I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
24
,
p
p
.
6
5
7
2
~
6
5
8
2
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
14
i
6
.
pp
6
5
7
2
-
6
5
8
2
6572
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Enha
ncing
senti
ment ana
ly
sis
in
K
a
nna
da
t
e
x
ts
by
f
ea
tur
e
selection
Su
nil
M
ug
a
lih
a
lli
E
s
hwa
ra
pp
a
1,
3
,
Vina
y
Sh
iv
a
s
ub
ra
m
a
ny
a
n
2
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
P
ES
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
V
i
s
v
e
sv
a
r
a
y
a
Te
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
s
i
t
y
,
B
e
l
a
g
a
v
i
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
P
ES C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
V
i
s
v
e
s
v
a
r
a
y
a
T
e
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
si
t
y
,
B
e
l
a
g
a
v
i
,
I
n
d
i
a
3
S
o
f
t
w
a
r
e
E
n
g
i
n
e
e
r
,
W
i
p
r
o
L
i
mi
t
e
d
,
B
a
n
g
a
l
o
r
e
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ap
r
1
8
,
2
0
2
4
R
ev
is
ed
J
u
l 3
0
,
2
0
2
4
Acc
ep
ted
Au
g
6
,
2
0
2
4
In
re
c
e
n
t
y
e
a
rs,
th
e
re
h
a
s
b
e
e
n
a
n
o
ti
c
e
a
b
le
su
rg
e
in
re
se
a
rc
h
a
c
ti
v
it
ies
fo
c
u
se
d
o
n
se
n
t
ime
n
t
a
n
a
l
y
sis
with
in
th
e
Ka
n
n
a
d
a
lan
g
u
a
g
e
d
o
m
a
in
.
T
h
e
e
x
isti
n
g
re
se
a
rc
h
h
i
g
h
l
ig
h
ts
a
lac
k
o
f
lab
e
ll
e
d
d
a
tas
e
ts
a
n
d
li
m
it
e
d
e
x
p
lo
ra
ti
o
n
i
n
fe
a
tu
re
se
lec
ti
o
n
fo
r
Ka
n
n
a
d
a
se
n
ti
m
e
n
t
a
n
a
l
y
sis,
h
in
d
e
rin
g
a
c
c
u
ra
te
se
n
ti
m
e
n
t
c
las
sifica
ti
o
n
.
T
o
a
d
d
re
ss
t
h
is
g
a
p
,
t
h
e
st
u
d
y
a
ims
t
o
in
tro
d
u
c
e
a
n
o
v
e
l
Ka
n
n
a
d
a
d
a
ta
se
t
a
n
d
d
e
v
e
lo
p
a
n
e
ffe
c
ti
v
e
c
las
sifier
fo
r
imp
ro
v
e
d
se
n
ti
m
e
n
t
a
n
a
ly
sis
i
n
Ka
n
n
a
d
a
tex
ts.
Th
e
st
u
d
y
p
re
se
n
ts
a
n
e
w
Ka
n
n
a
d
a
d
a
tas
e
t
fr
o
m
S
e
m
Ev
a
l
2
0
1
4
Tas
k
4
u
si
n
g
G
o
o
g
le
Tran
sla
t
e
.
It
t
h
e
n
in
tro
d
u
c
e
s a
m
o
d
ifi
e
d
b
id
irec
ti
o
n
a
l
e
n
c
o
d
e
r
re
p
re
se
n
tatio
n
fr
o
m
tran
sfo
rm
e
rs
BERT
fo
r
Ka
n
n
a
d
a
d
a
tas
e
t
c
a
ll
e
d
a
s
Ka
n
n
a
d
a
-
BERT
(K
-
BERT
).
F
u
rth
e
r,
a
p
ro
b
a
b
il
it
y
-
c
lu
ste
ri
n
g
(P
C)
a
p
p
r
o
a
c
h
is
p
re
se
n
ted
to
e
x
trac
t
th
e
t
o
p
ics
a
n
d
it
s
re
late
d
a
sp
e
c
ts.
B
o
th
t
h
e
K
-
BER
T
c
las
sifier
a
n
d
P
C
a
p
p
ro
a
c
h
we
re
m
e
rg
e
d
to
a
tt
a
i
n
a
K
-
BERT
-
P
C
c
las
sifier,
in
teg
ra
ti
n
g
a
m
o
d
ifi
e
d
BER
T
m
o
d
e
l
a
n
d
p
ro
b
a
b
il
it
y
c
l
u
ste
rin
g
a
p
p
ro
a
c
h
fo
r
a
c
h
iev
i
n
g
b
e
tt
e
r
re
su
lt
s.
Ex
p
e
rime
n
tal
re
su
lt
s
d
e
m
o
n
stra
te
t
h
a
t
K
-
BE
RT
-
P
C
a
c
h
iev
e
s
su
p
e
rio
r
p
e
rfo
rm
a
n
c
e
in
p
o
larit
y
a
n
d
se
n
ti
m
e
n
t
a
n
a
ly
sis
a
c
c
u
ra
c
y
,
with
a
n
imp
re
ss
iv
e
a
c
c
u
ra
c
y
ra
te
o
f
9
1
%
,
s
u
rp
a
ss
in
g
e
x
isti
n
g
c
las
si
fiers
.
Th
is
wo
r
k
c
o
n
tri
b
u
tes
b
y
p
r
o
v
i
d
in
g
a
so
lu
ti
o
n
to
th
e
sc
a
rc
it
y
o
f
lab
e
ll
e
d
d
a
tas
e
ts
fo
r
Ka
n
n
a
d
a
se
n
ti
m
e
n
t
a
n
a
ly
sis
a
n
d
i
n
tro
d
u
c
e
s
a
n
e
ffe
c
ti
v
e
c
las
sifier,
K
-
BERT
-
P
C,
fo
r
e
n
h
a
n
c
e
d
se
n
ti
m
e
n
t
a
n
a
ly
sis o
u
tco
m
e
s in
Ka
n
n
a
d
a
tex
ts.
K
ey
w
o
r
d
s
:
B
E
R
T
Kan
n
ad
a
Pro
b
ab
ilit
y
clu
s
ter
in
g
Sem
E
v
al
2
0
1
4
Sen
tim
en
t a
n
aly
s
is
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
:
Su
n
il M
u
g
alih
alli E
s
h
war
ap
p
a
D
e
p
a
r
t
m
e
n
t
o
f
C
o
m
p
u
t
e
r
S
c
ie
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
P
E
S
C
o
ll
eg
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
V
i
s
v
es
v
a
r
ay
a
T
e
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
s
it
y
B
elag
av
i
-
5
9
0
0
1
8
,
I
n
d
ia
E
m
ail: su
n
i.m
g
h
alli@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
L
an
g
u
ag
e
d
iv
er
s
ity
is
a
f
u
n
d
am
en
tal
asp
ec
t
o
f
h
u
m
an
co
m
m
u
n
icatio
n
,
r
ef
lectin
g
th
e
r
ich
cu
ltu
r
al
d
iv
er
s
ities
th
at
d
ef
in
es
s
o
cieties
wo
r
ld
wid
e.
Fro
m
Asi
a
to
Af
r
ica,
E
u
r
o
p
e
to
th
e
Am
er
ic
as,
p
eo
p
le
s
p
ea
k
a
p
leth
o
r
a
o
f
la
n
g
u
a
g
es,
ea
ch
ca
r
r
y
in
g
its
u
n
iq
u
e
cu
ltu
r
es,
ex
p
r
ess
io
n
s
,
an
d
s
en
tim
en
ts
[
1
]
.
T
h
is
lin
g
u
is
tic
d
iv
er
s
ity
n
o
t
o
n
ly
en
r
ic
h
es
th
eir
g
lo
b
al
h
er
itag
e
b
u
t
also
p
r
esen
ts
in
tr
ig
u
in
g
ch
allen
g
es,
p
ar
ticu
lar
ly
in
th
e
r
ea
lm
o
f
s
en
tim
en
t
a
n
aly
s
is
.
I
n
a
c
o
u
n
tr
y
lik
e
I
n
d
ia,
k
n
o
wn
f
o
r
its
cu
ltu
r
al
m
o
s
aic,
lin
g
u
i
s
tic
d
iv
er
s
ity
tak
es
o
n
an
ev
e
n
m
o
r
e
p
r
o
n
o
u
n
ce
d
f
o
r
m
.
W
ith
2
8
s
tates
an
d
8
u
n
io
n
ter
r
ito
r
ies,
I
n
d
ia
h
as
a
s
tag
g
er
in
g
v
ar
iety
o
f
lan
g
u
ag
es,
with
ea
ch
s
tate
o
f
t
en
h
a
v
in
g
its
d
is
tin
ct
lin
g
u
is
tic
id
en
tity
.
Fro
m
Hin
d
i
in
t
h
e
n
o
r
th
to
T
am
il
in
th
e
s
o
u
th
,
f
r
o
m
B
en
g
ali
in
th
e
ea
s
t
to
Gu
jar
ati
in
th
e
west,
th
e
lin
g
u
is
tic
lan
d
s
ca
p
e
o
f
I
n
d
ia
i
s
a
test
am
en
t
to
its
v
ib
r
an
t
h
er
itag
e
an
d
r
e
g
io
n
al
d
iv
er
s
ity
.
I
n
to
d
ay
'
s
d
ig
ital
a
g
e,
wh
er
e
i
n
d
iv
id
u
als
ex
p
r
ess
th
eir
o
p
i
n
io
n
s
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
cin
g
s
en
timen
t a
n
a
lysi
s
in
K
a
n
n
a
d
a
texts b
y
fea
tu
r
e
s
elec
tio
n
(
S
u
n
il Mu
g
a
lih
a
lli E
s
h
w
a
r
a
p
p
a
)
6573
em
o
tio
n
s
o
n
a
m
y
r
iad
o
f
o
n
l
in
e
p
latf
o
r
m
s
,
s
en
tim
en
t
an
al
y
s
is
p
lay
s
a
cr
u
cial
r
o
le
[
2
]
.
Peo
p
le
s
h
ar
e
th
eir
s
en
tim
en
ts
o
n
e
-
co
m
m
er
ce
p
latf
o
r
m
s
ab
o
u
t
p
r
o
d
u
cts
th
ey
p
u
r
ch
ase
[
3
]
,
v
o
ice
th
eir
c
o
n
ce
r
n
s
o
n
s
o
cial
m
ed
ia
p
latf
o
r
m
s
r
eg
ar
d
i
n
g
p
o
liti
ca
l
i
s
s
u
es
[
4
]
,
a
n
d
en
g
a
g
e
i
n
d
is
cu
s
s
io
n
s
o
n
v
id
eo
en
ter
tain
m
e
n
t
s
ites
ab
o
u
t
v
ar
io
u
s
co
n
ten
t,
b
e
it
m
o
v
ies,
m
u
s
ic,
o
r
v
lo
g
s
[
5
]
.
T
h
ese
o
n
lin
e
i
n
ter
ac
tio
n
s
p
r
o
v
id
e
a
tr
ea
s
u
r
e
tr
o
v
e
o
f
d
ata
f
o
r
u
n
d
er
s
tan
d
i
n
g
p
u
b
lic
s
en
tim
en
ts
an
d
o
p
in
io
n
s
[
6
]
.
Ho
wev
er
,
a
s
ig
n
if
ican
t
c
h
allen
g
e
ar
is
es
wh
en
th
ese
s
en
tim
en
ts
ar
e
ex
p
r
ess
ed
in
n
ativ
e
lan
g
u
ag
es
[
7
]
.
Ma
n
y
s
en
tim
en
t
an
aly
s
is
class
if
ier
s
s
tr
u
g
g
le
to
co
m
p
r
eh
en
d
an
d
an
aly
ze
tex
t
wr
itten
in
la
n
g
u
ag
es
o
t
h
er
th
an
E
n
g
lis
h
,
wh
ich
ca
n
lim
it
th
eir
ef
f
ec
tiv
en
ess
in
ca
p
tu
r
in
g
t
h
e
tr
u
e
s
en
tim
en
t
o
f
th
e
u
s
er
s
[
8
]
.
T
h
is
lan
g
u
ag
e
b
ar
r
ier
p
o
s
es
an
is
s
u
e
in
ac
cu
r
ately
co
llectin
g
p
u
b
lic
s
en
tim
en
ts
,
esp
ec
ially
in
m
u
ltil
in
g
u
al
an
d
m
u
lticu
ltu
r
a
l
s
o
cieties l
ik
e
I
n
d
ia.
T
o
ad
d
r
es
s
th
is
ch
allen
g
e,
r
esear
ch
er
s
a
n
d
d
ata
s
cien
tis
ts
h
av
e
tu
r
n
ed
to
m
ac
h
in
e
-
lear
n
i
n
g
(
ML
)
[
9
]
,
[
1
0
]
an
d
d
ee
p
-
lear
n
in
g
(
DL
)
[
1
1
]
,
[
1
2
]
ap
p
r
o
ac
h
e
s
.
B
y
co
llectin
g
d
ata
f
r
o
m
n
e
wsp
ap
er
s
,
web
s
ites
,
b
lo
g
s
,
an
d
o
th
er
o
n
lin
e
s
o
u
r
c
es,
th
ey
tr
ain
m
o
d
els
to
class
i
f
y
s
en
tim
en
ts
ex
p
r
ess
ed
in
n
ativ
e
lan
g
u
ag
es
s
u
ch
as
Hin
d
i,
Ma
lay
alam
,
T
am
il,
an
d
Kan
n
a
d
a
(
p
ar
t
o
f
th
e
Dr
a
v
id
ian
lan
g
u
ag
e
f
am
ily
)
[
1
3
]
.
Ho
wev
er
,
d
esp
ite
th
ese
ef
f
o
r
ts
,
th
e
av
ailab
ilit
y
o
f
s
tan
d
ar
d
d
atasets
f
o
r
s
en
tim
en
t
an
aly
s
is
in
Kan
n
ad
a,
f
o
r
ex
am
p
le,
r
em
ain
s
lim
ited
[
1
4
]
.
Fu
r
th
er
m
o
r
e,
ex
is
tin
g
d
atasets
o
f
ten
lack
lab
elled
s
en
tim
en
t
p
o
lar
ity
,
wh
i
ch
is
ess
en
tial
f
o
r
tr
ain
in
g
s
en
tim
en
t
a
n
aly
s
is
m
o
d
els.
L
ab
ellin
g
d
atasets
with
p
o
s
itiv
e
an
d
n
eg
ativ
e
s
en
tim
en
t
r
eq
u
ir
es
s
ig
n
if
ican
t
tim
e
a
n
d
ef
f
o
r
t,
f
u
r
th
er
co
m
p
licatin
g
t
h
e
p
r
o
ce
s
s
o
f
b
u
ild
in
g
ac
cu
r
ate
cl
ass
if
ier
s
f
o
r
n
ativ
e
lan
g
u
ag
es
[
1
5
]
.
A
d
d
itio
n
ally
,
wh
ile
s
o
m
e
class
if
ier
s
m
ay
s
h
o
w
p
r
o
m
is
in
g
r
esu
lts
,
th
ey
m
a
y
f
o
cu
s
o
n
s
p
ec
if
ic
d
o
m
ain
s
lik
e
e
-
c
o
m
m
er
ce
s
ites
an
d
o
v
er
l
o
o
k
o
th
e
r
p
latf
o
r
m
s
lik
e
s
o
cial
m
e
d
ia
o
r
v
id
eo
c
o
n
ten
t
s
ites
,
lead
in
g
to
a
lack
o
f
d
iv
e
r
s
ity
in
s
en
tim
en
t
an
aly
s
is
ap
p
r
o
ac
h
es
[
1
6
]
.
As
a
r
esu
lt,
th
er
e
is
a
p
r
ess
in
g
n
ee
d
f
o
r
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
r
esear
ch
t
h
at
co
n
s
id
er
s
d
iv
er
s
e
lin
g
u
is
tic
co
n
tex
ts
,
s
elec
ts
r
elev
an
t
f
ea
tu
r
es
ac
r
o
s
s
v
ar
io
u
s
d
o
m
ain
s
,
an
d
a
d
d
r
ess
es
th
e
ch
allen
g
es
p
o
s
ed
b
y
n
ativ
e
lan
g
u
ag
e
s
en
tim
en
t
an
aly
s
is
.
