I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
10
,
No
.
4
,
A
u
g
u
s
t
2020
,
p
p
.
3
7
5
1
~
3
7
5
6
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
10
i
4
.
p
p
3
7
5
1
-
3
7
5
6
3751
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m/in
d
ex
.
p
h
p
/I
JE
C
E
Predic
ting dep
res
sio
n using
deep
l
e
a
rning
and ense
mble
a
lg
o
rith
m
s o
n
ra
w
t
w
itt
er data
Nis
ha
P
.
Sh
et
t
y
,
B
a
la
cha
nd
ra
M
un
iy
a
l,
Arshi
a
Ana
nd
,
S
us
ha
nt
K
u
m
a
r,
Su
s
ha
nt
P
ra
bh
u
De
p
a
rtme
n
t
o
f
In
f
o
rm
a
ti
o
n
a
n
d
C
o
m
m
u
n
ica
ti
o
n
T
e
c
h
n
o
l
o
g
y
,
M
a
n
ip
a
l
In
stit
u
te
o
f
T
e
c
h
n
o
l
o
g
y
,
M
a
n
ip
a
l
A
c
a
d
e
m
y
o
f
Hi
g
h
e
r
Ed
u
c
a
ti
o
n
,
I
n
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
1
1
,
2
0
1
9
R
ev
i
s
ed
J
an
1
3
,
2
0
2
0
A
cc
ep
ted
Feb
2
,
2
0
2
0
S
o
c
ial
n
e
tw
o
rk
a
n
d
m
i
c
ro
b
lo
g
g
in
g
sites
su
c
h
a
s
Tw
it
ter
a
re
w
id
e
sp
re
a
d
a
m
o
n
g
st
a
ll
g
e
n
e
ra
ti
o
n
s
n
o
w
a
d
a
y
s
w
h
e
re
p
e
o
p
le
c
o
n
n
e
c
t
a
n
d
sh
a
re
th
e
ir
f
e
e
li
n
g
s,
e
m
o
ti
o
n
s,
p
u
rsu
it
s
e
tc.
De
p
re
ss
io
n
,
o
n
e
o
f
th
e
m
o
st
c
o
m
m
o
n
m
e
n
tal
d
iso
rd
e
r,
is
a
n
a
c
u
te
sta
te
o
f
s
a
d
n
e
ss
w
h
e
r
e
p
e
rso
n
lo
se
s
in
tere
st
in
a
ll
a
c
ti
v
it
ies
.
I
f
n
o
t
trea
ted
i
m
m
e
d
iate
l
y
th
is ca
n
r
e
su
lt
in
d
ire co
n
se
q
u
e
n
c
e
s su
c
h
a
s
d
e
a
th
.
In
th
is
e
ra
o
f
v
irt
u
a
l
w
o
rld
,
p
e
o
p
le
a
re
m
o
re
c
o
m
fo
rtab
le
in
e
x
p
re
ss
in
g
th
e
ir
e
m
o
ti
o
n
s
in
su
c
h
sites
a
s
th
e
y
h
a
v
e
b
e
c
o
m
e
a
p
a
rt
a
n
d
p
a
rc
e
l
o
f
e
v
e
r
y
d
a
y
li
v
e
s.
T
h
e
re
se
a
rc
h
p
u
t
f
o
rth
t
h
u
s,
e
m
p
lo
y
s
m
a
c
h
in
e
lea
rn
in
g
c
las
si
f
iers
o
n
th
e
tw
it
ter
d
a
ta
s
e
t
to
d
e
tec
t
if
a
p
e
rso
n
’s
t
w
e
e
t
in
d
ica
tes
a
n
y
sig
n
o
f
d
e
p
re
ss
io
n
o
r
n
o
t
.
K
ey
w
o
r
d
s
:
Dep
r
ess
io
n
L
ST
M
Ma
ch
i
n
e
lear
n
i
n
g
Sen
ti
m
e
n
t a
n
a
l
y
s
is
So
cial
m
ed
ia
Co
p
y
rig
h
t
©
2
0
2
0
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
B
alac
h
an
d
r
a
Mu
n
i
y
al,
Dep
ar
t
m
en
t o
f
I
n
f
o
r
m
atio
n
an
d
C
o
m
m
u
n
icatio
n
T
ec
h
n
o
lo
g
y
,
Ma
n
ip
al
I
n
s
tit
u
te
o
f
T
ec
h
n
o
lo
g
y
,
Ma
n
ip
al
A
ca
d
e
m
y
o
f
Hi
g
h
er
E
d
u
ca
tio
n
,
Ma
n
ip
al
-
5
7
6
1
0
4
I
n
d
ia.
E
m
ail: b
ala.
ch
a
n
d
r
a@
m
a
n
ip
al
.
ed
u
1.
I
NT
RO
D
UCT
I
O
N
So
cial
n
et
w
o
r
k
i
n
g
s
ite
s
h
a
v
e
b
ec
o
m
e
a
h
ab
itu
al
co
m
p
o
n
en
t,
w
it
h
s
ites
li
k
e
T
w
it
ter
an
d
Face
b
o
o
k
b
ein
g
t
h
e
7
th
a
n
d
2
nd
f
a
v
o
r
ite
s
ites
h
av
i
n
g
m
illi
o
n
s
o
f
s
u
b
s
cr
i
b
er
s
[
1
]
.
Su
ch
s
ites
h
a
v
e
b
ec
o
m
e
a
n
o
p
en
d
ais
f
o
r
p
eo
p
le
t
o
r
ea
ch
o
u
t
an
d
ex
p
r
ess
th
eir
f
ee
lin
g
s
,
li
k
es,
r
o
u
ti
n
es
etc.
Un
li
k
e
m
o
o
d
f
lu
c
tu
a
tio
n
s
,
d
ep
r
ess
io
n
is
a
co
m
m
o
n
m
e
n
tal
d
i
s
o
r
d
er
w
h
ic
h
b
r
u
tall
y
a
f
f
ec
t
s
a
p
er
s
o
n
’
s
d
ail
y
r
o
u
ti
n
e
li
f
e.
A
n
y
p
er
s
o
n
w
h
o
h
a
s
u
n
d
er
g
o
n
e
s
o
m
e
ad
v
er
s
e
e
x
p
er
ien
ce
s
li
k
e
s
u
d
d
en
d
ea
th
,
u
n
e
m
p
lo
y
m
e
n
t
etc.
is
l
iab
le
to
it.
A
cc
o
r
d
in
g
to
W
HO,
m
o
r
e
t
h
an
3
0
0
m
illi
o
n
p
eo
p
le
o
f
all
a
g
e
g
r
o
u
p
s
s
u
f
f
er
f
r
o
m
it,
w
it
h
o
n
l
y
f
e
w
er
t
h
an
1
0
%
r
ec
ei
v
i
n
g
s
u
itab
le
tr
ea
t
m
en
t,
o
w
i
n
g
to
r
ea
s
o
n
s
s
u
c
h
as
lack
o
f
s
u
itab
l
e
h
ea
lth
ca
r
e,
s
o
cial
h
u
m
iliati
o
n
an
d
ti
m
el
y
an
d
r
ig
h
t
d
ia
g
n
o
s
is
[
2
]
.
Fra
n
ce
s
A
et
al
.
