
u
is
h
in
g
b
etwe
en
p
o
s
itiv
e
an
d
n
eg
ativ
e
s
en
tim
en
ts
d
em
o
n
s
tr
ates
o
u
r
m
o
d
el'
s
ab
ilit
y
to
ca
p
tu
r
e
s
u
b
tle
em
o
t
io
n
a
l
cu
es
in
ed
u
ca
tio
n
al
d
is
cu
s
s
io
n
s
.
T
h
is
n
u
an
ce
d
u
n
d
er
s
tan
d
i
n
g
co
u
ld
b
e
p
ar
ti
cu
lar
ly
v
alu
a
b
le
f
o
r
id
en
tif
y
i
n
g
s
tu
d
en
t
e
n
g
ag
e
m
en
t,
s
atis
f
ac
tio
n
,
o
r
p
o
ten
tial
is
s
u
es
in
o
n
lin
e
lear
n
in
g
en
v
ir
o
n
m
en
ts
.
T
h
e
m
o
d
el'
s
ab
ilit
y
to
ef
f
ec
tiv
ely
g
e
n
er
alize
to
u
n
s
ee
n
MO
OC
p
o
s
t
s
is
cr
u
cial
f
o
r
r
ea
l
-
wo
r
l
d
ap
p
l
icatio
n
s
,
wh
er
e
it
will
en
co
u
n
ter
n
ew
an
d
d
iv
er
s
e
tex
t
in
p
u
ts
.
T
h
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
s
u
g
g
ests
it
co
u
ld
b
e
a
v
alu
ab
le
to
o
l
f
o
r
ed
u
ca
t
o
r
s
an
d
ad
m
in
is
tr
ato
r
s
in
u
n
d
er
s
tan
d
in
g
s
tu
d
en
t
s
en
tim
en
t
at
s
ca
le.
B
y
an
aly
zi
n
g
lar
g
e
v
o
lu
m
es
o
f
f
o
r
u
m
p
o
s
ts
,
in
s
titu
tio
n
s
co
u
ld
g
ain
i
n
s
ig
h
ts
in
to
s
tu
d
e
n
t
ex
p
er
ien
ce
s
,
id
en
tify
ar
ea
s
o
f
co
n
ce
r
n
,
a
n
d
m
ak
e
d
ata
-
d
r
iv
e
n
d
ec
is
io
n
s
to
im
p
r
o
v
e
co
u
r
s
e
co
n
ten
t
o
r
lear
n
in
g
ex
p
er
ien
ce
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.