B
y
o
v
er
co
m
i
n
g
th
ese
is
s
u
es,
m
o
r
e
d
ee
p
er
in
s
ig
h
ts
in
to
p
u
b
lic
s
en
tim
en
ts
ac
r
o
s
s
d
if
f
er
en
t
c
u
ltu
r
al
an
d
lin
g
u
is
tic
lan
d
s
ca
p
es,
f
ac
ilit
atin
g
m
o
r
e
ac
cu
r
ate
s
en
t
im
en
t
an
aly
s
is
in
d
iv
e
r
s
e
d
ig
it
al
en
v
ir
o
n
m
en
ts
ca
n
b
e
ac
q
u
ir
ed
.
T
o
a
d
d
r
ess
th
e
af
o
r
em
en
tio
n
ed
ch
allen
g
es,
th
is
wo
r
k
m
a
k
es
th
e
f
o
llo
win
g
co
n
tr
ib
u
tio
n
s
:
i)
Pro
v
id
in
g
a
s
tan
d
ar
d
Kan
n
ad
a
d
ataset
th
at
in
clu
d
es
s
en
tim
en
t
asp
ec
ts
,
p
o
lar
ity
lab
el
s
,
an
d
o
v
er
all
s
en
tim
en
t
a
n
aly
s
is
o
f
r
ev
iews
;
ii)
C
o
n
s
id
er
in
g
v
ar
io
u
s
f
ea
tu
r
es
an
d
s
elec
tin
g
th
e
m
o
s
t
r
el
ev
an
t
o
n
es
to
ac
h
iev
e
o
p
tim
al
r
esu
lts
with
in
th
e
g
iv
en
d
o
m
ain
;
iii)
Pro
p
o
s
in
g
a
n
o
v
el
class
if
ier
s
p
ec
if
ic
ally
d
esig
n
ed
f
o
r
Kan
n
a
d
a
l
an
g
u
ag
e
s
en
tim
en
t
an
aly
s
is
;
an
d
i
v
)
C
o
n
d
u
ctin
g
a
co
m
p
ar
ativ
e
a
n
aly
s
is
o
f
th
e
p
r
o
p
o
s
ed
class
if
ier
with
e
x
is
tin
g
ML
ap
p
r
o
ac
h
es,
ev
alu
atin
g
p
e
r
f
o
r
m
an
ce
m
etr
ic
s
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.
B
y
co
m
b
in
in
g
th
ese
co
n
tr
ib
u
tio
n
s
,
th
is
wo
r
k
aim
s
to
ad
v
an
ce
th
e
f
ield
o
f
s
en
tim
en
t
an
aly
s
is
in
Kan
n
ad
a
a
n
d
a
d
d
r
ess
th
e
u
n
iq
u
e
c
h
allen
g
es
p
o
s
ed
b
y
n
ativ
e
lan
g
u
ag
e
s
en
tim
e
n
t
c
lass
if
icatio
n
.
T
h
e
m
an
u
s
cr
ip
t
is
s
tr
u
ctu
r
ed
as
f
o
llo
ws:
s
ec
tio
n
2
p
r
esen
ts
th
e
liter
atu
r
e
s
u
r
v
ey
,
p
r
o
v
id
in
g
in
s
ig
h
ts
in
to
r
ec
en
t
wo
r
k
s
an
d
ad
v
a
n
ce
m
en
ts
in
Kan
n
ad
a
s
en
tim
en
t
an
aly
s
is
,
s
ec
tio
n
3
d
el
v
es
in
to
th
e
m
et
h
o
d
o
lo
g
y
ad
o
p
ted
in
th
is
wo
r
k
,
o
u
tlin
in
g
t
h
e
o
v
er
al
l
s
en
tim
en
t
class
if
icat
io
n
p
r
o
c
ess
b
y
co
n
s
id
er
in
g
r
elev
a
n
t
f
e
atu
r
es.
I
n
s
ec
tio
n
4
,
th
e
r
esu
lts
o
f
t
h
e
p
r
o
p
o
s
ed
cl
ass
if
ier
ar
e
p
r
esen
ted
an
d
d
is
cu
s
s
ed
.
Fin
ally
,
s
ec
tio
n
5
p
r
o
v
i
d
es
th
e
co
n
clu
s
io
n
o
f
th
e
wo
r
k
,
s
u
m
m
a
r
izin
g
th
e
k
ey
f
in
d
in
g
s
,
co
n
tr
ib
u
tio
n
s
,
lim
itatio
n
s
,
an
d
f
u
tu
r
e
d
i
r
ec
tio
n
s
o
f
th
e
r
esear
ch
.
2.
L
I
T
E
R
AT
U
RE
SU
RVE
Y
T
h
e
r
esear
ch
co
n
d
u
cte
d
in
[
1
7
]
ex
p
l
o
r
ed
th
e
d
o
m
ai
n
o
f
s
e
n
tim
en
t
p
r
ed
ictio
n
b
y
u
tili
zin
g
a
d
ataset
co
n
s
is
tin
g
o
f
co
d
e
-
m
ix
ed
Ka
n
n
ad
a
-
E
n
g
lis
h
tex
t
s
o
u
r
ce
d
e
x
tr
ac
ted
f
r
o
m
T
witter
.
T
h
is
d
ataset
was
an
aly
ze
d
with
co
r
r
esp
o
n
d
in
g
s
en
tim
en
ts
f
o
r
ev
er
y
p
o
s
t.
T
h
is
wo
r
k
u
tili
ze
d
d
if
f
er
en
t
f
ea
tu
r
es
s
u
ch
as
r
ep
etitiv
e
ch
ar
ac
ter
s
,
wo
r
d
an
d
ch
a
r
ac
te
r
n
-
g
r
am
s
.
T
h
e
ef
f
icie
n
cy
o
f
t
h
ese
ap
p
r
o
ac
h
es
was
s
u
b
s
eq
u
en
tly
ass
ess
ed
u
s
in
g
s
u
p
p
o
r
t
-
v
ec
to
r
-
m
ac
h
in
es (
SVM)
an
d
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
.
T
h
e
r
esu
lts
o
f
th
e
s
tu
d
y
in
d
icate
d
th
at
SVM
attain
ed
an
av
er
a
g
e
p
r
ec
is
io
n
v
alu
e
th
at
r
a
n
g
ed
f
r
o
m
0
.
2
2
t
o
0
.
2
9
.
C
h
an
d
r
ik
a
et
a
l.
[
1
8
]
aim
ed
t
o
d
el
v
e
in
to
th
e
c
o
m
p
lex
r
ea
lm
o
f
s
en
t
im
en
ts
co
n
v
ey
e
d
th
r
o
u
g
h
th
e
b
io
g
r
ap
h
y
o
r
n
ar
r
ativ
es
p
u
b
lis
h
ed
in
th
e
Kan
n
ad
a
lan
g
u
ag
e.
T
h
eir
p
r
o
p
o
s
ed
m
eth
o
d
in
v
o
lv
e
d
th
e
u
s
e
o
f
a
r
u
le
-
b
ased
s
tr
ateg
y
t
o
ca
teg
o
r
ize
s
en
tim
en
ts
in
Kan
n
ad
a
d
o
c
u
m
en
ts
.
T
h
e
ac
h
i
ev
ed
ac
cu
r
ac
y
s
co
r
e
o
f
8
5
%
was
q
u
ite
n
o
tewo
r
th
y
,
esp
ec
ia
lly
wh
en
co
m
p
a
r
ed
to
th
e
b
a
g
-
of
-
wo
r
d
s
(
B
o
W
)
m
eth
o
d
wh
ic
h
o
n
ly
ac
h
iev
e
d
an
ac
cu
r
ac
y
s
co
r
e
o
f
6
7
%.
T
h
e
n
ee
d
to
co
n
d
u
ct
an
aly
s
is
o
f
s
en
tim
en
t
o
n
Kan
n
ad
a
in
ter
n
et
p
a
g
es,
p
ar
ticu
la
r
ly
n
ews
web
s
ites
,
was
h
ig
h
li
g
h
ted
in
[
1
9
]
.
T
h
e
Dec
is
io
n
-
T
r
ee
(
DT
)
ap
p
r
o
ac
h
was
im
p
lem
en
ted
b
y
th
e
r
esear
ch
er
s
,
in
co
r
p
o
r
atin
g
an
o
p
tim
ized
d
ata
-
d
ictio
n
ar
y
.
T
h
e
p
r
ec
is
io
n
r
ate
ac
h
iev
ed
was
0
.
8
5
,
in
d
icatin
g
th
e
p
r
o
p
o
r
tio
n
o
f
co
r
r
e
ctly
p
r
ed
icted
p
o
s
itiv
e
in
s
tan
ce
s
o
u
t
o
f
th
e
to
tal
p
r
ed
i
cted
p
o
s
itiv
e
in
s
tan
ce
s
.
Su
n
il
an
d
Vin
ay
[
2
0
]
in
v
esti
g
ated
s
en
tim
en
t
ev
alu
atio
n
o
n
tr
a
n
s
lated
in
ter
n
et
-
m
o
v
ie
d
atab
ase
(
I
MD
B
)
m
o
v
ie
r
ev
i
ews
in
Kan
n
a
d
a,
in
a
d
d
itio
n
to
r
ev
iews
s
o
u
r
ce
d
f
r
o
m
m
u
ltip
le
r
e
p
u
tab
le
we
b
s
ites
.
T
h
e
r
esear
ch
er
s
p
u
t
f
o
r
war
d
an
en
s
em
b
le
class
if
icatio
n
ap
p
r
o
ac
h
,
wh
ic
h
in
v
o
lv
ed
co
m
b
in
in
g
m
u
ltip
le
class
if
icatio
n
ap
p
r
o
ac
h
es,
an
d
r
ep
o
r
ted
a
n
o
tab
le
ac
c
u
r
ac
y
s
co
r
e
o
f
8
9
%.
I
n
[
2
1
]
,
d
el
v
ed
in
to
th
e
d
o
m
ai
n
o
f
s
en
tim
en
t
ev
alu
atio
n
en
co
m
p
ass
in
g
th
e
Kan
n
ad
a,
Hin
d
i
an
d
E
n
g
lis
h
lan
g
u
ag
es.
T
o
ac
c
o
m
p
lis
h
th
i
s
task
,
a
n
eu
r
al
n
etwo
r
k
ar
c
h
itectu
r
e
b
ased
o
n
c
o
n
v
o
lu
tio
n
al
-
n
eu
r
al
-
n
etwo
r
k
(
C
NN)
an
d
L
STM
was
em
p
lo
y
ed
.
T
h
e
a
p
p
r
o
ac
h
d
ev
elo
p
ed
in
th
eir
r
esear
ch
d
em
o
n
s
tr
ated
r
em
ar
k
ab
le
lev
els
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
5
7
2
-
6
5
8
2
6574
o
f
ac
cu
r
ac
y
,
with
Hin
d
i d
atasets
ac
h
iev
in
g
a
r
em
ar
k
a
b
le
ac
cu
r
ac
y
lev
el
o
f
9
9
%,
E
n
g
lis
h
d
atasets
ac
h
iev
in
g
an
ac
cu
r
ac
y
lev
el
o
f
9
5
%,
an
d
Ka
n
n
ad
a
d
atasets
ac
h
iev
in
g
ac
cu
r
ac
y
lev
el
o
f
9
9
%.
C
h
ak
r
av
ar
th
i
et
a
l.
[
2
2
]
p
r
esen
ted
a
n
o
v
el
d
ataset
th
at
is
m
u
ltil
in
g
u
al
an
d
m
a
n
u
ally
class
if
ied
in
to
d
if
f
er
en
t
ca
teg
o
r
ies
o
f
p
o
lar
ity
.
T
h
is
d
ataset
f
o
cu
s
es
o
n
th
e
Dr
av
id
ian
lan
g
u
ag
es
an
d
w
as
o
b
tain
ed
th
r
o
u
g
h
s
o
cial
-
m
ed
ia
r
ev
iews.
T
h
e
r
esear
ch
er
s
em
p
lo
y
ed
c
o
n
v
e
n
tio
n
al
ML
ap
p
r
o
ac
h
es
alo
n
g
with
tr
an
s
f
o
r
m
er
ap
p
r
o
ac
h
es
s
u
ch
as
c
r
o
s
s
-
lin
g
u
al
lan
g
u
a
g
e
m
o
d
el,
s
elf
-
s
u
p
e
r
v
is
ed
cr
o
s
s
-
lin
g
u
al
m
o
d
el,
b
i
d
ir
ec
tio
n
al
en
c
o
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
)
,
ch
ar
ac
ter
B
E
R
T
,
d
is
til
B
E
R
T
,
an
d
o
p
tim
ized
B
E
R
T
(
R
o
B
E
R
T
A)
to
co
n
d
u
ct
s
en
tim
en
t
ev
alu
a
tio
n
an
d
in
ap
p
r
o
p
r
iate
lan
g
u
ag
e
id
en
tific
atio
n
.
T
h
e
ef
f
ec
tiv
en
ess
o
f
th
ese
ap
p
r
o
ac
h
es
was
ev
alu
ated
o
n
d
atasets
co
n
tain
in
g
co
d
e
-
m
i
x
ed
tex
t.
Mo
r
eo
v
er
,
th
r
o
u
g
h
th
e
u
tili
za
tio
n
o
f
a
co
n
s
tr
ain
ed
Kan
n
ad
a
c
o
r
p
o
r
a,
Sh
an
k
ar
a
n
d
Swam
y
[
2
3
]
em
p
lo
y
ed
SVM
to
e
n
h
an
ce
s
en
tim
en
t
ca
teg
o
r
izatio
n
.
T
h
eir
ap
p
r
o
ac
h
y
ield
e
d
n
o
ta
b
le
r
esu
lts
,
with
a
clas
s
if
icatio
n
ac
cu
r
ac
y
o
f
6
8
%,
p
r
ec
is
io
n
o
f
8
0
%,
r
ec
all
o
f
6
2
%,
alo
n
g
with
an
F
-
S
co
r
e
o
f
7
0
%.
Sh
a
n
k
ar
et
a
l.
[
2
4
]
i
n
v
esti
g
ated
th
e
ap
p
licatio
n
o
f
ML
an
d
en
s
em
b
le
ap
p
r
o
ac
h
es
f
o
r
s
en
tim
en
t
ass
ess
m
en
t
o
n
C
OVI
D
-
1
9
d
ata
in
th
e
Kan
n
ad
a
lan
g
u
ag
e.
Var
io
u
s
ap
p
r
o
ac
h
es
in
clu
d
in
g
lo
g
is
tic
r
eg
r
ess
io
n
(
L
R
)
,
r
an
d
o
m
f
o
r
est
(
R
F),
XGBo
o
s
t
(
XGB),
Ad
aBo
o
s
t
(
A
B
)
,
v
o
tin
g
class
if
ier
(
VC
)
,
an
d
g
r
a
d
ien
t
b
o
o
s
tin
g
(
GB
)
,
wer
e
em
p
lo
y
ed
t
o
an
aly
ze
th
e
d
ata.
T
h
e
r
esu
lts
in
d
icate
d
th
at
th
e
ac
h
iev
e
d
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F
-
Sco
r
e
r
an
g
ed
f
r
o
m
6
2
%
t
o
6
8
%.
Fin
ally
,
a
s
tu
d
y
co
n
d
u
cted
b
y
Sre
elak
s
h
m
i
et
a
l.
[
2
5
]
u
tili
ze
d
d
i
f
f
er
en
t
tr
an
s
f
o
r
m
er
-
b
ased
em
b
ed
d
in
g
a
p
p
r
o
ac
h
es
to
a
n
aly
ze
a
n
d
id
e
n
tify
s
u
ch
co
n
te
n
t.