[
3
]
,
s
tate
s
t
h
at
th
e
d
ep
r
ess
io
n
d
is
p
la
y
s
t
h
e
f
o
llo
w
i
n
g
s
y
m
p
to
m
s
i
n
t
h
e
g
iv
en
o
r
d
er
;
s
ad
m
o
o
d
,
elu
d
i
n
g
all
ac
ti
v
iti
es,
w
ei
g
h
t
an
d
s
leep
f
lu
ct
u
ati
o
n
s
,
b
o
d
y
ag
itatio
n
,
e
n
er
g
y
l
o
s
s
an
d
tire
d
n
es
s
,
f
ee
li
n
g
o
f
tr
i
v
ialit
y
,
lo
s
s
o
f
d
e
cisi
o
n
-
m
ak
i
n
g
ca
p
ac
it
y
an
d
f
i
n
all
y
s
u
icid
al
ten
d
en
cie
s
.
E
m
o
tio
n
an
d
s
en
ti
m
e
n
t
an
al
y
s
is
e
x
er
cise
s
m
ac
h
i
n
e
l
ea
r
n
in
g
al
g
o
r
ith
m
s
to
e
x
a
m
i
n
e
a
tex
t
w
it
h
r
esp
ec
t
to
th
e
e
m
o
tio
n
co
n
v
e
y
ed
.
Sen
te
n
ce
lev
e
l a
n
al
y
s
is
is
ap
p
l
ied
in
th
i
s
s
t
u
d
y
to
i
n
s
p
ec
t i
f
a
t
w
ee
t i
s
e
m
o
t
io
n
all
y
v
u
l
n
er
ab
le
o
r
n
o
t.
R
ec
e
n
tl
y
,
th
e
d
ea
th
o
f
a
1
6
-
y
ea
r
-
o
ld
M
ala
y
s
ia
n
teen
a
f
ter
ca
llin
g
f
o
r
a
p
o
ll
in
h
er
I
n
s
ta
g
r
a
m
[
4
]
(
w
h
er
e
m
a
n
y
p
eo
p
le
v
o
ted
f
o
r
“
Dea
t
h
”)
h
as
g
ath
er
ed
h
u
g
e
m
ed
ia.
Su
c
h
in
c
id
en
ts
s
u
p
p
o
r
t
th
e
f
ac
t
th
at
i
f
w
it
h
p
r
o
p
er
m
o
n
ito
r
in
g
o
f
s
u
c
h
s
ite
s
[
5
]
is
d
o
n
e
ti
m
el
y
h
el
p
an
d
ca
n
b
e
p
r
o
v
id
ed
,
h
en
ce
av
o
id
in
g
s
u
c
h
ca
tas
tr
o
p
h
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
4
,
A
u
g
u
s
t
2020
:
3
7
5
1
-
375
6
3752
2.
L
I
T
E
R
AT
U
RE
SU
RVE
Y
Haq
u
e
et
al.
[
7
]
s
cr
u
ti
n
ized
3
D
f
ac
ial
f
ea
t
u
r
es
an
d
lan
g
u
a
g
e
s
p
o
k
e
n
to
g
a
g
e
t
h
e
i
n
ten
s
it
y
o
f
d
ep
r
ess
io
n
.
T
h
e
y
co
m
p
ar
ed
t
h
eir
s
e
n
te
n
ce
le
v
el
e
m
b
ed
d
ed
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
et
w
o
r
k
(
C
NN
)
m
o
d
el
[
8
]
w
it
h
t
h
e
ex
is
tin
g
w
o
r
k
s
,
b
u
t
t
h
e
d
ata
w
a
s
co
llected
b
y
h
u
m
a
n
co
m
p
u
ter
in
ter
v
ie
w
s
w
h
ic
h
l
ac
k
ed
th
e
p
r
ec
is
io
n
of
a
f
o
r
m
al
d
iag
n
o
s
is
.
F
u
r
t
h
e
r
m
o
r
e,
th
e
au
th
o
r
s
p
lan
to
in
clu
d
e
m
o
r
e
p
ar
am
eter
s
s
u
c
h
as
d
ep
r
ess
io
n
tallies
f
r
o
m
i
n
ter
v
ie
w
s
ta
k
en
f
o
r
d
if
f
er
en
t p
er
io
d
s
o
f
ti
m
e.
Sh
ar
i
f
a
A
l
g
h
o
w
in
e
m
e
t
al
.
[
9
]
ex
tr
ac
ted
an
d
i
n
s
p
ec
ted
t
h
e
e
y
e
m
o
v
e
m
e
n
t
f
ea
t
u
r
es
f
o
r
s
ig
n
s
o
f
d
ep
r
ess
io
n
.
T
h
eir
m
et
h
o
d
o
lo
g
y
e
m
p
lo
y
ed
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
s
(
SVM)
an
d
Gau
s
s
ian
Mix
t
u
r
e
Mo
d
els
f
o
r
class
i
f
icatio
n
.
Ho
w
e
v
er
,
o
w
i
n
g
to
a
s
m
all
a
m
o
u
n
t
o
f
d
ata
s
e
t
th
e
y
d
id
n
o
t
g
et
a
f
air
l
y
h
i
g
h
ac
cu
r
ac
y
.
I
n
f
u
t
u
r
e
th
e
au
t
h
o
r
s
p
lan
to
i
n
teg
r
ate
f
a
ce
,
b
o
d
y
an
d
v
o
ice
f
ea
tu
r
es to
g
et
a
m
o
r
e
ap
t d
etec
tio
n
.
Qu
a
n
H
u
et
al.
[
1
0
]
ass
e
m
b
led
class
i
f
icatio
n
a
n
d
r
eg
r
es
s
io
n
m
o
d
els
to
i
n
s
p
ec
t
b
eh
a
v
io
r
al
an
d
d
ialec
tal
f
ea
t
u
r
es
f
r
o
m
s
o
cial
m
ed
ia
f
o
r
in
d
icatio
n
s
o
f
d
ep
r
ess
io
n
.
T
h
eir
r
esu
lt
s
ca
n
i
n
cr
ea
s
e
s
ig
n
i
f
ica
n
tl
y
b
y
tak
i
n
g
v
ar
ied
p
ar
ticip
an
ts
an
d
in
cr
ea
s
i
n
g
th
e
o
b
s
er
v
a
tio
n
p
er
io
d
f
o
r
im
p
r
o
v
ed
an
al
y
s
is
.
Ak
k
ap
o
n
W
o
n
g
k
o
b
lap
et
al.
[
1
1
]
u
tili
ze
d
d
ee
p
lea
r
n
in
g
m
o
d
el
(
5
-
f
o
ld
cr
o
s
s
v
alid
atio
n
[
1
2
]
)
to
in
v
esti
g
a
te
th
e
p
o
s
ts
i
n
s
o
cial
m
ed
ia.
T
h
e
o
b
tain
ed
ac
cu
r
ac
y
o
f
7
2
p
er
ce
n
t
ca
n
b
e
in
cr
ea
s
ed
b
y
in
c
lu
d
i
n
g
f
u
r
th
er
f
ea
tu
r
es
li
k
e
in
te
r
ac
tio
n
s
w
i
th
f
r
ien
d
s
,
co
m
m
en
t
s
/
r
ep
lies
etc.
Ma
n
d
ar
Desh
p
a
n
d
e
an
d
Vi
g
n
e
s
h
R
ao
[
1
3
]
ap
p
lied
n
atu
r
al
lan
g
u
a
g
e
p
r
o
ce
s
s
in
g
f
o
r
an
al
y
zin
g
t
h
e
s
e
n
ti
m
en
ts
o
f
th
e
t
w
e
ets.
Du
e
to
in
ac
cu
r
ac
y
o
f
p
r
o
p
er
lan
g
u
ag
e
s
t
y
le
i
n
s
o
cial
n
et
w
o
r
k
t
h
e
ac
cu
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
r
ed
u
ce
d
.
Gu
n
tu
k
u
et
al
.