T
h
e
r
esear
ch
er
s
d
is
co
v
er
ed
th
at
th
e
Mu
R
I
L
p
r
e
-
tr
ain
ed
em
b
ed
d
in
g
(
a
p
ar
t
o
f
B
E
R
T
)
ex
h
ib
ited
u
n
if
o
r
m
an
d
im
p
r
ess
iv
e
p
er
f
o
r
m
an
ce
th
r
o
u
g
h
o
u
t
v
ar
i
o
u
s
d
atasets
.
I
t
a
ch
iev
ed
r
em
ar
k
a
b
ly
h
ig
h
ac
c
u
r
ac
y
r
ates,
v
ar
y
i
n
g
f
r
o
m
6
6
%
to
9
6
%,
wh
en
a
p
p
lied
to
Kan
n
ad
a
,
T
am
il
an
d
Ma
lay
alam
d
atasets
.
Fro
m
th
e
ab
o
v
e
liter
atu
r
e
s
u
r
v
ey
,
it
is
s
ee
n
th
at
v
er
y
lim
ited
lab
elled
d
ata
is
av
ailab
l
e
f
o
r
Kan
n
ad
a
s
en
tim
en
t
an
al
y
s
is
.
Als
o
,
it
is
s
ee
n
th
at
m
o
s
t
o
f
th
e
wo
r
k
s
h
av
e
u
tili
ze
d
B
E
R
T
ap
p
r
o
ac
h
to
a
ch
iev
e
b
etter
ac
cu
r
ac
y
r
ates.
So
m
e
o
f
th
e
wo
r
k
s
h
av
e
u
s
ed
ML
class
if
ier
s
,
b
u
t
h
av
e
ac
h
iev
ed
less
ac
cu
r
ac
y
wh
ich
r
an
g
es
f
r
o
m
6
6
%
to
8
0
%.
Hen
ce
,
in
th
is
wo
r
k
,
we
p
r
esen
t
a
n
o
v
el
Ka
n
n
ad
a
d
ataset
b
y
co
n
v
er
tin
g
t
h
e
Sem
E
v
al
2
0
1
4
T
ask
4
d
at
aset.
Fu
r
th
er
,
in
t
h
is
wo
r
k
we
p
r
esen
t
a
m
o
d
el
b
ased
o
n
B
E
R
T
ca
lled
as
Kan
n
ad
a
-
B
E
R
T
p
r
o
b
ab
ilit
y
clu
s
ter
in
g
class
if
ier
(K
-
B
E
R
T
-
PC
)
.
T
h
e
m
eth
o
d
o
l
o
g
y
f
o
r
th
e
K
-
B
E
R
T
-
PC
cla
s
s
i
f
ier
is
d
is
cu
s
s
ed
in
th
e
n
ex
t se
ctio
n
.
3.
M
E
T
H
O
D
3
.
1
.
Arc
hite
ct
ure
T
h
e
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
s
e
n
t
i
m
e
n
t
c
l
a
s
s
i
f
i
c
at
i
o
n
u
s
i
n
g
K
-
B
E
R
T
-
P
C
c
la
s
s
i
f
i
e
r
is
p
r
e
s
e
n
te
d
i
n
F
i
g
u
r
e
1
.
I
t
b
e
g
i
n
s
wi
t
h
t
h
e
ac
q
u
i
s
it
i
o
n
o
f
r
a
w
t
e
x
t
d
a
t
a
f
r
o
m
t
h
e
S
e
m
E
v
a
l
2
0
1
4
T
as
k
4
d
at
a
s
et
,
r
en
o
w
n
e
d
f
o
r
i
ts
r
ic
h
a
s
p
e
c
ts
,
a
s
p
e
ct
t
e
r
m
s
,
a
n
d
p
o
la
r
i
t
y
a
n
n
o
t
a
t
i
o
n
s
.
T
h
e
n
e
x
t
s
tep
i
n
v
o
l
v
e
s
t
h
e
c
o
n
v
e
r
s
i
o
n
o
f
t
h
i
s
E
n
g
l
i
s
h
d
a
t
a
t
o
K
a
n
n
a
d
a
u
s
i
n
g
G
o
o
g
l
e
T
r
a
n
s
lat
e
,
e
x
p
a
n
d
i
n
g
t
h
e
l
i
n
g
u
is
ti
c
d
i
v
e
r
s
i
t
y
o
f
t
h
e
d
a
t
as
e
t
.
F
o
l
l
o
wi
n
g
d
a
t
a
,
t
h
e
r
aw
d
a
t
a
u
n
d
e
r
g
o
e
s
p
r
e
p
r
o
c
e
s
s
i
n
g
t
o
e
n
h
a
n
c
e
i
t
s
s
u
i
t
a
b
i
l
it
y
f
o
r
a
n
a
l
y
s
i
s
.
F
u
r
t
h
e
r
,
K
-
B
E
R
T
a
n
d
p
r
o
b
a
b
i
l
i
t
y
-
c
l
u
s
t
e
r
i
n
g
(
P
C
)
m
o
d
e
l
u
ti
l
iz
e
d
t
o
e
v
a
l
u
a
te
s
e
n
t
i
m
e
n
t
p
o
l
a
r
i
t
y
(
p
o
s
i
t
i
v
e
,
n
e
g
a
t
i
v
e
,
o
r
n
e
u
t
r
a
l
)
f
o
r
e
a
c
h
t
e
x
t
s
e
g
m
e
n
t
.
Fig
u
r
e
1
.
Ar
c
h
itectu
r
e
o
f
K
-
B
E
R
T
-
PC
clas
s
if
ier
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
cin
g
s
en
timen
t a
n
a
lysi
s
in
K
a
n
n
a
d
a
texts b
y
fea
tu
r
e
s
elec
tio
n
(
S
u
n
il Mu
g
a
lih
a
lli E
s
h
w
a
r
a
p
p
a
)
6575
3
.
2
.
Da
t
a
s
et
co
llect
i
o
n
I
n
th
is
s
tu
d
y
,
th
is
wo
r
k
in
itial
ly
f
o
c
u
s
ed
o
n
th
e
Sem
E
v
al
2
0
1
4
T
ask
4
d
ataset
[
2
6
]
.
T
h
is
d
ataset
is
r
en
o
wn
ed
f
o
r
its
r
ich
co
llecti
o
n
o
f
asp
ec
ts
,
asp
ec
t
ter
m
s
,
an
d
p
o
lar
ities
,
m
ak
in
g
it
a
v
a
lu
ab
le
r
eso
u
r
ce
f
o
r
s
en
tim
en
t
an
aly
s
is
an
d
r
elat
ed
task
s
.
T
o
f
u
r
th
er
en
h
a
n
c
e
th
e
d
ataset’
s
u
tili
ty
an
d
d
iv
er
s
ity
,
th
is
wo
r
k
tr
an
s
lated
th
e
Sem
E
v
al
d
atas
et
f
r
o
m
E
n
g
lis
h
to
Kan
n
ad
a
u
s
in
g
Go
o
g
le
T
r
an
s
late.
T
h
is
tr
an
s
latio
n
p
r
o
ce
s
s
allo
wed
u
s
to
cr
ea
te
b
o
th
tr
ai
n
in
g
an
d
test
in
g
d
ata
in
Ka
n
n
a
d
a,
th
er
e
b
y
ex
p
an
d
i
n
g
th
e
s
co
p
e
an
d
a
p
p
licab
ilit
y
o
f
th
e
wo
r
k
s
an
aly
s
is
.
T
h
e
d
ec
is
io
n
to
u
tili
ze
th
e
Sem
E
v
al
2
0
1
4
d
ataset
s
tem
s
f
r
o
m
its
co
m
p
r
eh
e
n
s
iv
e
co
v
er
ag
e
o
f
v
ar
io
u
s
asp
ec
ts
an
d
d
if
f
er
e
n
t
p
o
lar
ities
.
B
y
lev
er
ag
in
g
th
is
d
ataset
an
d
tr
an
s
latin
g
it
in
to
Kan
n
ad
a,
th
is
wo
r
k
aim
ed
t
o
ex
p
lo
r
e
s
en
tim
en
t
an
al
y
s
is
with
in
th
e
Kan
n
a
d
a
lan
g
u
a
g
e
d
o
m
ai
n
wh
ile
b
en
ef
itin
g
f
r
o
m
t
h
e
well
-
s
tr
u
ctu
r
ed
an
d
an
n
o
tated
Sem
E
v
al
d
ataset.
3
.
3
.
P
re
pro
ce
s
s
ing
I
n
th
e
p
r
ep
r
o
ce
s
s
in
g
s
tag
e,
th
e
f
ir
s
t
s
tep
in
v
o
lv
e
d
to
k
e
n
iz
atio
n
,
wh
ich
is
th
e
p
r
o
ce
s
s
o
f
b
r
ea
k
i
n
g
d
o
wn
a
s
en
ten
ce
i
n
to
in
d
iv
id
u
al
wo
r
d
s
o
r
to
k
en
s
.
T
h
is
s
tep
aid
s
in
s
tr
u
ctu
r
in
g
t
h
e
tex
t
f
o
r
f
u
r
th
er
an
aly
s
is
b
y
s
ep
ar
atin
g
it
i
n
to
its
c
o
n
s
titu
en
t
elem
en
ts
.
Fo
llo
win
g
t
o
k
en
i
za
tio
n
,
n
o
r
m
aliza
tio
n
was
p
e
r
f
o
r
m
ed
to
co
n
v
er
t
u
p
p
er
ca
s
e
wo
r
d
s
to
lo
wer
ca
s
e,
en
s
u
r
in
g
u
n
i
f
o
r
m
ity
in
th
e
tex
t
d
ata.
Ad
d
itio
n
ally
,
s
y
m
b
o
ls
wer
e
r
em
o
v
ed
f
r
o
m
th
e
r
aw
d
ata
to
elim
in
at
e
n
o
is
e
a
n
d
s
tr
ea
m
lin
e
th
e
tex
t.
Su
b
s
eq
u
e
n
tly
,
th
e
d
ata
u
n
d
er
wen
t
s
tem
m
in
g
,
lem
m
atiza
tio
n
,
an
d
s
to
p
wo
r
d
r
em
o
v
al
p
r
o
ce
s
s
es.
Stem
m
in
g
in
v
o
lv
ed
r
ed
u
cin
g
wo
r
d
s
to
th
eir
r
o
o
t
o
r
b
ase
f
o
r
m
,
lem
m
atiza
tio
n
in
v
o
l
v
e
d
g
r
o
u
p
i
n
g
in
f
lecte
d
f
o
r
m
s
o
f
wo
r
d
s
to
th
eir
lem
m
a
o
r
d
ictio
n
a
r
y
f
o
r
m
,
an
d
s
to
p
wo
r
d
r
em
o
v
al
in
v
o
l
v
ed
elim
in
atin
g
co
m
m
o
n
wo
r
d
s
th
at
d
o
n
o
t
c
o
n
tr
ib
u
te
s
ig
n
if
ican
tly
to
t
h
e
m
ea
n
in
g
o
f
th
e
tex
t.
3
.
4
.
K
-
B
E
RT
T
h
e
B
E
R
T
ap
p
r
o
ac
h
is
an
ML
ap
p
r
o
ac
h
th
at
u
tili
ze
s
tr
an
s
f
o
r
m
er
s
f
o
r
p
r
e
-
tr
ain
in
g
tex
tu
al
d
ata
in
th
e
f
ield
o
f
n
atu
r
al
-
la
n
g
u
a
g
e
-
p
r
o
c
ess
in
g
(
NL
P)
[
2
7
]
.
T
h
is
wo
r
k
aim
s
to
o
p
tim
ize
th
e
p
e
r
f
o
r
m
an
ce
o
f
B
E
R
T
b
y
in
co
r
p
o
r
atin
g
ad
d
itio
n
al
lay
er
s
s
u
ch
as
d
e
n
s
e
an
d
class
if
icatio
n
lay
er
s
a
n
d
p
r
esen
tin
g
a
n
o
v
el
B
E
R
T
m
o
d
el
ca
lled
as K
-
B
E
R
T
f
o
r
th
e
Kan
n
ad
a
s
en
tim
en
t a
n
aly
s
is
as p
r
e
s
en
ted
in
Fig
u
r
e
2
.
Fig
u
r
e
2
.
K
-
B
E
R
T
class
if
ier
Fu
r
th
er
m
o
r
e
,
th
e
tr
a
n
s
f
o
r
m
er
-
en
co
d
e
r
u
tili
ze
d
in
K
-
B
E
R
T
ap
p
r
o
ac
h
h
as
b
ee
n
s
u
b
jecte
d
to
f
in
e
-
tu
n
in
g
s
p
ec
if
ically
f
o
r
th
e
Ka
n
n
ad
a
d
ataset
in
th
is
s
tu
d
y
.
I
n
itially
,
in
th
is
wo
r
k
,
th
e
p
r
e
-
p
r
o
ce
s
s
ed
r
ev
iew
was
tak
en
as
in
p
u
t
f
o
r
th
e
K
-
B
E
R
T
.
I
n
s
tead
o
f
u
s
in
g
th
e
u
n
i
g
r
am
s
as
in
p
u
t,
t
h
e
K
-
B
E
R
T
ap
p
r
o
ac
h
u
s
es
th
e
’
’
(
)
p
ar
t
-
of
-
s
p
ee
ch
(
Po
S)
tag
g
in
g
m
eth
o
d
[
2
8
]
,
[
2
9
]
.
As
p
r
esen
ted
in
Fig
u
r
e
2
,
th
e
r
e
ar
e
two
to
k
en
s
p
r
esen
t
i
n
in
p
u
t,
i
.
e.
,
at
r
ig
h
t
a
s
ep
ar
ato
r
to
k
en
(
SEP)
an
d
at
lef
t
a
class
if
icatio
n
to
k
en
(
C
L
S).
T
h
e
f
u
n
ctio
n
o
f
SEP
is
to
ac
t
as
a
s
ep
ar
ato
r
b
etwe
en
ea
ch
tr
ig
r
am
in
th
e
in
p
u
t
tex
t.
T
h
e
SEP
to
k
en
en
s
u
r
es
th
at
ea
ch
tr
ig
r
a
m
is
d
is
tin
ct
an
d
id
en
tifia
b
le
with
in
th
e
i
n
p
u
t
s
eq
u
e
n
ce
.
Als
o
,
th
e
f
u
n
ctio
n
o
f
C
L
S
is
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
5
7
2
-
6
5
8
2
6576
class
if
y
o
r
r
ep
r
esen
t e
ac
h
tr
ig
r
am
in
s
u
ch
a
way
th
at
it c
o
n
tr
i
b
u
tes to
ac
h
iev
in
g
a
m
ea
n
in
g
f
u
l c
lass
if
icatio
n
f
o
r
th
e
en
tire
s
en
ten
ce
o
r
te
x
t.
T
h
e
C
L
S
to
k
en
is
ty
p
ically
u
s
e
d
in
B
E
R
T
m
o
d
els
as
th
e
f
ir
s
t
to
k
en
i
n
th
e
i
n
p
u
t
s
eq
u
en
ce
a
n
d
is
ass
o
ciate
d
with
task
s
lik
e
s
en
tim
en
t
an
aly
s
is
,
wh
er
e
th
e
m
o
d
el
lear
n
s
to
class
if
y
th
e
s
en
tim
en
t
o
f
th
e
e
n
tire
s
en
ten
ce
b
ased
o
n
t
h
e
r
e
p
r
esen
tatio
n
g
en
e
r
ated
b
y
th
e
C
L
S
to
k
e
n
.
Ad
d
itio
n
ally
,
t
h
is
wo
r
k
co
n
s
id
er
s
o
n
e
-
h
o
t
en
c
o
d
in
g
ap
p
r
o
ac
h
f
o
r
tr
ig
r
am
s
.
T
h
is
en
co
d
in
g
m
eth
o
d
h
el
p
ed
in
cr
ea
tin
g
a
n
u
m
er
ical
r
ep
r
esen
tatio
n
o
f
tr
ig
r
am
s
th
at
wer
e
f
ed
in
to
K
-
B
E
R
T
f
o
r
s
en
tim
en
t
a
n
aly
s
is
task
s
.
Fu
r
th
er
,
th
e
tr
ig
r
am
s
,
wer
e
co
n
v
er
ted
i
n
to
p
o
s
itio
n
al
an
d
to
k
en
em
b
ed
d
i
n
g
s
.
I
n
p
o
s
itio
n
al
em
b
e
d
d
in
g
,
th
e
in
p
u
t
s
en
ten
ce
is
d
ef
in
ed
as
=
{
1
,
2
,
3
,
.
.
.
}
∈
×
,
wh
er
e
r
ep
r
es
en
ts
th
e
p
o
s
itio
n
o
f
wo
r
d
,
r
ep
r
esen
ts
r
ev
iew
s
en
ten
ce
an
d
E
r
ep
r
esen
ts
d
im
en
s
io
n
o
f
p
o
s
itio
n
em
b
e
d
d
in
g
.
Mo
r
eo
v
e
r
,
th
e
p
o
s
itio
n
al
em
b
ed
d
in
g
is
ev
alu
ated
as (
1
)
,
(
2
)
.
(
,
2
)
=
(
1000
2
ℎ
)
(
1
)
(
,
2
1
)
=
(
1000
2
ℎ
)
(
2
)
wh
er
e,
r
ep
r
esen
ts
th
e
p
o
s
itio
n
,
r
ep
r
esen
ts
em
b
ed
d
in
g
d
i
m
en
s
io
n
,
2
an
d
2
1
in
d
icate
alter
n
atin
g
d
im
en
s
io
n
s
o
r
in
d
ices
with
in
th
e
p
o
s
itio
n
al
em
b
ed
d
i
n
g
.
T
he
eq
u
atio
n
(
1
)
an
d
(
2
)
g
en
e
r
a
tes
a
s
in
u
s
o
id
al
o
r
cy
clica
l
p
atter
n
in
th
e
p
o
s
itio
n
al
em
b
ed
d
in
g
b
ased
o
n
th
e
p
o
s
itio
n
o
f
t
o
k
en
s
in
t
h
e
in
p
u
t
s
eq
u
en
ce
.
Fu
r
th
e
r
,
th
e
to
k
en
em
b
ed
d
in
g
s
ar
e
d
ef
in
ed
as
W
=
{
W
1
,
W
2
,
W
3
,
.
.
.
W
n
}
∈
×
W
,
wh
er
e
r
ep
r
esen
ts
to
k
en
s
f
o
r
th
e
wo
r
d
s
a
n
d
W
r
ep
r
esen
ts
d
i
m
en
s
io
n
o
f
to
k
e
n
em
b
ed
d
in
g
.
B
o
th
th
ese
em
b
e
d
d
in
g
s
,
i.e
.
,
an
d
ar
e
co
n
ca
ten
ated
.
Hen
ce
,
t
h
e
co
n
c
aten
ated
em
b
e
d
d
in
g
s
ar
e
ac
h
iev
ed
as
=
E
+
E
.
T
h
e
is
u
s
ed
as
an
in
p
u
t
f
o
r
t
h
e
tr
an
s
f
o
r
m
er
en
co
d
er
lay
e
r
.
I
n
th
e
s
tan
d
ar
d
B
E
R
T
tr
an
s
f
o
r
m
er
en
c
o
d
in
g
lay
er
,
ea
c
h
co
m
p
o
n
en
t
p
lay
s
a
cr
u
cial
r
o
le
in
p
r
o
ce
s
s
in
g
th
e
in
p
u
t
d
ata
an
d
ex
tr
ac
tin
g
m
ea
n
in
g
f
u
l
r
ep
r
esen
tatio
n
s
.
T
h
e
in
p
u
t
em
b
ed
d
in
g
lay
er
co
n
v
er
ts
to
k
en
s
o
r
wo
r
d
s
in
th
e
in
p
u
t
s
eq
u
en
ce
in
to
h
ig
h
-
d
im
en
s
io
n
al
v
ec
to
r
r
ep
r
esen
tatio
n
s
,
en
co
d
in
g
s
e
m
an
tic
in
f
o
r
m
atio
n
b
ased
o
n
p
r
e
-
tr
ain
e
d
o
r
lear
n
ed
em
b
e
d
d
in
g
s
.
T
h
e
p
o
s
itio
n
al
en
co
d
i
n
g
lay
e
r
p
r
o
v
id
es
r
elativ
e
p
o
s
itio
n
al
in
f
o
r
m
atio
n
,
aid
in
g
in
d
is
tin
g
u
is
h
in
g
to
k
en
o
r
d
er
a
n
d
en
s
u
r
in
g
d
is
tin
ct
em
b
ed
d
i
n
g
s
f
o
r
t
o
k
en
s
with
s
im
ilar
m
ea
n
in
g
s
b
u
t
d
if
f
er
en
t
p
o
s
itio
n
s
.
T
h
e
m
u
lti
-
h
ea
d
atten
tio
n
lay
er
c
o
m
p
u
tes
atten
tio
n
s
co
r
es
b
etwe
en
to
k
e
n
s
,
allo
win
g
f
o
cu
s
o
n
r
ele
v
an
t
p
ar
ts
o
f
th
e
in
p
u
t
s
eq
u
en
ce
an
d
ca
p
tu
r
in
g
co
m
p
lex
d
ep
en
d
e
n
cies
with
m
u
ltip
le
atten
tio
n
h
ea
d
s
.
Fin
ally
,
t
h
e
f
u
lly
co
n
n
ec
ted
lay
er
,
c
o
m
p
r
is
in
g
th
e
“a
d
d
an
d
n
o
r
m
”
lay
er
a
n
d
th
e
f
ee
d
-
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
,
in
te
g
r
ates
atten
tio
n
o
u
t
p
u
ts
with
in
p
u
t
e
m
b
ed
d
in
g
s
,
ap
p
lies
n
o
r
m
aliza
tio
n
,
an
d
u
tili
ze
s
n
o
n
-
lin
ea
r
tr
an
s
f
o
r
m
atio
n
s
to
ca
p
tu
r
e
in
tr
icate
p
atter
n
s
,
th
u
s
en
ab
lin
g
th
e
ex
tr
ac
tio
n
o
f
h
ier
ar
ch
i
ca
l r
ep
r
esen
tatio
n
s
v
ital
f
o
r
v
ar
i
o
u
s
n
atu
r
al
la
n
g
u
ag
e
p
r
o
ce
s
s
in
g
task
s
.
I
n
t
h
is
wo
r
k
,
th
e
m
u
lti
-
h
ea
d
atten
tio
n
lay
er
h
as
b
ee
n
f
in
e
-
tu
n
ed
f
o
r
s
en
tim
en
t a
n
al
y
s
is
.
T
h
e
m
u
lti
-
h
ea
d
atten
tio
n
lay
e
r
h
as b
ee
n
s
h
o
wn
in
Fig
u
r
e
3
.
Fig
u
r
e
3
.
Mu
lti
-
h
ea
d
atten
tio
n
f
r
am
ewo
r
k
I
n
th
e
m
u
lti
-
h
ea
d
atten
tio
n
p
r
o
ce
s
s
,
d
is
tin
ct
q
u
er
y
v
ec
to
r
s
ar
e
d
ef
in
ed
f
o
r
d
if
f
er
e
n
t
s
eq
u
en
ce
task
s
.
T
y
p
ically
,
th
e
in
p
u
t
t
o
an
atte
n
tio
n
f
u
n
ctio
n
c
o
m
p
r
is
es
a
q
u
er
y
an
d
a
s
et
o
f
k
ey
s
,
w
h
er
e
e
ac
h
k
ey
is
d
en
o
ted
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
cin
g
s
en
timen
t a
n
a
lysi
s
in
K
a
n
n
a
d
a
texts b
y
fea
tu
r
e
s
elec
tio
n
(
S
u
n
il Mu
g
a
lih
a
lli E
s
h
w
a
r
a
p
p
a
)
6577
as
=
{
1
,
2
,
3
,
.
.
.
}
.
It
i
s
im
p
o
r
tan
t
to
n
o
te
th
at
in
NL
P
task
s
,
=
.
T
h
e
s
ca
led
d
o
t
p
r
o
d
u
c
t
atten
tio
n
f
o
r
m
u
lti
-
h
ea
d
atten
t
io
n
is
co
m
p
u
te
d
b
y
ca
lcu
latin
g
th
e
weig
h
ts
th
r
o
u
g
h
th
e
d
o
t
p
r
o
d
u
ct
o
p
e
r
atio
n
b
etwe
en
th
e
an
d
th
e
q
u
er
y
.
T
h
is
p
r
o
ce
s
s
r
esu
lts
in
a
d
is
tr
ib
u
tio
n
o
f
atten
tio
n
s
co
r
es,
wh
ich
is
d
eter
m
in
ed
u
s
in
g
(
3
)
.
I
n
(
3
)
,
th
e
v
ar
iab
le
s
ig
n
if
ies
th
e
s
e
m
an
tic
g
ain
,
a
p
ar
am
ete
r
cr
u
c
ial
f
o
r
d
eter
m
in
in
g
th
e
s
em
an
tic
r
elev
an
ce
b
etwe
en
a
wo
r
d
a
n
d
its
co
n
tex
t.
T
h
is
ca
lcu
latio
n
p
r
o
ce
s
s
is
r
ef
e
r
r
ed
as
th
e
s
co
r
in
g
p
r
o
ce
s
s
wh
ich
p
lay
s
a
p
iv
o
tal
r
o
le
in
ass
es
s
in
g
th
e
s
ig
n
if
ic
an
ce
o
f
th
e
wo
r
d
with
in
its
s
u
r
r
o
u
n
d
in
g
co
n
tex
t
.
T
h
e
s
co
r
in
g
p
r
o
ce
s
s
its
elf
is
co
m
p
u
ted
u
s
in
g
(
4
)
.
(
,
)
=
(
(
,
)
)
(
3
)
=
ℎ
(
[
,
]
.
)
(
4
)
Her
e,
d
en
o
tes
th
e
s
em
an
tic
g
ain
o
r
r
ele
v
an
ce
s
co
r
e,
r
ep
r
esen
ts
th
e
k
ey
v
ec
to
r
,
[
,
]
r
ep
r
e
s
en
ts
th
e
co
n
ca
ten
atio
n
o
f
th
e
k
ey
v
ec
t
o
r
an
d
q
u
e
r
y
v
ec
to
r
,
r
ep
r
esen
ts
th
e
weig
h
t
m
atr
ix
ass
o
ciate
d
with
th
e
atten
tio
n
p
r
o
ce
s
s
,
an
d
ℎ
d
en
o
tes th
e
h
y
p
er
b
o
lic
tan
g
en
t f
u
n
cti
o
n
.
T
h
e
d
o
t p
r
o
d
u
ct
[
,
]
.
f
o
llo
wed
b
y
th
e
h
y
p
er
b
o
lic
tan
g
e
n
t
ac
tiv
a
tio
n
f
u
n
ctio
n
co
m
p
u
tes
th
e
f
in
al
r
elev
an
ce
s
co
r
e
b
ased
o
n
th
e
in
ter
ac
tio
n
b
etwe
en
th
e
k
ey
a
n
d
q
u
er
y
v
ec
to
r
s
weig
h
ted
b
y
th
e
atte
n
tio
n
m
atr
i
x
.
I
n
th
is
wo
r
k
,
eig
h
t
h
ea
d
s
we
r
e
em
p
lo
y
ed
with
in
th
e
m
u
lti
-
h
e
ad
atten
tio
n
p
r
o
ce
s
s
,
as
in
d
iv
i
d
u
al
p
a
r
am
eter
s
ca
n
n
o
t
ef
f
ec
tiv
ely
s
h
ar
e
a
s
in
g
le
h
ea
d
.
T
h
is
d
esig
n
c
h
o
ice
s
tem
s
f
r
o
m
th
e
u
tili
za
tio
n
o
f
lin
ea
r
p
r
o
jectio
n
s
to
g
lean
in
f
o
r
m
atio
n
f
r
o
m
d
is
tin
ct
s
u
b
s
p
ac
es,
wh
er
e
th
e
v
alu
es o
f
k
ey
s
(
)
an
d
q
u
er
ies
(
)
ten
d
to
ex
h
ib
it
r
ap
id
f
lu
ctu
atio
n
s
.
T
h
e
f
u
n
ctio
n
in
g
o
f
th
e
m
u
lti
-
h
ea
d
atten
tio
n
p
r
o
ce
s
s
is
f
u
r
th
er
d
ef
in
ed
th
r
o
u
g
h
(
5
)
an
d
(
6
)
,
ex
p
lai
n
in
g
its
r
o
le
in
p
r
o
ce
s
s
in
g
a
n
d
co
n
tex
tu
alizin
g
i
n
p
u
t
d
ata
ef
f
e
ctiv
ely
.
ℎ
=
(
,
)
(
5
)
=
(
ℎ
1
⨁
ℎ
2
⨁
ℎ
3
…
⨁
ℎ
ℎ
)
.
(
6
)
T
h
is
(
5
)
r
ep
r
esen
ts
th
e
ca
lcu
latio
n
o
f
an
atten
tio
n
h
ea
d
(
ℎ
)
with
in
th
e
m
u
lti
-
h
ea
d
atten
t
io
n
p
r
o
ce
s
s
.
T
h
e
atten
tio
n
f
u
n
ctio
n
is
ap
p
lied
t
o
th
e
k
ey
(
)
an
d
q
u
er
y
(
)
to
co
m
p
u
te
t
h
e
atte
n
tio
n
s
co
r
es
.
T
h
e
eq
u
atio
n
(
6
)
r
ep
r
esen
ts
th
e
m
u
lti
-
h
ea
d
atten
tio
n
o
u
tp
u
t.
T
h
e
o
u
tp
u
ts
o
f
all
atten
tio
n
h
ea
d
s
(
ℎ
1
,
ℎ
2
,
ℎ
3
,
…
,
ℎ
)
ar
e
co
n
ca
ten
ated
⊕
an
d
th
en
m
u
ltip
lied
b
y
a
weig
h
t
m
atr
ix
to
o
b
tain
th
e
f
in
al
o
u
tp
u
t
o
f
t
h
e
m
u
lti
-
h
ea
d
atten
tio
n
p
r
o
ce
s
s
.
Als
o
,
ℎ
∈
[
1
,
8
]
s
ig
n
if
ies
th
e
r
an
g
e
o
f
v
alu
e
s
f
o
r
th
e
weig
h
t
wh
ich
b
elo
n
g
s
to
(
(
ℎ
)
)
×
(
(
ℎ
)
)
.
Her
e,
(
ℎ
)
r
ep
r
esen
ts
th
e
h
id
d
en
d
im
en
s
io
n
.
T
h
e
p
u
r
p
o
s
e
o
f
in
tr
o
d
u
cin
g
th
e
atten
tio
n
f
u
n
cti
o
n
is
t
o
o
f
f
er
a
m
o
r
e
d
ir
ec
t
p
ath
way
f
o
r
i
n
p
u
ts
,
th
u
s
ad
d
r
ess
in
g
th
e
v
an
is
h
in
g
-
g
r
ad
ien
t
p
r
o
b
le
m
.
T
h
is
atten
tio
n
m
ec
h
an
is
m
d
ir
ec
ts
f
o
cu
s
b
y
en
h
a
n
cin
g
th
e
h
id
d
en
an
d
o
u
tp
u
t
s
tates
f
r
o
m
p
r
ec
ed
in
g
b
l
o
ck
s
with
a
co
n
tex
t
v
ec
to
r
as
d
ef
in
ed
b
y
(
7
)
.
T
h
e
e
q
u
atio
n
(
7
)
r
ep
r
esen
ts
th
e
co
m
p
u
tatio
n
o
f
th
e
c
o
n
tex
t
v
ec
to
r
,
wh
ich
is
o
b
tain
ed
b
y
s
u
m
m
i
n
g
u
p
th
e
at
ten
tio
n
-
weig
h
ted
r
ep
r
esen
tatio
n
s
o
f
th
e
in
p
u
t
t
o
k
en
s
ℎ
b
ased
o
n
th
eir
r
elev
a
n
ce
s
co
r
es
f
o
r
th
e
ℎ
o
u
tp
u
t
t
o
k
en
.