[
1
4
]
ap
p
r
aised
th
e
s
t
u
d
ies
t
h
at
p
r
ed
icted
th
e
m
e
n
tal
h
e
alt
h
o
f
p
eo
p
le
b
ased
o
n
th
e
s
u
r
v
e
y
r
esp
o
n
s
e
s
,
p
o
s
ts
an
d
g
r
o
u
p
s
i
n
ter
ac
ted
in
s
o
c
ial
m
ed
ia.
T
s
u
g
a
w
a
et
al.
[
1
5
]
co
n
s
id
er
ed
u
s
er
ac
tiv
itie
s
in
s
o
cial
m
ed
ia
s
u
c
h
as
f
r
eq
u
en
cie
s
o
f
w
o
r
d
s
r
elate
d
to
m
ela
n
c
h
o
l
y
i
n
a
t
w
ee
t
,
to
p
ics
t
w
ee
ted
o
n
,
p
o
s
tin
g
r
eg
u
lar
it
y
e
tc.
to
ch
ec
k
f
o
r
th
e
m
a
n
i
f
es
tatio
n
o
f
d
e
p
r
ess
io
n
.
No
n
et
h
eles
s
,
b
y
e
m
p
lo
y
i
n
g
tech
n
iq
u
e
s
lik
e
p
r
in
c
ip
le
co
m
p
o
n
e
n
t
a
n
al
y
s
i
s
t
h
e
f
ea
tu
r
e
s
et
u
s
ed
co
u
ld
b
e
i
m
p
r
o
v
ed
.
T
h
e
m
et
h
o
d
s
s
u
ch
as
d
ee
p
lear
n
in
g
an
d
en
s
e
m
b
le
m
et
h
o
d
s
ar
e
ex
p
ec
ted
to
o
f
f
er
b
etter
r
esu
lts
t
h
an
S
VM
.
Ma
r
y
a
m
Mo
h
a
m
m
ed
A
ld
ar
w
i
s
h
an
d
Haf
iz
Far
o
o
q
Ah
m
ed
[
1
6
]
u
s
ed
SVM
an
d
Naï
v
e
B
a
y
es
Mo
d
els
o
n
th
e
p
r
ep
r
o
ce
s
s
ed
p
o
s
ts
o
b
tain
ed
f
r
o
m
s
o
cial
n
et
w
o
r
k
p
lat
f
o
r
m
s
.
T
h
e
ac
cu
r
ac
y
o
b
tain
ed
ca
n
b
e
in
cr
ea
s
ed
b
y
tr
ai
n
i
n
g
a
n
d
co
n
s
tr
u
c
tin
g
b
etter
m
o
d
els.
3.
M
E
T
H
O
DO
L
O
G
Y
Fig
u
r
e
1
h
i
g
h
lig
h
t
s
t
h
e
m
et
h
o
d
o
lo
g
y
f
o
llo
w
ed
in
t
h
e
p
ap
er
.
T
h
e
ai
m
o
f
th
e
p
r
o
p
o
s
ed
w
o
r
k
is
to
p
r
ed
ict
d
e
p
r
ess
io
n
in
i
n
d
iv
id
u
als
u
s
in
g
th
e
ir
b
eh
av
io
r
o
n
li
n
e
(
o
n
t
w
itter
s
p
ec
i
f
icall
y
)
[
1
7
-
1
9
]
.
T
h
is
is
d
o
n
e
in
t
w
o
m
a
in
s
ta
g
e
s
.
First
b
ein
g
th
e
s
ta
g
e
w
h
er
e
s
e
n
ti
m
en
t
an
al
y
s
i
s
[
2
0
]
is
ap
p
lied
o
n
a
p
a
r
ticu
lar
in
d
i
v
id
u
al
's
t
w
i
tter
p
o
s
ts
to
p
r
ed
ict
b
in
ar
y
class
es
(
i.e
.
d
ep
r
ess
ed
/n
o
t
d
e
p
r
ess
ed
)
.
T
h
e
t
w
itter
p
o
s
ts
wer
e
o
b
tain
ed
u
s
i
n
g
th
e
t
w
i
tter
A
P
I
f
r
o
m
a
d
e
v
elo
p
er
t
w
it
ter
ac
co
u
n
t.
A
d
ee
p
le
ar
n
in
g
m
o
d
u
le
k
n
o
w
n
as
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
L
ST
M)
[
2
1
,
2
2
]
is
em
p
lo
y
ed
.
T
h
e
p
r
o
p
o
s
ed
L
ST
M
m
o
d
el
u
s
ed
a
Kag
g
le
d
ataset
o
n
t
w
it
t
er
t
w
ee
ts
r
elate
d
to
d
ep
r
ess
io
n
to
lear
n
an
d
v
alid
at
e.
Fig
u
r
e
1
.
Ov
er
all
m
e
th
o
d
o
lo
g
y
P
r
ep
r
o
ce
s
s
in
g
as
d
ep
icted
i
n
Fig
u
r
e
2
i
n
clu
d
e
d
r
o
p
p
in
g
e
m
p
t
y
t
w
ee
ts
,
r
e
m
o
v
in
g
p
u
n
ctu
atio
n
s
,
cr
ea
tin
g
a
d
ictio
n
ar
y
to
m
ap
w
o
r
d
s
to
in
te
g
er
v
a
lu
e
s
an
d
f
in
d
in
g
m
ax
i
m
u
m
le
n
g
t
h
o
f
th
e
t
w
ee
t
s
[
2
3
]
.
Fin
al
g
en
er
ated
f
ea
t
u
r
es
tr
i
m
m
ed
t
o
th
e
s
eq
u
e
n
ce
le
n
g
t
h
ar
e
f
e
d
in
to
t
h
e
m
o
d
el.
T
h
e
s
eq
u
e
n
tial
L
ST
M
m
o
d
el
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
-
8708
P
r
ed
ictin
g
d
ep
r
ess
io
n
u
s
in
g
d
ee
p
lea
r
n
in
g
a
n
d
e
n
s
emb
le
a
lg
o
r
ith
ms o
n
r
a
w
tw
itte
r
d
a
ta
(
N
is
h
a
P
.
S
h
etty
)
3753
n
et
w
o
r
k
ar
ch
itect
u
r
e
en
co
m
p
ass
es
a
n
e
m
b
ed
d
in
g
la
y
er
,
wh
ich
h
a
s
in
p
u
t
d
i
m
e
n
s
io
n
s
e
t
to
to
tal
n
u
m
b
er
o
f
t
w
ee
t
s
,
o
u
tp
u
t
d
i
m
en
s
io
n
s
et
to
2
0
0
an
d
in
p
u
t
len
g
t
h
s
et
to
s
eq
u
en
ce
len
g
t
h
as
s
tated
ea
r
lier
.
T
h
e
n
ex
t
p
ar
t
in
c
l
u
d
es
an
L
ST
M
la
y
er
w
it
h
5
0
0
u
n
its
,
d
r
o
p
o
u
t
0
.
2
an
d
r
ec
u
r
r
en
t
d
r
o
p
o
u
t
also
0
.
2
.
L
a
s
t
p
ar
t
o
f
t
h
e
m
o
d
el
n
et
w
o
r
k
i
n
cl
u
d
es
a
d
en
s
e
la
y
er
w
it
h
o
n
e
u
n
it
an
d
a
s
i
g
m
o
id
ac
ti
v
atio
n
to
co
n
cise
t
h
e
g
en
er
ated
o
u
tp
u
t
b
et
w
ee
n
ze
r
o
an
d
o
n
e.