T
h
is
atten
tio
n
co
m
p
u
tatio
n
is
d
o
n
e
u
s
in
g
(
8
)
an
d
(
9
)
.
=
Σ
=
1
ℎ
(
7
)
=
(
)
=
(
)
Σ
=
1
(
)
(
8
)
=
(
−
1
,
ℎ
)
(
9
)
T
h
e
(
8
)
co
m
p
u
tes
th
e
atten
ti
o
n
weig
h
ts
f
o
r
ea
ch
to
k
en
p
air
(
,
)
in
th
e
in
p
u
t
s
eq
u
en
ce
.
T
h
e
s
o
f
tm
ax
f
u
n
ctio
n
en
s
u
r
es
th
at
th
e
weig
h
ts
s
u
m
u
p
to
1
,
r
ep
r
esen
tin
g
th
e
im
p
o
r
tan
ce
o
r
r
elev
an
ce
o
f
ea
c
h
to
k
en
in
th
e
s
eq
u
en
ce
with
r
esp
ec
t
to
th
e
c
u
r
r
e
n
t
to
k
e
n
b
ein
g
p
r
o
ce
s
s
ed
.
Fu
r
th
e
r
,
in
(
9
)
,
th
e
f
u
n
ctio
n
r
ep
r
esen
ts
th
e
alig
n
m
en
t
m
o
d
el,
wh
ich
ass
ess
e
s
th
e
s
co
r
es
in
d
icatin
g
h
o
w
well
t
h
e
in
p
u
t
s
s
u
r
r
o
u
n
d
in
g
th
e
o
u
tp
u
t
at
th
e
p
o
s
itio
n
alig
n
o
r
co
r
r
elate
.
T
h
e
h
id
d
en
s
tate
−
1
is
d
er
iv
ed
f
r
o
m
th
e
p
r
ev
io
u
s
t
im
e
s
tep
an
d
co
n
tr
ib
u
tes
t
o
th
e
co
m
p
u
tati
o
n
o
f
th
e
alig
n
m
en
t
s
co
r
es.
Mo
r
eo
v
er
,
th
is
wo
r
k
h
as
e
x
p
an
d
e
d
u
p
o
n
th
e
estab
lis
h
ed
B
E
R
T
f
r
am
ewo
r
k
,
as
illu
s
tr
ated
in
Fig
u
r
e
2
,
b
y
in
teg
r
atin
g
a
class
if
icatio
n
lay
er
an
d
a
d
e
n
s
e
lay
er
.
T
h
e
m
ain
p
u
r
p
o
s
e
o
f
t
h
is
en
h
an
ce
m
e
n
t
is
to
e
m
p
o
w
er
th
e
m
o
d
el
t
o
p
r
o
d
u
ce
tag
s
eq
u
en
ce
s
f
o
r
in
p
u
t
s
en
ten
ce
s
.
T
o
m
itig
ate
th
e
p
o
ten
tial
o
f
o
v
er
f
itti
n
g
,
a
d
r
o
p
o
u
t
n
o
r
m
aliza
tio
n
tech
n
iq
u
e
h
a
s
b
ee
n
s
p
ec
if
ically
ap
p
lied
to
th
e
d
e
n
s
e
lay
er
.
Af
ter
th
e
K
-
B
E
R
T
p
r
o
ce
s
s
th
e
s
en
tim
en
t
p
o
lar
ity
is
ac
h
iev
ed
,
i.e
.
,
wh
eth
er
it
is
p
o
s
itiv
e,
n
eg
ativ
e
o
r
n
e
u
tr
al.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
5
7
2
-
6
5
8
2
6578
3
.
5
.
P
r
o
ba
bil
it
y
clus
t
er
ing
Pro
b
ab
ilit
y
-
clu
s
ter
in
g
(
PC
)
h
a
s
em
er
g
ed
as
a
p
o
wer
f
u
l
to
o
l
f
o
r
ex
tr
ac
tin
g
asp
ec
ts
an
d
d
i
s
co
v
er
in
g
h
id
d
en
p
atter
n
s
[
3
0
]
,
[
3
1
]
th
r
o
u
g
h
d
if
f
er
e
n
t
r
ev
iews
.
T
h
e
PC
p
o
s
s
e
s
s
es
th
e
ca
p
ab
ilit
y
to
p
er
f
o
r
m
d
ata
an
aly
s
is
an
d
g
en
er
ate
to
p
ic
-
v
e
cto
r
s
b
y
u
tili
zin
g
a
s
eq
u
en
ce
o
f
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
s
.
T
h
e
PC
ap
p
r
o
ac
h
h
as
d
em
o
n
s
tr
ated
r
e
m
ar
k
ab
le
e
f
f
i
ca
cy
in
th
e
d
o
m
ain
o
f
tex
t
-
cl
u
s
ter
in
g
an
d
t
o
p
ic
-
m
o
d
elin
g
.
I
n
PC
,
th
e
η
an
d
r
ep
r
esen
ts
th
e
Dir
ich
let
p
ar
a
m
eter
(
p
r
o
b
a
b
ilit
y
d
is
tr
ib
u
tio
n
)
,
r
ep
r
esen
ts
to
p
ics
p
r
esen
t
in
th
e
d
ataset
clu
s
ter
,
d
e
n
o
tes
th
e
ℎ
p
o
s
itio
n
ed
wo
r
d
in
t
h
e
co
r
p
u
s
r
ev
iew
,
,
d
en
o
tes
th
e
o
b
s
er
v
ed
w
o
r
d
s
in
th
e
clu
s
ter
o
f
r
ev
iews,
,
d
en
o
tes
p
e
r
-
wo
r
d
t
o
p
ic
an
d
r
ep
r
esen
ts
p
er
-
d
o
c
u
m
en
t
to
p
ic.
T
h
is
wo
r
k
ass
u
m
es
th
e
p
r
io
r
-
d
is
tr
ib
u
tio
n
o
f
r
ev
iews
as
Dir
ich
let
-
d
is
tr
ib
u
tio
n
[
3
2
]
.
Hen
ce
,
th
e
o
v
er
all
d
is
tr
ib
u
ti
o
n
f
o
r
PC
m
o
d
el
is
ev
alu
ated
u
s
in
g
Gib
b
s
-
Sam
p
li
n
g
[
3
3
]
as
p
r
esen
ted
in
(
1
0
)
.
Als
o
,
th
e
co
h
er
en
ce
s
co
r
e
is
e
v
alu
ated
u
s
in
g
(
1
1
)
[
3
2
]
.
(
,
,
,
|
,
)
=
∏
(
|
=
1
)
(
|
)
(
|
)
(
|
)
(
1
0
)
(
⃗
,
⃗
⃗
)
=
∑
∙
|
|
=
1
|
|
⃗
⃗
|
|
2
∙
|
|
⃗
⃗
|
|
2
(
1
1
)
wh
er
e
r
ep
r
esen
ts
th
e
co
n
f
ir
m
atio
n
m
ea
s
u
r
e,
r
ep
r
esen
ts
th
e
p
air
o
f
v
ec
to
r
s
an
d
.
⃗
an
d
⃗
⃗
r
ep
r
esen
t
co
s
in
e
v
ec
to
r
s
im
ilar
ity
.
Af
ter
th
e
ev
alu
atio
n
o
f
th
e
co
h
er
en
ce
s
co
r
e,
th
e
clu
s
ter
o
f
to
p
ics
h
av
in
g
h
ig
h
est
co
h
er
en
ce
s
co
r
e
is
s
elec
ted
an
d
th
en
th
e
to
p
ics ar
e
s
elec
ted
a
n
d
its
r
esp
ec
tiv
e
asp
ec
ts
ar
e
e
x
tr
ac
ted
.
3
.
6
.
K
-
B
E
RT
-
PC
T
o
en
h
an
ce
s
en
tim
e
n
t
class
if
icatio
n
an
d
to
p
ic
m
in
in
g
e
x
tr
ac
tio
n
,
t
h
is
wo
r
k
h
as
in
te
g
r
ated
th
e
K
-
B
E
R
T
clas
s
if
ier
with
PC
u
s
in
g
a
c
o
m
b
in
e
d
r
e
p
r
esen
tatio
n
ap
p
r
o
ac
h
.
T
h
e
c
o
m
b
i
n
ed
r
e
p
r
esen
tatio
n
ap
p
r
o
ac
h
m
e
r
g
es
th
e
s
tr
en
g
t
h
s
o
f
K
-
B
E
R
T
an
d
PC
to
a
ch
iev
e
co
m
p
r
e
h
en
s
iv
e
tex
t
a
n
aly
s
is
.
K
-
B
E
R
T
,
a
cu
s
to
m
ized
B
E
R
T
m
o
d
el
f
o
r
Kan
n
ad
a
s
en
tim
en
t
a
n
aly
s
is
,
ex
ce
ls
in
u
n
d
e
r
s
tan
d
in
g
co
n
t
ex
tu
al
in
f
o
r
m
atio
n
an
d
s
en
tim
en
t
ex
p
r
ess
io
n
s
with
in
tex
t
d
ata.
On
th
e
o
th
er
h
an
d
,
PC
is
b
etter
f
o
r
to
p
ic
m
o
d
elin
g
an
d
asp
ec
t
ex
tr
ac
tio
n
,
u
n
c
o
v
er
in
g
h
id
d
e
n
p
atter
n
s
an
d
th
em
atic
s
tr
u
ctu
r
es
with
in
tex
tu
al
co
n
ten
t.
B
y
in
teg
r
atin
g
th
ese
two
m
o
d
els,
th
e
co
m
b
in
ed
r
e
p
r
esen
tatio
n
ap
p
r
o
ac
h
le
v
er
ag
es
K
-
B
E
R
T
’
s
d
ee
p
co
n
tex
tu
al
u
n
d
er
s
tan
d
i
n
g
f
o
r
s
en
tim
en
t
class
if
icatio
n
an
d
PC
’
s
ca
p
ab
ilit
y
to
ex
tr
ac
t
m
ea
n
in
g
f
u
l
to
p
ics
an
d
asp
ec
ts
f
r
o
m
tex
t
d
ata.
T
h
is
en
ab
les
a
m
o
r
e
h
o
lis
tic
an
aly
s
is
,
p
r
o
v
id
in
g
v
alu
ab
le
in
s
ig
h
ts
in
to
s
en
tim
en
t
tr
en
d
s
,
th
e
m
atic
co
n
ten
t,
a
n
d
u
n
d
er
ly
i
n
g
s
en
tim
en
ts
ass
o
ciate
d
with
s
p
ec
if
ic
to
p
ics
o
r
a
s
p
ec
ts
in
tex
tu
al
co
n
ten
t.
Ov
er
all,
th
e
co
m
b
in
e
d
r
ep
r
esen
tatio
n
a
p
p
r
o
ac
h
m
a
x
i
m
izes
th
e
an
aly
tical
ca
p
a
b
ilit
ies
an
d
u
tili
ty
o
f
b
o
th
K
-
B
E
R
T
an
d
PC
,
en
h
an
cin
g
th
e
o
v
e
r
all
tex
t
an
aly
s
is
p
r
o
ce
s
s
.
I
n
th
e
n
e
x
t
s
ec
tio
n
,
th
e
r
esu
lts
f
o
r
th
e
K
-
B
E
R
T
-
PC
class
if
ier
ar
e
d
is
cu
s
s
ed
an
d
co
m
p
ar
ed
with
e
x
is
tin
g
w
o
r
k
s
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
K
-
B
E
R
T
-
PC
clas
s
if
ier
w
as
ev
alu
ated
o
n
a
W
in
d
o
ws
1
1
o
p
er
atin
g
s
y
s
tem
co
n
s
is
tin
g
o
f
1
6
GB
o
f
R
AM
.
T
h
e
An
ac
o
n
d
a
p
latf
o
r
m
was
co
n
s
id
er
e
d
f
o
r
b
u
ild
in
g
th
e
class
if
ier
.
T
h
e
An
ac
o
n
d
a
b
u
ilt
in
Py
th
o
n
f
r
am
ewo
r
k
was
u
s
ed
.
Mo
s
t
o
f
th
e
Natu
r
al
L
an
g
u
ag
e
T
o
o
lk
it
(
NL
T
K
)
lib
r
ar
ies
wer
e
u
s
ed
f
o
r
e
v
alu
atio
n
.
Fu
r
th
er
,
f
o
r
ev
al
u
atio
n
o
f
K
-
B
E
R
T
-
PC
,
th
e
f
o
llo
win
g
p
er
f
o
r
m
an
ce
m
etr
ics we
r
e
u
s
ed
.
=
+
+
+
+
(
1
2
)
=
+
(
1
3
)
=
+
(
1
4
)
−
=
2
×
×
+
(
1
5
)
wh
er
e
is
tr
u
e
p
o
s
itiv
e
,
is
f
alse
p
o
s
itiv
e
,
is
tr
u
e
n
eg
ativ
e
,
an
d
is
f
al
s
e
n
eg
ativ
e
.
T
h
e
r
esu
lts
wer
e
co
m
p
ar
ed
with
d
if
f
er
e
n
t
clas
s
if
ier
s
p
r
esen
ted
in
[
2
4
]
.
Ad
d
itio
n
ally
,
T
a
b
le
1
p
r
esen
ts
a
s
u
b
s
et
o
f
test
in
g
ex
am
p
les
ex
tr
ac
ted
f
r
o
m
th
e
d
ataset.
T
h
is
tab
le
o
f
f
er
s
a
g
lim
p
s
e
in
to
th
e
ac
tu
al
d
ata
u
s
e
d
f
o
r
ev
alu
atin
g
th
e
class
if
ier
'
s
p
er
f
o
r
m
an
ce
,
s
h
o
w
ca
s
in
g
a
r
ep
r
esen
tativ
e
s
am
p
le
o
f
in
s
tan
ce
s
with
th
eir
co
r
r
esp
o
n
d
in
g
s
en
tim
en
t
lab
els.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
cin
g
s
en
timen
t a
n
a
lysi
s
in
K
a
n
n
a
d
a
texts b
y
fea
tu
r
e
s
elec
tio
n
(
S
u
n
il Mu
g
a
lih
a
lli E
s
h
w
a
r
a
p
p
a
)
6579
T
ab
le
1
.
T
est
d
ataset
S
L.
N
o
.
En
g
l
i
sh
K
a
n
n
a
d
a
1
S
t
r
o
n
g
b
u
i
l
d
t
h
o
u
g
h
w
h
i
c
h
r
e
a
l
l
y
a
d
d
s
t
o
i
t
s
d
u
r
a
b
i
l
i
t
y
.
ಬ
ಲ
ವ
ಾ
ದ
ಾ
ಾ
ಣ
ಆ
ದ
ರೂ
ಇ
ದ
ು
ಜ
ವ
ಾ
ಯೂ
ಅ
ದ
ರ
ಬ
ಾ
ಕ
ೆಗೆ
ಸ
ೆೇ
ಸ
ು
ತದ
ೆ
.
2
I
c
a
n
sa
y
t
h
a
t
I
a
m fu
l
l
y
s
a
t
i
sf
i
e
d
w
i
t
h
t
h
e
p
e
r
f
o
r
m
a
n
c
e
t
h
a
t
t
h
e
c
o
m
p
u
t
e
r
h
a
s
su
p
p
l
i
e
d
.
ಕಂ
ಯ
ೂ
ಟ
ಒ
ದ
ದ
ಕ
ಾ
ಯಾ
ಮತ
ೆ
ಯಂ
ಗೆ
ನ
ಾ
ನ
ು
ಸ
ಂ
ಯ
ಣ
ಾ
ವ
ಾ
ೃ
ತ
ನ
ಾ
ದ
ೆದೇ
ನ
ೆ
ಎ
ಂ
ದ
ು
ನ
ಾ
ನ
ು
ಹ
ೆೇ
ಳ
ಬ
ೆೆ
.
3
Th
e
b
a
t
t
e
r
y
l
i
f
e
i
s
e
x
c
e
l
l
e
n
t
-
6
-
7
h
o
u
r
s
w
i
t
h
o
u
t
c
h
a
r
g
i
n
g
.
ಬ
ಾ
ೂ
ಟ
ಬ
ಾ
ಕ
ೆ
ಅ
ು
ೂ
ತಮ
ವ
ಾ
ದ
ೆ
-
ಚ
ಾ
ಾ
ಾ
ಡ
ದ
ೆಯೇ
6
-
7
ಗಂ
ಟ
ೆಗ
ಳ
ು
.
4
Th
e
A
p
p
l
e
e
n
g
i
n
e
e
r
s
h
a
v
e
n
o
t
y
e
t
d
i
s
c
o
v
e
r
e
d
t
h
e
d
e
l
e
t
e
k
e
y
.
ಆ
ಇ
ಂ
ಯ
ಗಳ
ು
ಇ
ನ
ೂ
ೂ
ಲ
ೇ
ಕ
ೇ
ಯನ
ು
ೂ
ಕಂಡ
ು
ಲ
ೆ
.
5
O
n
l
y
p
r
o
b
l
e
m
t
h
a
t
I
h
a
d
w
a
s
t
h
a
t
t
h
e
t
r
a
c
k
p
a
d
w
a
s
n
o
t
v
e
r
y
g
o
o
d
f
o
r
m
e
,
I
o
n
l
y
h
a
d
a
p
r
o
b
l
e
m
o
n
c
e
o
r
t
w
i
c
e
w
i
t
h
i
t
,
but
p
r
o
b
a
b
l
y
my
c
o
m
p
u
t
e
r
w
a
s
a
b
i
t
d
e
f
e
c
t
i
v
e
.
ನ
ಾ
ನ
ು
ಹ
ೆ
ೂ
ಂ
ದ
ದ
ಒ
ಂದ
ೆೇ
ಸ
ಮ
ಸ
ೊ
ಯಂ
ದ
ರೆ
ಟ
ಾ
ಯಾ
ಾ
ೂ
ನ
ನ
ಗೆ
ು
ಂಬ
ಾ
ಒ
ೆ
ೆ
ಯ
ದ
ಲ
ೆ
,
ನ
ನ
ಗೆ
ಅ
ದ
ರೆ
ೂ
ಂ
ಗೆ
ಒ
ಂ
ದ
ು
ಅ
ಥ
ವ
ಾ
ಎ
ರ
ಡ
ು
ಬ
ಾ
ಾ
ಯ
ಸ
ಮಸ
ೆ
ೂ
ಇ
ು
ತ
,
ಆ
ದ
ರೆ
ಬ
ಹ
ು
ಶಃ
ನನೂ
ಕಂ
ಯ
ೂ
ಟ
ಸ
ವ
ಲ
ಪ
ದ
ೆ
ೂ
ೇ
ಷ
ಯು
ಕತ
ವ
ಾ
ು
ತ
.
6
La
st
l
y
,
W
i
n
d
o
w
s 8
i
s
a
n
n
o
y
i
n
g
.
ಕ
ೊ
ನ
ೆಯದ
ಾ
,
ವ
ಂ
ಡ
ೆ
ೂ
ೇ
8
ಕಕ
.
T
ab
le
2
co
n
s
is
ts
o
f
s
ix
ex
a
m
p
les
in
b
o
th
E
n
g
lis
h
an
d
Kan
n
ad
a
lan
g
u
a
g
es,
alo
n
g
with
th
eir
co
r
r
esp
o
n
d
in
g
p
o
la
r
ity
s
co
r
es
an
d
s
en
tim
en
ts
.
U
p
o
n
an
aly
s
is
,
it
is
o
b
s
er
v
ed
th
at
ex
a
m
p
les
1
,
3
,
an
d
4
co
n
v
ey
p
o
s
itiv
e
s
en
tim
en
ts
with
h
i
g
h
p
o
lar
ity
s
co
r
es
(
0
.
9
5
6
4
,
0
.
8
9
5
4
,
an
d
0
.
9
8
7
5
,
r
esp
e
ctiv
ely
)
,
in
d
icatin
g
s
atis
f
ac
tio
n
o
r
ap
p
r
o
v
al.
C
o
n
v
er
s
ely
,
ex
a
m
p
les
2
,
5
,
an
d
6
ex
p
r
ess
n
e
g
ativ
e
s
en
tim
en
ts
with
lo
wer
p
o
lar
it
y
s
co
r
es
(
0
.
2
3
5
8
,
0
.
1
2
8
5
,
an
d
0
.
0
2
5
,
r
esp
ec
tiv
ely
)
,
h
i
g
h
li
g
h
tin
g
d
is
s
atis
f
ac
tio
n
o
r
cr
iticis
m
.
Ad
d
itio
n
ally
,
ex
am
p
le
5
ex
h
ib
its
d
if
f
er
e
n
t sen
tim
en
t,
m
en
tio
n
in
g
a
s
p
ec
if
ic
p
r
o
b
lem
wh
ile
ac
k
n
o
wled
g
i
n
g
th
e
p
o
s
s
ib
ilit
y
o
f
a
d
ef
ec
tiv
e
p
r
o
d
u
ct.
O
v
er
all,
th
e
d
ataset
s
h
o
wca
s
es
a
r
a
n
g
e
o
f
s
en
tim
en
ts
ex
p
r
ess
ed
b
y
u
s
er
s
,
p
r
o
v
id
i
n
g
v
alu
ab
le
in
s
ig
h
ts
in
to
th
eir
ex
p
er
ien
ce
s
an
d
p
er
ce
p
tio
n
s
.
T
ab
le
2
.
Sen
tim
en
t
e
v
alu
atio
n
S
L.
N
o
.
En
g
l
i
sh
K
a
n
n
a
d
a
P
o
l
a
r
i
t
y
S
c
o
r
e
/
S
e
n
t
i
me
n
t
1
S
t
r
o
n
g
b
u
i
l
d
t
h
o
u
g
h
w
h
i
c
h
r
e
a
l
l
y
a
d
d
s
t
o
i
t
s
d
u
r
a
b
i
l
i
t
y
.
ಬ
ಲ
ವ
ಾ
ದ
ಾ
ಾ
ಣ
ಆ
ದ
ರೂ
ಇ
ದ
ು
ಜ
ವ
ಾ
ಯೂ
ಅ
ದ
ರ
ಬ
ಾ
ಕ
ೆಗೆ
ಸ
ೆೇ
ಸ
ು
ತ
ದ
ೆ
.
0
.
9
5
6
4
/
P
o
si
t
i
v
e
2
Th
e
A
p
p
l
e
e
n
g
i
n
e
e
r
s
h
a
v
e
n
o
t
y
e
t
d
i
s
c
o
v
e
r
e
d
t
h
e
d
e
l
e
t
e
k
e
y
.
ಆ
ಇ
ಂ
ಯ
ಗಳ
ು
ಇ
ನ
ೂ
ೂ
ಲ
ೇ
ಕ
ೇ
ಯನ
ು
ೂ
ಕಂಡ
ು
ಲ
ೆ
.
0
.
2
3
5
8
/
N
e
g
a
t
i
v
e
3
I
c
a
n
sa
y
t
h
a
t
I
a
m fu
l
l
y
s
a
t
i
sf
i
e
d
w
i
t
h
t
h
e
p
e
r
f
o
r
ma
n
c
e
t
h
a
t
t
h
e
c
o
m
p
u
t
e
r
h
a
s s
u
p
p
l
i
e
d
.
ಕಂ
ಯ
ೂ
ಟ
ಒ
ದ
ದ
ಕ
ಾ
ಯಾ
ಮತ
ೆ
ಯಂ
ಗೆ
ನ
ಾ
ನ
ು
ಸ
ಂ
ಯ
ಣ
ಾ
ವ
ಾ
ೃ
ತ
ನ
ಾ
ದ
ೆದೇ
ನ
ೆ
ಎ
ಂ
ದ
ು
ನ
ಾ
ನ
ು
ಹ
ೆೇ
ಳ
ಬ
ೆೆ
.
0
.
8
9
5
4
/
P
o
si
t
i
v
e
4
Th
e
b
a
t
t
e
r
y
l
i
f
e
i
s
e
x
c
e
l
l
e
n
t
-
6
-
7
h
o
u
r
s w
i
t
h
o
u
t
c
h
a
r
g
i
n
g
.
ಬ
ಾ
ೂ
ಟ
ಬ
ಾ
ಕ
ೆ
ಅ
ು
ೂ
ತಮ
ವ
ಾ
ದ
ೆ
-
ಚ
ಾ
ಾ
ಾ
ಡ
ದ
ೆಯೇ
6
-
7
ಗಂ
ಟ
ೆಗ
ಳ
ು
.
0
.
9
8
7
5
/
P
o
si
t
i
v
e
5
O
n
l
y
p
r
o
b
l
e
m
t
h
a
t
I
h
a
d
w
a
s
t
h
a
t
t
h
e
t
r
a
c
k
p
a
d
w
a
s
n
o
t
v
e
r
y
g
o
o
d
f
o
r
me,
I
o
n
l
y
h
a
d
a
p
r
o
b
l
e
m
o
n
c
e
o
r
t
w
i
c
e
w
i
t
h
i
t
,
b
u
t
p
r
o
b
a
b
l
y
m
y
c
o
m
p
u
t
e
r
w
a
s
a
b
i
t
d
e
f
e
c
t
i
v
e
.
ನ
ಾ
ನ
ು
ಹ
ೆ
ೂ
ಂ
ದ
ದ
ಒ
ಂದ
ೆೇ
ಸ
ಮ
ಸ
ೊ
ಯಂ
ದ
ರೆ
ಟ
ಾ
ಯಾ
ಾ
ೂ
ನ
ನ
ಗೆ
ು
ಂ
ಬ
ಾ
ಒ
ೆ
ೆ
ಯದ
ಲ
ೆ
,
ನ
ನ
ಗೆ
ಅ
ದ
ರೆ
ೂ
ಂ
ಗೆ
ಒ
ಂ
ದ
ು
ಅ
ಥ
ವ
ಾ
ಎ
ರಡ
ು
ಬ
ಾ
ಾ
ಯ
ಸ
ಮ
ಸ
ೊ
ಇ
ು
ತ
,
ಆ
ದ
ರೆ
ಬ
ಹ
ು
ಶಃ
ನನೂ
ಕಂ
ಯ
ೂ
ಟ
ಸ
ವ
ಲ
ಪ
ದ
ೆ
ೂ
ೇ
ಷ
ಯು
ಕತವ
ಾ
ು
ತ
.
0
.
1
2
8
5
/
N
e
g
a
t
i
v
e
6
La
st
l
y
,
W
i
n
d
o
w
s 8
i
s
a
n
n
o
y
i
n
g
.
ಕ
ೊ
ನ
ೆಯದ
ಾ
,
ವ
ಂ
ಡ
ೆ
ೂ
ೇ
8
ಕಕ
.
0
.
0
2
5
/
N
e
g
a
t
i
v
e
T
h
e
p
r
o
v
id
ed
Fig
u
r
e
4
p
r
es
en
ts
th
e
ac
cu
r
ac
y
s
co
r
es
o
f
d
if
f
er
en
t
class
if
ier
s
u
s
ed
f
o
r
s
en
tim
en
t
an
aly
s
is
.
Fro
m
th
e
f
ig
u
r
e
,
it is
o
b
s
er
v
ed
th
at
th
e
t
r
ad
itio
n
al
ML
class
if
ier
s
s
u
ch
as R
F,
L
R
,
XGB,
AB
,
an
d
GB
ac
h
iev
e
ac
cu
r
ac
y
s
co
r
es
r
an
g
in
g
f
r
o
m
0
.
6
to
0
.
6
7
.
No
tab
l
y
,
th
e
K
-
B
E
R
T
-
P
C
cla
s
s
if
ier
o
u
tp
er
f
o
r
m
s
all
o
th
er
class
if
ier
s
with
an
im
p
r
ess
iv
e
ac
cu
r
ac
y
s
co
r
e
o
f
0
.
9
1
.
Fig
u
r
e
5
p
r
esen
ts
th
e
p
r
ec
is
io
n
s
co
r
es
o
f
d
if
f
er
en
t
class
if
ier
s
u
s
ed
f
o
r
s
en
tim
en
t
an
aly
s
is
.
Fro
m
f
ig
u
r
e,
it
is
o
b
s
er
v
ed
th
at
th
e
tr
a
d
itio
n
al
ML
class
if
ier
s
s
u
ch
as
R
F,
L
R
,
XG
B
,
AB
,
an
d
GB
ac
h
iev
e
p
r
ec
is
io
n
s
co
r
es
r
an
g
in
g
f
r
o
m
0
.
6
to
0
.
6
7
.
T
h
e
K
-
B
E
R
T
-
P
C
class
if
ier
s
tan
d
s
o
u
t
with
th
e
h
ig
h
est
p
r
ec
is
io
n
s
co
r
e
o
f
0
.
9
2
am
o
n
g
all
class
if
ier
s
.
Fig
u
r
e
6
p
r
esen
ts
th
e
r
ec
all
s
co
r
es
o
f
d
if
f
er
en
t
class
if
ier
s
u
s
ed
f
o
r
s
en
tim
en
t
an
al
y
s
is
.
An
aly
zin
g
th
e
f
ig
u
r
e
r
ev
ea
ls
th
at
tr
ad
itio
n
al
ML
class
if
ier
s
s
u
ch
as
R
F,
L
R
,
XG
B
,
AB
,
an
d
GB
ac
h
iev
e
r
ec
all
s
co
r
es
r
an
g
in
g
f
r
o
m
0
.
6
to
0
.
6
7
.
T
h
e
K
-
B
E
R
T
-
PC
class
if
ier
s
tan
d
s
o
u
t
with
th
e
h
ig
h
est
r
ec
all
s
co
r
e
o
f
0
.
9
1
a
m
o
n
g
all
class
if
ier
s
.
Fig
u
r
e
7
s
h
o
ws
th
e
F1
-
S
co
r
e.
Fro
m
f
ig
u
r
e
,
it
is
o
b
s
er
v
ed
th
at
tr
ad
itio
n
al
ML
class
if
ier
s
s
u
ch
as
R
F,
L
R
,
XG
B
,
AB
,
a
n
d
GB
ac
h
iev
e
F1
-
S
co
r
es
r
an
g
in
g
f
r
o
m
0
.
6
to
0
.
6
7
.
T
h
e
K
-
B
E
R
T
-
PC
cla
s
s
if
i
er
s
tan
d
s
o
u
t
with
th
e
h
ig
h
est
F1
-
S
co
r
e
o
f
0
.
9
3
am
o
n
g
all
class
if
ier
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
5
7
2
-
6
5
8
2
6580
Fig
u
r
e
4
.
Acc
u
r
ac
y
Fig
u
r
e
5
.
Pre
cisi
o
n
Fig
u
r
e
6
.
R
ec
all
Fig
u
r
e
7
.
F
-
Sco
r
e
5.
CO
NCLU
SI
O
N
I
n
c
o
n
clu
s
io
n
,
it
is
e
v
id
en
t
f
r
o
m
ex
is
tin
g
r
esear
ch
th
at
th
e
r
e
ex
is
ts
a
s
ca
r
city
o
f
lab
elle
d
d
atasets
s
p
ec
if
ically
d
esig
n
ed
f
o
r
Kan
n
ad
a
s
en
tim
en
t
an
al
y
s
is
.
Ad
d
itio
n
ally
,
th
er
e
h
as
b
ee
n
lim
i
ted
ex
p
lo
r
atio
n
i
n
ter
m
s
o
f
f
ea
tu
r
e
s
elec
tio
n
f
o
r
s
en
tim
en
t
an
aly
s
is
in
Kan
n
ad
a.