T
h
is
c
o
m
p
leted
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
'
s
ar
ch
itect
u
r
e
a
n
d
th
e
m
o
d
el
is
co
m
p
iled
u
s
i
n
g
A
d
a
m
o
p
ti
m
izer
an
d
lo
s
s
i
s
d
ea
lt
b
y
u
s
i
n
g
b
i
n
ar
y
cr
o
s
s
e
n
tr
o
p
y
w
it
h
ac
cu
r
ac
y
m
etr
ic
s
elec
tio
n
to
o
b
s
er
v
e
an
d
ev
al
u
ate.
P
o
s
t
co
m
p
ilatio
n
,
t
h
e
m
o
d
el
is
f
itted
o
n
th
e
ea
r
lier
g
en
er
ated
f
ea
tu
r
es
an
d
v
alid
ated
o
n
lab
els
s
p
ec
if
ied
i
n
t
h
e
d
atase
t.
A
v
a
lid
atio
n
s
p
lit
o
f
0
.
3
(
7
0
%
tr
ai
n
in
g
d
ata,
3
0
%
te
s
t
d
ata)
i
s
i
n
tr
o
d
u
ce
d
an
d
t
h
is
m
o
d
el
i
s
tr
ain
ed
f
o
r
5
ep
o
ch
s
.
Fin
all
y
t
h
i
s
m
o
d
el
is
s
to
r
ed
in
a
J
SON
f
ile
f
o
r
f
u
t
u
r
e
u
s
e.
E
v
er
y
n
e
w
t
w
ee
t
o
b
tain
ed
f
r
o
m
th
e
t
w
itter
A
P
I
g
o
es
th
r
o
u
g
h
th
e
s
a
m
e
p
r
ep
r
o
ce
s
s
i
n
g
p
r
o
ce
d
u
r
e
m
e
n
tio
n
ed
ab
o
v
e
b
ef
o
r
e
b
ein
g
f
o
r
w
ar
d
ed
to
th
e
L
ST
M
m
o
d
el.
T
h
e
r
esu
lt
o
f
th
e
m
o
d
el
is
o
b
tain
ed
f
r
o
m
t
h
e
m
o
d
el.
p
r
ed
i
ct
(
)
f
u
n
ctio
n
an
d
i
s
r
o
u
n
d
ed
to
an
in
te
g
er
v
al
u
e
(
0
o
r
1
in
th
is
ca
s
e)
.
T
h
e
o
b
tain
ed
ac
cu
r
a
c
y
is
co
m
p
ar
ed
to
s
eq
u
e
n
tia
l
C
o
n
v
o
l
u
tio
n
al
Neu
r
al
Ne
t
w
o
r
k
(
C
N
N
)
[
2
4
,
2
5
]
.
T
h
e
p
r
e
p
r
o
ce
s
s
in
g
s
tep
s
ar
e
ex
ac
tl
y
t
h
e
s
a
m
e.
T
h
e
n
et
w
o
r
k
ar
ch
itectu
r
e
i
n
clu
d
es
a
n
e
m
b
ed
d
in
g
la
y
er
s
i
m
i
lar
to
th
e
L
ST
M
e
m
b
ed
d
in
g
.
T
h
e
o
n
l
y
d
if
f
er
en
ce
i
s
a
w
ei
g
h
t
s
ar
g
u
m
e
n
t
is
g
iv
e
n
an
e
m
b
e
d
d
in
g
m
a
tr
ix
,
w
it
h
r
an
d
o
m
v
alu
e
s
i
n
t
h
e
r
a
n
g
e
o
f
2
0
0
to
to
tal
n
u
m
b
er
o
f
t
w
ee
ts
m
u
ltip
lied
b
y
0
.
0
1
.
T
h
is
i
s
f
o
ll
o
w
ed
b
y
a
d
r
o
p
o
u
t
o
f
0
.
4
w
h
ic
h
f
ee
d
s
t
h
e
d
ata
to
a
to
tal
o
f
f
o
u
r
1
D
co
n
v
o
l
u
tio
n
la
y
er
s
.
E
ac
h
o
f
t
h
e
co
n
v
o
lu
tio
n
la
y
er
s
h
as
k
er
n
e
l
s
ize
s
et
to
3
,
p
ad
d
in
g
s
et
to
v
a
lid
,
ac
tiv
atio
n
i
s
r
elu
a
n
d
s
tr
id
es
is
s
et
to
1
.
On
l
y
t
h
in
g
t
h
at
d
if
f
er
ed
is
t
h
e
f
ilter
(
d
im
e
n
s
io
n
alit
y
o
f
t
h
e
o
u
tp
u
t
s
p
ac
e)
.
I
t
is
d
ec
r
ea
s
ed
b
y
5
0
%
at
ea
ch
la
y
er
.
First
la
y
er
h
ad
f
ilter
s
et
to
6
0
0
,
s
ec
o
n
d
h
ad
3
0
0
,
th
ir
d
1
5
0
an
d
f
o
u
r
t
h
h
ad
7
5
.
Af
ter
t
h
i
s
,
f
latte
n
i
s
i
n
cl
u
d
ed
in
th
e
m
o
d
el
ar
ch
itect
u
r
e
to
f
lat
ten
th
e
i
n
p
u
t.
T
h
i
s
i
s
f
o
llo
w
ed
b
y
ad
d
in
g
a
d
en
s
e
la
y
er
w
it
h
6
0
0
u
n
it
s
,
d
r
o
p
o
u
t
o
f
0
.
5
,
ac
tiv
a
tio
n
s
et
to
r
elu
,
a
d
e
n
s
e
la
y
er
w
i
th
o
n
e
u
n
i
t,
an
d
f
i
n
all
y
,
an
ac
tiv
a
tio
n
s
et
to
s
i
g
m
o
id
.
C
NN
m
o
d
el
is
co
m
p
il
ed
u
s
i
n
g
ex
ac
t
l
y
s
a
m
e
ar
g
u
m
en
ts
f
o
r
lo
s
s
,
m
e
tr
ic
an
d
o
p
ti
m
izer
.
T
h
e
s
ec
o
n
d
s
tag
e
in
cl
u
d
es
tr
y
in
g
to
i
m
p
r
o
v
e
th
e
o
u
tco
m
e
o
f
th
e
p
r
o
p
o
s
ed
w
o
r
k
u
s
i
n
g
b
asi
c
m
ac
h
in
e
lear
n
in
g
clas
s
i
f
ier
s
[
2
6
,
2
7
]
an
d
a
f
e
w
o
p
ti
m
ized
en
s
e
m
b
les.
C
la
s
s
i
f
ie
r
s
u
s
ed
ar
e
lo
g
is
tic
r
eg
r
es
s
io
n
,
lin
ea
r
s
u
p
p
o
r
t
v
ec
to
r
class
i
f
ie
r
(
SVC
)
,
m
u
lti
n
o
m
ia
l
n
ai
v
e
b
a
y
es,
b
er
n
o
u
l
li
n
a
iv
e
b
a
y
es
al
o
n
g
w
it
h
e
n
s
e
m
b
les
lik
e
r
an
d
o
m
f
o
r
est cla
s
s
i
f
ier
an
d
g
r
ad
ien
t b
o
o
s
tin
g
cla
s
s
i
f
ier
.
P
r
ep
r
o
ce
s
s
in
g
as
d
ep
icted
in
Fig
u
r
e
3
in
cl
u
d
es
s
p
l
i
ttin
g
o
f
d
ata
in
to
tr
ai
n
an
d
tes
t
s
et.
T
h
is
is
f
o
llo
w
ed
b
y
v
ec
to
r
izin
g
t
h
e
t
w
ee
t
s
.
T
h
e
m
e
n
tio
n
ed
s
ta
g
e
p
r
o
v
id
es
th
e
clas
s
if
ier
s
w
i
th
th
r
ee
d
if
f
er
en
t
k
i
n
d
o
f
v
ec
to
r
izer
s
n
a
m
el
y
co
u
n
t
v
ec
to
r
izer
,
T
F
-
I
DF
an
d
n
-
g
r
a
m
s
[
2
8
]
.
Fo
r
C
o
u
n
t
Vec
to
r
izer
,
th
e
p
r
ep
r
o
ce
s
s
in
g
in
cl
u
d
es
f
it
tin
g
t
h
e
co
u
n
t
v
ec
t
o
r
izer
o
n
th
e
tr
ain
in
g
an
d
d
ata
th
en
tr
an
s
f
o
r
m
i
n
g
t
h
e
d
o
cu
m
en
ts
i
n
th
e
tr
ai
n
in
g
d
ata
to
d
o
cu
m
e
n
t
-
ter
m
m
atr
i
x
.
Fig
u
r
e
2
.
L
ST
M
p
r
ep
r
o
ce
s
s
in
g
Fig
u
r
e
3
.
Mo
d
el
v
ec
to
r
p
r
e
p
r
o
ce
s
s
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
4
,
A
u
g
u
s
t
2020
:
3
7
5
1
-
375
6
3754
T
h
e
o
b
tain
ed
tr
ain
in
g
d
ata
(
f
ea
tu
r
es)
is
f
ed
to
th
e
cu
r
r
en
t
m
o
d
el
s
elec
ted
f
r
o
m
t
h
e
lis
t.
A
f
ea
t
u
r
e
n
a
m
e
s
l
is
t
is
in
itialized
a
n
d
p
o
p
u
lated
u
s
in
g
C
o
u
n
t
Vec
to
r
izer
'
s
g
et_
f
ea
t
u
r
e_
n
a
m
es
(
)
f
u
n
ctio
n
.
R
e
s
u
l
ts
ar
e
p
r
ed
icted
u
s
in
g
p
r
ed
ict
f
u
n
ct
i
o
n
.
T
F
-
I
DF
Vec
to
r
i
ze
r
is
f
it
o
n
th
e
tr
ain
i
n
g
d
ata
u
s
i
n
g
m
i
n
_
d
f
=5
,
s
i
m
ilar
to
C
o
u
n
t
Vec
to
r
izer
,
d
o
cu
m
en
t
s
ar
e
tr
an
s
f
o
r
m
ed
an
d
th
e
tr
ain
i
n
g
d
ata
is
f
ed
to
th
e
cu
r
r
en
t
m
o
d
el
s
elec
ted
f
r
o
m
th
e
li
s
t.
A
lo
n
g
w
i
th
t
h
i
s
,
f
ea
t
u
r
e
n
a
m
es
a
n
d
s
o
r
ted
T
F
-
I
DF
in
d
ex
i
s
also
ca
lc
u
lated
to
f
i
n
d
th
e
s
m
allest
a
n
d
h
ig
h
e
s
t
T
F
-
I
DF,
(
least
a
n
d
m
o
s
t
i
m
p
o
r
tan
t
co
ef
f
icie
n
t
s
)
.
Fo
r
n
-
g
r
a
m
s
,
s
i
m
ilar
p
r
o
ce
d
u
r
e
w
a
s
f
o
llo
w
ed
an
d
i
s
f
it
o
n
th
e
tr
ai
n
i
n
g
d
ata
u
s
in
g
m
i
n
_
d
f
=5
an
d
n
-
g
r
a
m
_
r
a
n
g
e
(
1
,
2
)
.
Sim
ilar
to
o
th
er
v
ec
to
r
izer
s
,
s
m
alle
s
t
a
n
d
h
ig
h
e
s
t c
o
ef
f
icie
n
ts
ar
e
n
o
ted
an
d
r
esu
lt
s
en
t
i
m
e
n
t o
f
t
h
e
t
wee
t is p
r
ed
icted
.
E
ac
h
an
d
ev
er
y
clas
s
if
ier
an
d
en
s
e
m
b
le
m
e
n
tio
n
ed
ab
o
v
e
is
f
it
o
n
ea
c
h
o
f
t
h
e
th
r
ee
v
ec
to
r
izer
s
an
d
r
esu
lt
s
ar
e
n
o
ted
.
T
h
is
s
ta
g
e
is
lar
g
el
y
f
o
cu
s
ed
o
n
cr
o
s
s
-
ch
ec
k
i
n
g
p
r
ed
ictio
n
s
m
ad
e
b
y
t
h
e
f
ir
s
t
st
a
g
e.
E
v
er
y
n
e
w
t
w
ee
t
o
b
tain
ed
th
r
o
u
g
h
th
e
t
w
itter
A
P
I
is
s
en
t
t
h
r
o
u
g
h
f
ir
s
t
a
n
d
s
ec
o
n
d
s
tag
e.
I
n
th
e
s
ec
o
n
d
s
ta
g
e
ev
er
y
p
o
s
s
ib
ilit
y
o
f
m
o
d
el
a
n
d
v
ec
to
r
izer
is
e
x
ec
u
ted
an
d
r
esu
lts
ar
e
n
o
ted
.
T
h
e
f
i
n
al
r
esu
lt
in
cl
u
d
ed
th
e
w
ei
g
h
ted
m
ea
n
o
f
all
s
ec
o
n
d
s
tag
e
p
o
s
s
ib
ili
ties
.
T
h
is
v
al
u
e
is
cr
o
s
s
-
c
h
ec
k
ed
w
ith
t
h
e
f
ir
s
t
s
ta
g
e
p
r
ed
ictio
n
s
.
T
h
e
w
eig
h
t
s
ar
e
ass
i
g
n
ed
ac
co
r
d
in
g
to
ac
c
u
r
ac
y
o
f
th
e
m
o
d
el
o
n
t
h
e
d
ata.
4.
RE
SU
L
T
T
h
e
f
o
llo
w
i
n
g
tab
les
T
ab
les
1
-
7
s
h
o
w
t
h
e
r
es
u
lts
o
b
tai
n
ed
in
th
is
r
esear
ch
.
T
h
e
m
en
tio
n
ed
class
i
f
ier
s
's r
esu
lts
ar
e
co
m
p
ar
ed
h
er
e.
T
ab
le
1
.
L
ST
M
v
s
C
NN
C
l
a
ssi
f
i
e
r
V
a
l
i
d
a
t
i
o
n
A
c
c
u
r
a
c
y
T
e
st
A
c
c
u
r
a
c
y
L
S
T
M
0
.
7
0
.
9
3
C
N
N
0
.
6
8
0
.
9
5
T
ab
le
2
.
L
o
g
is
tic
r
eg
r
es
s
io
n
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
C
o
u
n
t
v
e
c
t
o
r
i
z
e
r
7
5
.
8
1
%
TF
-
I
D
F
7
6
.
2
2
%
n
-
g
r
a
ms
7
6
.
1
5
%
T
ab
le
3
.
R
an
d
o
m
f
o
r
est
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
C
o
u
n
t
v
e
c
t
o
r
i
z
e
r
7
0
%
TF
-
I
D
F
7
2
%
n
-
g
r
a
ms
7
2
%
T
ab
le
4
.
B
er
n
o
u
lli n
aï
v
e
b
a
y
e
s
(
NB
)
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
TF
-
I
D
F
7
3
.
9
5
%
n
-
g
r
a
ms
7
5
.
5
3
%
T
ab
le
5
.