T
o
ad
d
r
ess
th
is
is
s
u
e,
th
e
p
r
esen
t
s
tu
d
y
in
tr
o
d
u
ce
s
a
n
o
v
el
Kan
n
ad
a
d
ataset
d
er
iv
ed
f
r
o
m
th
e
Sem
E
v
al
2
0
1
4
T
ask
4
d
ataset
th
r
o
u
g
h
Go
o
g
le
T
r
an
s
late.
Fu
r
th
er
m
o
r
e
,
th
is
wo
r
k
in
tr
o
d
u
ce
s
th
e
K
-
B
E
R
T
-
PC
clas
s
if
i
er
,
wh
ich
in
c
o
r
p
o
r
ates
a
m
o
d
if
ied
B
E
R
T
m
o
d
el
alo
n
g
with
a
p
r
o
b
a
b
ilit
y
cl
u
s
ter
in
g
ap
p
r
o
ac
h
.
T
h
e
e
x
p
er
im
en
tal
r
esu
lts
d
em
o
n
s
tr
at
e
th
at
K
-
B
E
R
T
-
P
C
ac
h
iev
es
s
u
p
er
io
r
p
e
r
f
o
r
m
an
ce
in
ter
m
s
o
f
p
o
lar
ity
a
n
d
s
en
tim
en
t
an
aly
s
is
ac
cu
r
ac
y
.
Sp
ec
if
ically
,
th
e
K
-
B
E
R
T
-
P
C
clas
s
if
ier
ac
h
iev
es
an
im
p
r
ess
iv
e
ac
cu
r
ac
y
r
at
e
o
f
9
1
%,
o
u
t
p
er
f
o
r
m
in
g
e
x
is
tin
g
class
if
ier
s
.
Fo
r
f
u
tu
r
e
r
esear
c
h
wo
r
k
,
th
is
wo
r
k
ca
n
b
e
ex
p
a
n
d
ed
to
in
clu
d
e
test
in
g
o
n
v
ar
i
o
u
s
o
th
er
Kan
n
ad
a
s
o
u
r
ce
s
s
u
ch
as
d
o
cu
m
e
n
ts
,
n
ewsp
ap
er
s
,
an
d
r
ev
iews
f
r
o
m
d
if
f
er
en
t
we
b
s
ites
,
th
er
eb
y
co
n
tr
ib
u
tin
g
to
f
u
r
th
er
ad
v
an
ce
m
en
ts
in
Kan
n
ad
a
s
en
tim
en
t a
n
aly
s
is
.
RE
F
E
R
E
NC
E
S
[
1
]
S
.
K
u
maw
a
t
,
I
.
Y
a
d
a
v
,
N
.
P
a
h
a
l
,
a
n
d
D
.
G
o
e
l
,
“
S
e
n
t
i
me
n
t
a
n
a
l
y
s
i
s
u
s
i
n
g
l
a
n
g
u
a
g
e
m
o
d
e
l
s
:
a
st
u
d
y
,
”
i
n
2
0
2
1
1
1
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
l
o
u
d
C
o
m
p
u
t
i
n
g
,
D
a
t
a
S
c
i
e
n
c
e
&
E
n
g
i
n
e
e
ri
n
g
(
C
o
n
f
l
u
e
n
c
e
)
,
Ja
n
.
2
0
2
1
,
p
p
.
9
8
4
–
9
8
8
,
d
o
i
:
1
0
.
1
1
0
9
/
C
o
n
f
l
u
e
n
c
e
5
1
6
4
8
.
2
0
2
1
.
9
3
7
7
0
4
3
.
[
2
]
M
.
R
o
d
r
í
g
u
e
z
-
I
b
á
n
e
z
,
A
.
C
a
sá
n
e
z
-
V
e
n
t
u
r
a
,
F
.
C
a
st
e
j
ó
n
-
M
a
t
e
o
s,
a
n
d
P
.
-
M
.
C
u
e
n
c
a
-
J
i
mé
n
e
z
,
“
A
r
e
v
i
e
w
o
n
s
e
n
t
i
m
e
n
t
a
n
a
l
y
si
s
f
r
o
m so
c
i
a
l
me
d
i
a
p
l
a
t
f
o
r
ms
,
”
Ex
p
e
r
t
S
y
s
t
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
2
2
3
,
A
u
g
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
2
3
.
1
1
9
8
6
2
.
[
3
]
S
.
El
z
e
h
e
i
r
y
,
W
.
A
.
G
a
b
-
A
l
l
a
h
,
N
.
M
e
k
k
y
,
a
n
d
M
.
E
l
m
o
g
y
,
“
S
e
n
t
i
me
n
t
a
n
a
l
y
si
s
f
o
r
e
-
c
o
mm
e
r
c
e
p
r
o
d
u
c
t
r
e
v
i
e
w
s:
c
u
r
r
e
n
t
t
r
e
n
d
s
a
n
d
f
u
t
u
r
e
d
i
r
e
c
t
i
o
n
s
,
”
Pre
p
ri
n
t
s
,
2
0
2
3
,
2
0
2
3
0
5
1
6
4
9
,
d
o
i
:
1
0
.
2
0
9
4
4
/
p
r
e
p
r
i
n
t
s2
0
2
3
0
5
.
1
6
4
9
.
v
1
.
[
4
]
F
.
A
r
i
a
s
,
M
.
Z
.
N
u
n
e
z
,
A
.
G
u
e
r
r
a
-
A
d
a
m
e
s,
N
.
Te
j
e
d
o
r
-
F
l
o
r
e
s
,
a
n
d
M
.
V
a
r
g
a
s
-
L
o
m
b
a
r
d
o
,
“
S
e
n
t
i
me
n
t
a
n
a
l
y
si
s
o
f
p
u
b
l
i
c
so
c
i
a
l
med
i
a
a
s
a
t
o
o
l
f
o
r
h
e
a
l
t
h
-
r
e
l
a
t
e
d
t
o
p
i
c
s,”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
0
,
p
p
.
7
4
8
5
0
–
7
4
8
7
2
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
2
.
3
1
8
7
4
0
6
.
[
5
]
M
.
W
o
l
l
mer
e
t
a
l
.
,
“
Y
o
u
T
u
b
e
m
o
v
i
e
r
e
v
i
e
w
s:
s
e
n
t
i
me
n
t
a
n
a
l
y
si
s
i
n
a
n
a
u
d
i
o
-
v
i
su
a
l
c
o
n
t
e
x
t
,
”
I
EE
E
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s
,
v
o
l
.
2
8
,
0
0
.2
0
.
4
0
.
6
0
.
8
1
A
c
c
u
r
a
c
y
C
l
a
ssi
f
i
e
r
Ac
c
u
r
ac
y
Rand
om
F
or
e
s
t
L
ogis
t
ic
Re
gr
e
s
s
i
on
XGB
oos
t
AdaB
oos
t
Gradi
e
n
t
B
oos
t
Vot
in
g
Clas
s
if
ie
r
B
E
RT
K-
B
E
RT
-
P
C
0
0
.
2
0
.
4
0
.
6
0
.
8
1
Pr
e
c
i
si
o
n
C
l
a
ssi
f
i
e
r
P
r
e
c
i
si
on
Rand
om
F
or
e
s
t
L
ogis
t
ic
Re
gr
e
s
s
i
on
XG
Bo
os
t
AdaB
oos
t
Gradi
e
n
t
B
oos
t
Vot
in
g
Cl
as
s
if
ie
r
B
E
RT
K-
B
E
RT
-
P
C
0
0
.
2
0
.
4
0
.
6
0
.
8
1
R
e
c
a
l
l
C
l
a
ss
i
f
i
er
Re
c
al
l
Rand
om
F
or
e
s
t
L
ogis
t
ic
Re
gr
e
s
s
i
on
XGB
oos
t
AdaB
oos
t
Gradi
e
n
t
B
oos
t
Vot
in
g
Clas
s
if
ie
r
B
E
RT
K-
B
E
RT
-
P
C
0
0
.
2
0
.
4
0
.
6
0
.
8
1
F
-
S
c
o
r
e
C
l
a
ss
i
f
i
er
F
-
S
c
or
e
Rand
om
F
or
e
s
t
L
ogis
t
ic
Re
gr
e
s
s
i
on
XGB
oos
t
AdaB
oos
t
Gradi
e
n
t
B
oos
t
Vot
in
g
Clas
s
if
ie
r
B
E
RT
K-
B
E
RT
-
P
C
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
E
n
h
a
n
cin
g
s
en
timen
t a
n
a
lysi
s
in
K
a
n
n
a
d
a
texts b
y
fea
tu
r
e
s
elec
tio
n
(
S
u
n
il Mu
g
a
lih
a
lli E
s
h
w
a
r
a
p
p
a
)
6581
n
o
.
3
,
p
p
.
4
6
–
5
3
,
M
a
y
2
0
1
3
,
d
o
i
:
1
0
.
1
1
0
9
/
M
I
S
.
2
0
1
3
.
3
4
.
[
6
]
J.
S
h
a
y
n
n
-
Ly
K
w
a
n
a
n
d
K
.
H
u
i
Li
m,
“
U
n
d
e
r
s
t
a
n
d
i
n
g
p
u
b
l
i
c
s
e
n
t
i
m
e
n
t
s,
o
p
i
n
i
o
n
s
a
n
d
t
o
p
i
c
s
a
b
o
u
t
C
O
V
I
D
-
1
9
u
si
n
g
Tw
i
t
t
e
r
,
”
i
n
2
0
2
0
I
E
EE
/
AC
M
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
A
d
v
a
n
c
e
s
i
n
S
o
c
i
a
l
N
e
t
w
o
r
k
s
An
a
l
y
si
s
a
n
d
Mi
n
i
n
g
(
A
S
O
N
AM
)
,
D
e
c
.
2
0
2
0
,
p
p
.
6
2
3
–
6
2
6
,
d
o
i
:
1
0
.
1
1
0
9
/
A
S
O
N
A
M
4
9
7
8
1
.
2
0
2
0
.
9
3
8
1
3
8
4
.
[
7
]
K
.
R
a
k
s
h
i
t
h
a
,
R
.
H
M
,
M
.
P
a
v
i
t
h
r
a
,
A
.
H
D
,
a
n
d
M
.
H
e
g
d
e
,
“
S
e
n
t
i
m
e
n
t
a
l
a
n
a
l
y
si
s
o
f
I
n
d
i
a
n
r
e
g
i
o
n
a
l
l
a
n
g
u
a
g
e
s
o
n
so
c
i
a
l
me
d
i
a
,
”
G
l
o
b
a
l
T
r
a
n
s
i
t
i
o
n
s
Pr
o
c
e
e
d
i
n
g
s
,
v
o
l
.
2
,
n
o
.
2
,
p
p
.
4
1
4
–
4
2
0
,
N
o
v
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
g
l
t
p
.
2
0
2
1
.
0
8
.
0
3
9
.
[
8
]
J.
R
.
Ji
m
,
M
.
A
.
R
.
Ta
l
u
k
d
e
r
,
P
.
M
a
l
a
k
a
r
,
M
.
M
.
K
a
b
i
r
,
K
.
N
u
r
,
a
n
d
M
.
F
.
M
r
i
d
h
a
,
“
R
e
c
e
n
t
a
d
v
a
n
c
e
me
n
t
s
a
n
d
c
h
a
l
l
e
n
g
e
s
o
f
N
LP
-
b
a
s
e
d
s
e
n
t
i
m
e
n
t
a
n
a
l
y
si
s
:
a
st
a
t
e
-
of
-
t
h
e
-
a
r
t
r
e
v
i
e
w
,
”
N
a
t
u
r
a
l
L
a
n
g
u
a
g
e
Pr
o
c
e
ssi
n
g
J
o
u
r
n
a
l
,
v
o
l
.
6
,
M
a
r
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
n
l
p
.
2
0
2
4
.
1
0
0
0
5
9
.
[
9
]
R
.
K
.
D
a
s,
M
.
I
sl
a
m,
M
.
M
.
H
a
sa
n
,
S
.
R
a
z
i
a
,
M
.
H
a
s
sa
n
,
a
n
d
S
.
A
.
K
h
u
s
h
b
u
,
“
S
e
n
t
i
m
e
n
t
a
n
a
l
y
si
s
i
n
m
u
l
t
i
l
i
n
g
u
a
l
c
o
n
t
e
x
t
:
c
o
m
p
a
r
a
t
i
v
e
a
n
a
l
y
si
s
o
f
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
n
d
h
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
,
”
H
e
l
i
y
o
n
,
v
o
l
.
9
,
n
o
.
9
,
S
e
p
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
h
e
l
i
y
o
n
.
2
0
2
3
.
e
2
0
2
8
1
.
[
1
0
]
D
.
D
r
a
s
k
o
v
i
c
,
D
.
Z
e
c
e
v
i
c
,
a
n
d
B
.
N
i
k
o
l
i
c
,
“
D
e
v
e
l
o
p
me
n
t
o
f
a
mu
l
t
i
l
i
n
g
u
a
l
m
o
d
e
l
f
o
r
mac
h
i
n
e
se
n
t
i
me
n
t
a
n
a
l
y
si
s
i
n
t
h
e
S
e
r
b
i
a
n
l
a
n
g
u
a
g
e
,
”
M
a
t
h
e
m
a
t
i
c
s
,
v
o
l
.
1
0
,
n
o
.
1
8
,
S
e
p
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
mat
h
1
0
1
8
3
2
3
6
.
[
1
1
]
M
.
M
.
A
g
ü
e
r
o
-
To
r
a
l
e
s,
J.
I
.
A
b
r
e
u
S
a
l
a
s,
a
n
d
A
.
G
.
Ló
p
e
z
-
H
e
r
r
e
r
a
,
“
D
e
e
p
l
e
a
r
n
i
n
g
a
n
d
m
u
l
t
i
l
i
n
g
u
a
l
se
n
t
i
me
n
t
a
n
a
l
y
s
i
s
o
n
so
c
i
a
l
med
i
a
d
a
t
a
:
a
n
o
v
e
r
v
i
e
w
,
”
A
p
p
l
i
e
d
S
o
f
t
C
o
m
p
u
t
i
n
g
,
v
o
l
.
1
0
7
,
A
u
g
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
s
o
c
.
2
0
2
1
.
1
0
7
3
7
3
.
[
1
2
]
P
.
D
o
m
i
n
i
c
,
N
.
P
u
r
u
sh
o
t
h
a
ma
n
,
A
.
S
.
A
.
K
u
mar,
A
.
P
r
a
b
a
g
a
r
a
n
,
J
.
A
n
g
e
l
i
n
B
l
e
ss
y
,
a
n
d
J.
A
,
“
M
u
l
t
i
l
i
n
g
u
a
l
s
e
n
t
i
me
n
t
a
n
a
l
y
s
i
s
u
si
n
g
d
e
e
p
-
l
e
a
r
n
i
n
g
a
r
c
h
i
t
e
c
t
u
r
e
s,
”
i
n
2
0
2
3
5
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
S
m
a
rt
S
y
s
t
e
m
s
a
n
d
I
n
v
e
n
t
i
v
e
T
e
c
h
n
o
l
o
g
y
(
I
C
S
S
I
T
)
,
Jan
.
2
0
2
3
,
p
p
.
1
0
7
7
–
1
0
8
3
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
S
S
I
T5
5
8
1
4
.
2
0
2
3
.
1
0
0
6
0
9
9
3
.
[
1
3
]
B
.
R
.
C
h
a
k
r
a
v
a
r
t
h
i
e
t
a
l
.
,
“
F
i
n
d
i
n
g
s
o
f
t
h
e
se
n
t
i
m
e
n
t
a
n
a
l
y
si
s
o
f
D
r
a
v
i
d
i
a
n
l
a
n
g
u
a
g
e
s
i
n
c
o
d
e
-
m
i
x
e
d
t
e
x
t
,
”
a
r
Xi
v
p
re
p
ri
n
t
a
rXi
v
:
2
1
1
1
.
0
9
8
1
1
,
2
0
2
1
,
d
o
i
:
1
0
.
4
8
5
5
0
/
a
r
x
i
v
.
2
1
1
1
.
0
9
8
1
1
.