Mu
ltin
o
m
ia
l n
aï
v
e
b
ay
es (
NB
)
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
TF
-
I
D
F
7
4
.
2
2
3
%
n
-
g
r
a
ms
7
6
.
6
9
%
T
ab
le
6
.
L
in
ea
r
SVC
(
r
an
d
o
m
_
s
tate=
1
0
0
,
to
l=1
e
-
10)
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
C
o
u
n
t
v
e
c
t
o
r
i
z
e
r
7
3
.
7
1
%
TF
-
I
D
F
7
5
.
4
8
%
n
-
g
r
a
ms
7
3
.
5
3
%
T
ab
le
6
.
Gr
a
d
ien
t
B
o
o
s
tin
g
C
l
ass
i
f
ier
(
s
u
b
s
a
m
p
le=
0
.
8
,
lear
n
in
g
_
r
ate=
0
.
0
5
,
n
_
esti
m
a
to
r
s
=2
5
0
,
r
an
d
o
m
_
s
tate
=5
,
m
a
x
_
d
ep
th
=
2
0
,
m
ax
_
lea
f
_
n
o
d
es=1
5
0
)
C
l
a
ssi
f
i
e
r
A
c
c
u
r
a
c
y
n
-
g
r
a
ms
7
4
.
4
2
%
5.
CO
NCLU
SI
O
N
AND
F
U
T
U
RE
WO
RK
Of
te
n
tr
ad
itio
n
al
s
u
r
v
e
y
-
b
a
s
ed
q
u
esti
o
n
s
f
ail
to
u
n
co
v
e
r
th
e
e
x
ten
t
o
f
u
s
er
s
‘
m
en
tal
h
ea
lt
h
d
ep
r
ec
iatio
n
.
No
w
ad
a
y
s
,
s
o
ci
al
m
ed
ia
i
s
a
co
m
m
o
n
p
lat
f
o
r
m
,
w
h
ic
h
p
eo
p
le
u
s
e
to
r
e
ac
h
o
u
t.
T
h
er
ef
o
r
e,
th
e
ab
o
v
e
p
r
o
p
o
s
ed
m
et
h
o
d
o
lo
g
y
ca
s
h
e
s
in
th
i
s
p
o
p
u
lar
it
y
o
f
SN
S
a
n
d
e
v
alu
ate
s
t
h
e
d
ep
r
ess
io
n
le
v
els
o
f
th
e
u
s
er
s
b
y
e
m
p
lo
y
i
n
g
n
at
u
r
a
l
lan
g
u
a
g
e
a
n
d
m
ac
h
i
n
e
lear
n
i
n
g
tech
n
iq
u
e
s
.
I
n
f
u
t
u
r
e,
b
y
co
llectin
g
m
o
r
e
d
ata,
m
o
r
e
f
r
eq
u
en
tl
y
w
i
th
t
h
e
ai
m
o
f
i
m
p
r
o
v
in
g
t
h
e
ac
cu
r
ac
y
o
f
th
e
w
o
r
k
to
g
i
v
e
a
b
etter
d
iag
n
o
s
is
.
S
u
ch
to
o
l
s
w
h
ic
h
ca
n
p
r
ed
ict
v
ar
iatio
n
s
i
n
p
er
s
o
n
’
s
m
o
o
d
ca
n
b
e
an
i
m
p
o
r
tan
t
m
et
h
o
d
f
o
r
b
o
th
clin
ic
al
o
b
s
er
v
atio
n
s
a
n
d
s
elf
-
d
ia
g
n
o
s
i
n
g
.
T
h
e
m
et
h
o
d
ca
n
b
e
ti
m
e
co
n
s
u
m
i
n
g
a
n
d
h
en
ce
s
tep
s
m
u
s
t
b
e
ta
k
e
n
i
n
t
h
is
r
e
g
ar
d
to
i
m
p
r
o
v
e
u
p
o
n
.
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
-
8708
P
r
ed
ictin
g
d
ep
r
ess
io
n
u
s
in
g
d
ee
p
lea
r
n
in
g
a
n
d
e
n
s
emb
le
a
lg
o
r
ith
ms o
n
r
a
w
tw
itte
r
d
a
ta
(
N
is
h
a
P
.
S
h
etty
)
3755
RE
F
E
R
E
NC
E
S
[1
]
S
tatista,
“
M
o
st
p
o
p
u
l
a
r
so
c
i
a
l
n
e
tw
o
rk
s
w
o
rld
w
id
e
a
s
o
f
Oc
to
b
e
r,
”
2
0
1
9
.
[
O
n
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s:/
/www
.
sta
ti
sta
.
c
o
m
/statisti
c
s/2
7
2
0
1
4
/g
lo
b
a
l
-
so
c
ial
-
n
e
tw
o
rk
s
-
ra
n
k
e
d
-
by
-
n
u
m
b
e
r
-
of
-
u
se
rs/
[2
]
W
HO
,
“
W
HO
D
e
p
re
ss
io
n
,
”
2
0
1
9
.
[
O
n
li
n
e
].
A
v
a
il
a
b
le:
h
tt
p
s://
w
ww
.
w
h
o
.
in
t/
n
e
ws
-
ro
o
m
/
f
a
c
t
-
sh
e
e
ts/d
e
tail/
d
e
p
re
ss
io
n
[3
]
F
ra
n
c
e
s,
A
.
,
P
in
c
u
s,
H.,
&
F
irst,
M
.
,
“
M
a
jo
r
De
p
re
ss
iv
e
Ep
iso
d
e
.
In
Dia
g
n
o
stic
a
n
d
sta
ti
stica
l
m
a
n
u
a
l
o
f
m
e
n
tal
d
iso
rd
e
rs:
D
S
M
-
IV
,
”
W
a
sh
i
n
g
t
o
n
,
DC: A
me
ric
a
n
Psy
c
h
i
a
tric A
ss
o
c
ia
ti
o
n
,
1
9
9
4
.
[4
]
M
a
il
o
n
li
n
e
,
“
M
a
lay
sia
n
te
e
n
a
g
e
r
,
1
6
,
'
d
ied
j
u
m
p
in
g
f
ro
m
th
e
th
ird
f
lo
o
r
o
f
a
sh
o
p
a
f
ter
c
o
n
d
u
c
ti
n
g
a
n
In
sta
g
ra
m
p
o
ll
o
n
w
h
e
th
e
r
sh
e
sh
o
u
ld
k
il
l
h
e
rse
lf'
-
a
n
d
6
9
p
e
r
c
e
n
t
c
h
o
se
'
d
e
a
th
'
,
”
2
0
1
9
.
[
On
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s:/
/www
.
d
a
il
y
m
a
il
.
c
o
.
u
k
/n
e
ws
/article
7
0
3
1
0
6
7
/
M
a
lay
sia
n
-
tee
n
a
g
e
r
-
16
-
k
il
led
-
c
o
n
d
u
c
ti
n
g
-
In
sta
g
ra
m
-
p
o
ll
.
h
tm
l.
[5
]
S
e
re
n
a
G
o
rd
o
n
,
“
F
a
c
e
b
o
o
k
P
o
sts
M
a
y
Hin
t
a
t
De
p
re
ss
io
n
,
”
2
0
1
8
.
[
On
li
n
e
]
A
v
a
il
a
b
le:
h
tt
p
s:/
/www
.
w
e
b
m
d
.
c
o
m
/d
e
p
re
ss
io
n
/
n
e
w
s/2
0
1
8
1
0
1
5
/f
a
c
e
b
o
o
k
-
p
o
sts
-
m
a
y
-
h
in
t
-
at
-
d
e
p
re
ss
io
n
#
1
[6
]
Ja
c
k
Ca
r
fa
g
n
o
,
“
S
o
c
ial
M
e
d
ia
P
o
sts
Ca
n
He
lp
I
d
e
n
ti
f
y
De
p
re
ss
io
n
,
”
2
0
1
9
.
[
On
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s:/
/www
.
d
o
c
w
iren
e
w
s.c
o
m
/d
o
c
w
ire
-
p
ick
/so
c
ial
-
m
e
d
ia
-
p
o
sts
-
can
-
h
e
lp
-
i
d
e
n
ti
f
y
-
d
e
p
re
ss
io
n
/
[7
]
Ha
q
u
e
,
A
lb
e
rt,
M
ic
h
e
ll
e
G
u
o
,
A
d
a
m
S
.
M
in
e
r
a
n
d
L
i
F
e
i
-
F
e
i,
“
M
e
a
su
rin
g
De
p
re
ss
io
n
S
y
m
p
to
m
S
e
v
e
rit
y
f
ro
m
S
p
o
k
e
n
L
a
n
g
u
a
g
e
a
n
d
3
D
F
a
c
ial
Ex
p
re
ss
io
n
s,”
2
0
1
8
.
[
O
n
li
n
e
]
A
v
a
il
a
b
le:
a
rXiv.o
rg
a
rXiv:
1
8
1
1
.
0
8
5
9
2
[8
]
S
.
A
lb
a
w
i,
T
.
A
.
M
o
h
a
m
m
e
d
a
n
d
S
.
A
l
-
Zaw
i,
“
Un
d
e
rsta
n
d
in
g
o
f
a
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
tw
o
rk
,
”
2
0
1
7
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
E
n
g
i
n
e
e
rin
g
a
n
d
T
e
c
h
n
o
l
o
g
y
(
ICET
)
,
p
p
.
1
-
6
,
2
0
1
7
.
[9
]
S
.
A
lg
h
o
w
in
e
m
,
R.
G
o
e
c
k
e
,
M
.
W
a
g
n
e
r,
G
.
P
a
rk
e
r
a
n
d
M
.
Bre
a
k
sp
e
a
r,
“
E
y
e
m
o
v
e
m
e
n
t
a
n
a
l
y
sis
f
o
r
d
e
p
re
ss
io
n
d
e
tec
ti
o
n
,
”
2
0
1
3
IEE
E
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Im
a
g
e
Pro
c
e
ss
in
g
,
M
e
lb
o
u
rn
e
,
p
p
.
4
2
2
0
-
4
2
2
4
,
2
0
1
3
.
[1
0
]
Q.
Hu
,
A
.
L
i,
F
.
He
n
g
,
J.
L
i
a
n
d
T
.
Zh
u
,
“
P
re
d
icti
n
g
De
p
re
ss
io
n
o
f
S
o
c
ial
M
e
d
ia
Us
e
r
o
n
Diff
e
re
n
t
Ob
se
rv
a
ti
o
n
W
in
d
o
w
s,”
2
0
1
5
IEE
E
/W
IC/A
CM
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
W
e
b
In
telli
g
e
n
c
e
a
n
d
In
telli
g
e
n
t
Ag
e
n
t
T
e
c
h
n
o
l
o
g
y
(
W
I
-
IAT
),
S
in
g
a
p
o
re
,
p
p
.
3
6
1
-
3
6
4
,
2
0
1
5
.
DO
I:
1
0
.
1
1
0
9
/W
I
-
IA
T
.
2
0
1
5
.
1
6
6
.
[1
1
]
A
.
W
o
n
g
k
o
b
lap
,
M
.
A
.
V
a
d
il
lo
a
n
d
V
.
Cu
rc
i
n
,
“
Clas
sify
in
g
D
e
p
re
ss
e
d
Us
e
rs
w
it
h
M
u
lt
ip
le
In
sta
n
c
e
L
e
a
rn
in
g
f
ro
m
S
o
c
ial
Ne
tw
o
rk
D
a
ta,”
2
0
1
8
IEE
E
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
H
e
a
lt
h
c
a
re
In
f
o
rm
a
ti
c
s
(
ICHI),
p
p
.
4
3
6
-
4
3
6
,
2
0
1
8
.
DO
I:
1
0
.
1
1
0
9
/
ICHI.2
0
1
8
.
0
0
0
8
8
.
[1
2
]
J.
D.
Ro
d
rig
u
e
z
,
A
.
P
e
re
z
a
n
d
J.
A
.
L
o
z
a
n
o
,
“
S
e
n
siti
v
it
y
A
n
a
l
y
si
s
o
f
k
-
F
o
ld
Cro
ss
V
a
li
d
a
ti
o
n
in
P
re
d
ictio
n
Err
o
r
Esti
m
a
ti
o
n
,
”
in
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Pa
tt
e
rn
A
n
a
lys
is
a
n
d
M
a
c
h
in
e
In
telli
g
e
n
c
e
,
v
o
l.
3
2
,
n
o
.
3
,
p
p
.
5
6
9
-
5
7
5
,
M
a
rc
h
2
0
1
0
.
d
o
i:
1
0
.
1
1
0
9
/T
P
A
M
I.
2
0
0
9
.
1
8
7
.
[1
3
]
M
.
De
sh
p
a
n
d
e
a
n
d
V
.
Ra
o
,
“
De
p
re
ss
io
n
d
e
tec
ti
o
n
u
sin
g
e
m
o
ti
o
n
a
rti
f
icia
l
in
telli
g
e
n
c
e
,
”
2
0
1
7
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
In
tell
ig
e
n
t
S
u
sta
i
n
a
b
le
S
y
ste
ms
(
I
CIS
S
),
P
a
ll
a
d
a
m
,
p
p
.
8
5
8
-
8
6
2
,
2
0
1
7
.
DO
I:
1
0
.
1
1
0
9
/
IS
S
1
.
2
0
1
7
.
8
3
8
9
2
9
9
.
[1
4
]
S
h
a
ra
th
Ch
a
n
d
ra
G
u
n
tu
k
u
,
Da
v
id
B.
Ya
d
e
n
,
M
a
rg
a
re
t
L
.
Ke
rn
,
Ly
le
H.
Un
g
a
r,
Jo
h
a
n
n
e
s
C.
Ei
c
h
sta
e
d
t,
“
De
tec
ti
n
g
d
e
p
re
ss
io
n
a
n
d
m
e
n
tal
il
ln
e
ss
o
n
so
c
ial
m
e
d
ia:
a
n
in
teg
ra
ti
v
e
re
v
iew
,
”
Cu
rr
e
n
t
Op
in
i
o
n
i
n
Beh
a
v
io
ra
l
S
c
ien
c
e
s
,
v
o
l.
1
8
,
p
p
.
4
3
-
4
9
,
2
0
1
7
.
[1
5
]
S
h
o
T
su
g
a
w
a
,
Yu
su
k
e
Kik
u
c
h
i,
F
u
m
io
Kish
in
o
,
K
o
su
k
e
Na
k
a
ji
m
a
,
Yu
ich
i
It
o
h
,
a
n
d
Hir
o
y
u
k
i
Oh
sa
k
i
,
“
Re
c
o
g
n
izin
g
De
p
re
ss
io
n
f
ro
m
Tw
it
ter
Ac
ti
v
it
y
,
”
In
Pro
c
e
e
d
in
g
s
o
f
th
e
3
3
rd
An
n
u
a
l
ACM
Co
n
fer
e
n
c
e
o
n
Hu
ma
n
Fa
c
to
rs
i
n
C
o
mp
u
ti
n
g
S
y
ste
ms
(
CHI
'
1
5
)
.