[
1
4
]
A
.
H
a
n
d
e
,
R
.
P
r
i
y
a
d
h
a
r
sh
i
n
i
,
a
n
d
B
.
R
.
C
h
a
k
r
a
v
a
r
t
h
i
,
“
K
a
n
C
M
D
:
K
a
n
n
a
d
a
C
o
d
e
M
i
x
e
d
d
a
t
a
set
f
o
r
s
e
n
t
i
men
t
a
n
a
l
y
s
i
s
a
n
d
o
f
f
e
n
s
i
v
e
l
a
n
g
u
a
g
e
d
e
t
e
c
t
i
o
n
,
”
i
n
P
ro
c
e
e
d
i
n
g
s
o
f
t
h
e
T
h
i
r
d
Wo
r
k
s
h
o
p
o
n
C
o
m
p
u
t
a
t
i
o
n
a
l
M
o
d
e
l
i
n
g
o
f
P
e
o
p
l
e
’
s
O
p
i
n
i
o
n
s
,
Pe
rso
n
a
l
i
t
y
,
a
n
d
Em
o
t
i
o
n
’
s i
n
S
o
c
i
a
l
Me
d
i
a
,
2
0
2
0
,
p
p
.
5
4
–
6
3
.
[
1
5
]
J.
H
a
r
t
ma
n
n
,
M
.
H
e
i
t
m
a
n
n
,
C
.
S
i
e
b
e
r
t
,
a
n
d
C
.
S
c
h
a
mp
,
“
M
o
r
e
t
h
a
n
a
f
e
e
l
i
n
g
:
a
c
c
u
r
a
c
y
a
n
d
a
p
p
l
i
c
a
t
i
o
n
o
f
s
e
n
t
i
m
e
n
t
a
n
a
l
y
si
s
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
R
e
se
a
rc
h
i
n
Ma
r
k
e
t
i
n
g
,
v
o
l
.
4
0
,
n
o
.
1
,
p
p
.
7
5
–
8
7
,
M
a
r
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
j
r
e
sm
a
r
.
2
0
2
2
.
0
5
.
0
0
5
.
[
1
6
]
I
.
D
i
a
s,
R
.
R
e
i
,
P
.
P
e
r
e
i
r
a
,
a
n
d
L
.
C
o
h
e
u
r
,
“
T
o
w
a
r
d
s
a
s
e
n
t
i
me
n
t
-
a
w
a
r
e
c
o
n
v
e
r
sat
i
o
n
a
l
a
g
e
n
t
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
2
n
d
AC
M
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
I
n
t
e
l
l
i
g
e
n
t
Vi
r
t
u
a
l
A
g
e
n
t
s
,
S
e
p
.
2
0
2
2
,
p
p
.
1
–
3
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
1
4
1
9
7
.
3
5
4
9
6
9
2
.
[
1
7
]
A
.
R
.
A
p
p
i
d
i
,
V
.
K
.
S
r
i
r
a
n
g
a
m,
D
.
S
u
h
a
s,
a
n
d
M
.
S
h
r
i
v
a
s
t
a
v
a
,
“
C
r
e
a
t
i
o
n
o
f
c
o
r
p
u
s
a
n
d
a
n
a
l
y
s
i
s
i
n
c
o
d
e
-
m
i
x
e
d
K
a
n
n
a
d
a
-
E
n
g
l
i
sh
Tw
i
t
t
e
r
d
a
t
a
f
o
r
e
m
o
t
i
o
n
p
r
e
d
i
c
t
i
o
n
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
8
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
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
2
0
,
p
p
.
6
7
0
3
–
6
7
0
9
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
2
0
2
0
.
c
o
l
i
n
g
-
m
a
i
n
.
5
8
7
.
[
1
8
]
C
.
P
.
C
h
a
n
d
r
i
k
a
,
J
.
S
.
K
a
l
l
i
ma
n
i
,
H
.
P
.
A
d
a
r
s
h
a
,
A
.
N
a
g
a
b
h
u
sh
a
n
,
a
n
d
A
.
C
h
a
v
a
n
,
“
R
u
l
e
-
b
a
se
d
a
p
p
r
o
a
c
h
f
o
r
e
mo
t
i
o
n
d
e
t
e
c
t
i
o
n
f
o
r
K
a
n
n
a
d
a
t
e
x
t
,
”
2
0
2
1
,
p
p
.
4
6
3
–
4
7
2
.
[
1
9
]
P
.
R
a
n
j
i
t
h
a
a
n
d
K
.
N
.
B
h
a
n
u
,
“
I
mp
r
o
v
e
d
se
n
t
i
me
n
t
a
n
a
l
y
s
i
s
f
o
r
D
r
a
v
i
d
i
a
n
l
a
n
g
u
a
g
e
-
K
a
n
n
a
d
a
u
si
n
g
d
e
c
i
si
o
n
t
r
e
e
a
l
g
o
r
i
t
h
m
w
i
t
h
e
f
f
i
c
i
e
n
t
d
a
t
a
d
i
c
t
i
o
n
a
r
y
,
”
I
O
P
C
o
n
f
e
re
n
c
e
S
e
ri
e
s:
Ma
t
e
r
i
a
l
s
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
1
2
3
,
n
o
.
1
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
8
8
/
1
7
5
7
-
8
9
9
X
/
1
1
2
3
/
1
/
0
1
2
0
3
9
.
[
2
0
]
M
.
E.
S
u
n
i
l
a
n
d
S
.
V
i
n
a
y
,
“
K
a
n
n
a
d
a
sen
t
i
m
e
n
t
a
n
a
l
y
si
s
u
s
i
n
g
v
e
c
t
o
r
i
z
a
t
i
o
n
a
n
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
i
n
S
e
n
t
i
m
e
n
t
a
l
A
n
a
l
y
s
i
s
a
n
d
D
e
e
p
L
e
a
rn
i
n
g
:
Pro
c
e
e
d
i
n
g
s
o
f
I
C
S
A
D
L
2
0
2
1
,
2
0
2
2
,
p
p
.
6
7
7
–
6
8
9
.
[
2
1
]
S
.
S
h
e
t
t
y
e
t
a
l
.
,
“
S
e
n
t
i
me
n
t
a
n
a
l
y
s
i
s
o
f
Tw
i
t
t
e
r
p
o
st
s
i
n
En
g
l
i
s
h
,
K
a
n
n
a
d
a
a
n
d
H
i
n
d
i
l
a
n
g
u
a
g
e
s,”
i
n
Re
c
e
n
t
Ad
v
a
n
c
e
s
i
n
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
D
a
t
a
E
n
g
i
n
e
e
ri
n
g
:
S
e
l
e
c
t
Pr
o
c
e
e
d
i
n
g
s
o
f
AI
D
E
2
0
2
0
,
2
0
2
2
,
p
p
.
3
6
1
–
3
7
5
.
[
2
2
]
B
.
R
.
C
h
a
k
r
a
v
a
r
t
h
i
e
t
a
l
.
,
“
D
r
a
v
i
d
i
a
n
C
o
d
e
M
i
x
:
se
n
t
i
me
n
t
a
n
a
l
y
si
s
a
n
d
o
f
f
e
n
si
v
e
l
a
n
g
u
a
g
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
d
a
t
a
set
f
o
r
D
r
a
v
i
d
i
a
n
l
a
n
g
u
a
g
e
s
i
n
c
o
d
e
-
m
i
x
e
d
t
e
x
t
,
”
L
a
n
g
u
a
g
e
Re
s
o
u
rc
e
s
a
n
d
E
v
a
l
u
a
t
i
o
n
,
v
o
l
.
5
6
,
n
o
.
3
,
p
p
.
7
6
5
–
8
0
6
,
S
e
p
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
5
7
9
-
0
2
2
-
0
9
5
8
3
-
7.
[
2
3
]
S
h
a
n
k
a
r
R
a
n
d
S
.
S
w
a
m
y
,
“
H
y
b
r
i
d
s
e
n
t
i
m
e
n
t
a
n
a
l
y
s
i
s:
a
n
o
v
e
l
i
n
t
e
g
r
a
t
i
o
n
o
f
c
o
r
p
u
s
-
d
r
i
v
e
n
a
n
d
mac
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
e
s
t
o
p
e
r
f
o
r
m
se
n
t
i
me
n
t
a
n
a
l
y
s
i
s
o
f
K
a
n
n
a
d
a
P
o
l
i
t
i
c
a
l
Tw
e
e
t
s,
”
T
u
i
j
i
n
J
i
sh
u
/
J
o
u
r
n
a
l
o
f
Pro
p
u
l
s
i
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
4
4
,
n
o
.
5
,
p
p
.
3
2
4
6
–
3
2
5
9
,
2
0
2
3
.
[
2
4
]
S
h
a
n
k
a
r
R
,
S
.
S
w
a
my
,
a
n
d
S
.
H
e
g
d
e
,
“
E
x
p
l
o
r
i
n
g
se
n
t
i
me
n
t
a
n
a
l
y
si
s
i
n
K
a
n
n
a
d
a
l
a
n
g
u
a
g
e
:
a
c
o
mp
r
e
h
e
n
s
i
v
e
st
u
d
y
o
n
C
O
V
I
D
-
1
9
d
a
t
a
u
si
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
n
d
e
n
sem
b
l
e
a
l
g
o
r
i
t
h
ms
,
”
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
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
2
,
n
o
.
1
1
,
p
p
.
2
1
–
2
9
,
2
0
2
4
.
[
2
5
]
K
.
S
r
e
e
l
a
k
sh
mi
,
B
.
P
r
e
m
j
i
t
h
,
B
.
R
.
C
h
a
k
r
a
v
a
r
t
h
i
,
a
n
d
K
.
P
.
S
o
m
a
n
,
“
D
e
t
e
c
t
i
o
n
o
f
h
a
t
e
s
p
e
e
c
h
a
n
d
o
f
f
e
n
si
v
e
l
a
n
g
u
a
g
e
C
o
d
e
M
i
x
t
e
x
t
i
n
D
r
a
v
i
d
i
a
n
l
a
n
g
u
a
g
e
s
u
s
i
n
g
c
o
s
t
-
se
n
si
t
i
v
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
I
EEE
A
c
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
2
0
0
6
4
–
2
0
0
9
0
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
3
5
8
8
1
1
.
[
2
6
]
M
.
P
o
n
t
i
k
i
,
D
.
G
a
l
a
n
i
s,
J.
P
a
v
l
o
p
o
u
l
o
s,
H
.
P
a
p
a
g
e
o
r
g
i
o
u
,
I
.
A
n
d
r
o
u
t
so
p
o
u
l
o
s,
a
n
d
S
.
M
a
n
a
n
d
h
a
r
,
“
S
e
mE
v
a
l
-
2
0
1
4
Ta
sk
4
:
a
s
p
e
c
t
b
a
s
e
d
s
e
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
8
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
Wo
r
k
s
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
E
v
a
l
2
0
1
4
)
,
2
0
1
4
,
p
p
.
2
7
–
3
5
,
d
o
i
:
1
0
.
3
1
1
5
/
v
1
/
S
1
4
-
2
0
0
4
.
[
2
7
]
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
,
”
a
rX
i
v
p
r
e
p
r
i
n
t
a
rX
i
v
:
1
8
1
0
.
0
4
8
0
5
.
,
2
0
1
9
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
N
1
9
-
1
4
2
3
.
[
2
8
]
T.
B
r
a
n
t
s,
“
T
n
T
-
a
st
a
t
i
st
i
c
a
l
p
a
r
t
-
of
-
sp
e
e
c
h
t
a
g
g
e
r
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
s
i
x
t
h
c
o
n
f
e
r
e
n
c
e
o
n
A
p
p
l
i
e
d
n
a
t
u
r
a
l
l
a
n
g
u
a
g
e
p
ro
c
e
ss
i
n
g
-
,
2
0
0
0
,
p
p
.
2
2
4
–
2
3
1
,
d
o
i
:
1
0
.
3
1
1
5
/
9
7
4
1
4
7
.
9
7
4
1
7
8
.
[
2
9
]
T.
H
a
r
i
y
a
n
t
i
,
S
.
A
i
d
a
,
a
n
d
H
.
K
a
me
d
a
,
“
S
a
m
a
w
a
l
a
n
g
u
a
g
e
p
a
r
t
o
f
sp
e
e
c
h
t
a
g
g
i
n
g
w
i
t
h
p
r
o
b
a
b
i
l
i
st
i
c
a
p
p
r
o
a
c
h
:
c
o
m
p
a
r
i
so
n
o
f
U
n
i
g
r
a
m,
H
M
M
a
n
d
T
n
T
m
o
d
e
l
s,”
J
o
u
rn
a
l
o
f
Ph
y
si
c
s:
C
o
n
f
e
re
n
c
e
S
e
r
i
e
s
,
v
o
l
.
1
2
3
5
,
n
o
.
1
,
Ju
n
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
8
8
/
1
7
4
2
-
6
5
9
6
/
1
2
3
5
/
1
/
0
1
2
0
1
3
.
[
3
0
]
B
.
O
z
y
u
r
t
a
n
d
M
.
A
.
A
k
c
a
y
o
l
,
“
A
n
e
w
t
o
p
i
c
m
o
d
e
l
i
n
g
b
a
s
e
d
a
p
p
r
o
a
c
h
f
o
r
a
sp
e
c
t
e
x
t
r
a
c
t
i
o
n
i
n
a
s
p
e
c
t
b
a
s
e
d
se
n
t
i
m
e
n
t
a
n
a
l
y
s
i
s:
SS
-
LD
A
,
”
Ex
p
e
r
t
S
y
s
t
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
6
8
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
2
0
.
1
1
4
2
3
1
.
[
3
1
]
M
.
V
e
n
u
g
o
p
a
l
a
n
a
n
d
D
.
G
u
p
t
a
,
“
A
n
e
n
h
a
n
c
e
d
g
u
i
d
e
d
LD
A
m
o
d
e
l
a
u
g
me
n
t
e
d
w
i
t
h
B
ER
T
b
a
s
e
d
sem
a
n
t
i
c
st
r
e
n
g
t
h
f
o
r
a
sp
e
c
t
t
e
r
m
e
x
t
r
a
c
t
i
o
n
i
n
s
e
n
t
i
m
e
n
t
a
n
a
l
y
si
s,”
K
n
o
w
l
e
d
g
e
-
B
a
se
d
S
y
s
t
e
m
s
,
v
o
l
.
2
4
6
,
J
u
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
k
n
o
sy
s
.
2
0
2
2
.
1
0
8
6
6
8
.
[
3
2
]
S
.
S
y
e
d
a
n
d
M
.
S
p
r
u
i
t
,
“
F
u
l
l
-
t
e
x
t
o
r
a
b
s
t
r
a
c
t
?
e
x
a
mi
n
i
n
g
t
o
p
i
c
c
o
h
e
r
e
n
c
e
s
c
o
r
e
s
u
si
n
g
l
a
t
e
n
t
D
i
r
i
c
h
l
e
t
a
l
l
o
c
a
t
i
o
n
,
”
i
n
2
0
1
7
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
D
a
t
a
S
c
i
e
n
c
e
a
n
d
A
d
v
a
n
c
e
d
An
a
l
y
t
i
c
s
(
D
S
A
A)
,
O
c
t
.
2
0
1
7
,
p
p
.
1
6
5
–
1
7
4
,
d
o
i
:
1
0
.
1
1
0
9
/
D
S
A
A
.
2
0
1
7
.
6
1
.
[
3
3
]
K
.
S
.
A
r
u
n
a
n
d
V
.
K
.
G
o
v
i
n
d
a
n
,
“
A
h
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
l
a
t
e
n
t
t
o
p
i
c
-
b
a
se
d
i
ma
g
e
r
e
t
r
i
e
v
a
l
,
”
D
a
t
a
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
3
,
n
o
.
2
,
p
p
.
1
6
6
–
1
9
5
,
J
u
n
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
0
7
/
s
4
1
0
1
9
-
018
-
0
0
6
3
-
7.
Evaluation Warning : The document was created with Spire.PDF for Python.