ACM
,
Ne
w
Y
o
rk
,
NY
,
US
A
,
p
p
.
3
1
8
7
-
3
1
9
6
,
2
0
1
5
.
DO
I:
ht
tp
s:/
/d
o
i.
o
rg
/1
0
.
1
1
4
5
/2
7
0
2
1
2
3
.
2
7
0
2
2
8
0
.
[1
6
]
M
.
M
.
A
ld
a
rw
ish
a
n
d
H.
F
.
A
h
m
a
d
,
“
P
re
d
ictin
g
De
p
re
ss
io
n
L
e
v
e
ls
Us
in
g
S
o
c
ial
M
e
d
ia
P
o
sts,”
2
0
1
7
IEE
E
1
3
t
h
In
ter
n
a
t
io
n
a
l
S
y
mp
o
siu
m
o
n
Au
to
n
o
mo
u
s
De
c
e
n
tra
li
ze
d
S
y
ste
m
(
IS
ADS
),
p
p
.
2
7
7
-
2
8
0
,
2
0
1
7
.
DO
I:
1
0
.
1
1
0
9
/
IS
A
DS.
2
0
1
7
.
41.
[1
7
]
F
a
rig
S
a
d
e
q
u
e
,
Do
n
g
f
a
n
g
X
u
,
a
n
d
S
tev
e
n
Be
th
a
rd
,
“
M
e
a
su
ri
n
g
th
e
L
a
ten
c
y
o
f
De
p
re
s
sio
n
De
tec
ti
o
n
i
n
S
o
c
ial
M
e
d
ia,”
In
Pro
c
e
e
d
i
n
g
s
o
f
th
e
El
e
v
e
n
th
ACM
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
W
e
b
S
e
a
rc
h
a
n
d
Da
t
a
M
i
n
in
g
(
W
S
D
M
’1
8
)
,
p
p
.
4
9
5
–
5
0
3
,
2
0
1
8
.
[1
8
]
Bu
rd
isso
,
S
e
rg
io
G
.
,
M
a
rc
e
lo
E
.
,
a
n
d
M
.
M
o
n
tes
-
y
-
G
ó
m
e
z
,
“
A
T
e
x
t
Clas
si
f
ica
ti
o
n
F
ra
m
e
w
o
rk
fo
r
S
im
p
le
a
n
d
Eff
e
c
ti
v
e
Earl
y
De
p
re
ss
io
n
De
tec
ti
o
n
o
v
e
r
S
o
c
ial
M
e
d
ia S
trea
m
s,”
a
rXiv.o
rg
a
rXiv:
1
9
0
5
.
0
8
7
7
2
,
p
p
.
1
-
2
0
,
2
0
1
9
.
[1
9
]
V
iri
d
ian
a
Ro
m
e
ro
M
.
,
“
A
m
a
c
h
in
e
lea
rn
in
g
a
p
p
ro
a
c
h
f
o
r
th
e
d
e
tec
t
io
n
o
f
d
e
p
re
ss
io
n
a
n
d
m
e
n
tal
il
ln
e
ss
in
Tw
it
ter,”
2
0
1
9
.
[
On
l
in
e
]
.
A
v
a
il
a
b
le:
h
tt
p
s://
m
e
d
iu
m
.
c
o
m
/d
a
tad
riv
e
n
in
v
e
sto
r/a
-
m
a
c
h
in
e
-
lea
rn
in
g
-
a
p
p
ro
a
c
h
-
f
o
r
-
d
e
tec
ti
o
n
-
of
-
d
e
p
re
ss
io
n
-
a
n
d
-
m
e
n
tal
-
il
ln
e
ss
-
in
-
tw
it
ter
-
3
f
3
a
3
2
a
4
d
f
6
0
.
[2
0
]
S
.
A
.
El
Ra
h
m
a
n
,
F
.
A
.
A
lOtaib
i
a
n
d
W
.
A
.
A
lS
h
e
h
ri,
“
S
e
n
ti
m
e
n
t
A
n
a
l
y
sis
o
f
Tw
it
ter
Da
ta,”
2
0
1
9
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
a
n
d
In
f
o
rm
a
ti
o
n
S
c
ien
c
e
s (
ICCIS
)
,
p
p
.
1
-
4
,
2
0
1
9
.
[2
1
]
J.
W
a
n
g
a
n
d
Z.
Ca
o
,
“
Ch
in
e
se
tex
t
se
n
ti
m
e
n
t
a
n
a
l
y
sis
u
sin
g
L
S
TM
n
e
tw
o
rk
b
a
se
d
o
n
L
2
a
n
d
Na
d
a
m
,
”
20
1
7
IE
EE
1
7
t
h
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Co
mm
u
n
ic
a
ti
o
n
T
e
c
h
n
o
lo
g
y
(
ICCT
),
Ch
e
n
g
d
u
,
p
p
.
1
8
9
1
-
1
8
9
5
,
2
0
1
7
.
DO
I:
1
0
.
1
1
0
9
/
ICCT
.
2
0
1
7
.
8
3
5
9
9
5
8
.
[2
2
]
F
.
L
iu
,
X
.
Z
h
o
u
,
J.
Ca
o
,
Z.
W
a
n
g
,
H.
W
a
n
g
a
n
d
Y.
Zh
a
n
g
,
“
A
LS
T
M
a
n
d
CNN
Ba
se
d
A
ss
e
m
b
le
Ne
u
ra
l
Ne
tw
o
r
k
F
ra
m
e
w
o
rk
f
o
r
A
rr
h
y
th
m
ias
Cla
ss
if
ic
a
ti
o
n
,
”
ICA
S
S
P
2
0
1
9
-
2
0
1
9
IEE
E
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Aco
u
stics
,
S
p
e
e
c
h
a
n
d
S
ig
n
a
l
Pro
c
e
ss
in
g
(
ICAS
S
P),
Br
ig
h
to
n
,
Un
it
e
d
Ki
n
g
d
o
m
,
p
p
.
1
3
0
3
-
1
3
0
7
,
2
0
1
9
.
DO
I:
1
0
.
1
1
0
9
/
ICA
S
S
P
.
2
0
1
9
.
8
6
8
2
2
9
9
.
[2
3
]
D.
Zi
m
b
ra
,
M
.
G
h
ias
si
a
n
d
S
.
L
e
e
,
“
Bra
n
d
-
Re
late
d
Tw
it
ter
S
e
n
ti
m
e
n
t
A
n
a
l
y
sis
Us
in
g
F
e
a
tu
re
En
g
i
n
e
e
rin
g
a
n
d
th
e
D
y
n
a
m
ic
A
rc
h
it
e
c
tu
re
f
o
r
A
rti
f
ic
ial
Ne
u
ra
l
Ne
tw
o
rk
s,”
4
9
th
Ha
w
a
ii
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
S
y
ste
m
S
c
ien
c
e
s
(
HICS
S
),
p
p
.
1
9
3
0
-
1
9
3
8
,
2
0
1
6
.
d
o
i:
1
0
.
1
1
0
9
/HIC
S
S
.
2
0
1
6
.
2
4
4
.
[2
4
]
A
b
d
e
lg
h
a
n
i
Da
h
o
u
,
M
o
h
a
m
e
d
Ab
d
El
a
z
iz,
Ju
n
w
e
i
Zh
o
u
,
a
n
d
S
h
e
n
g
w
u
X
io
n
g
,
“
A
ra
b
ic
S
e
n
ti
m
e
n
t
Clas
sif
i
c
a
ti
o
n
Us
in
g
Co
n
v
o
lu
ti
o
n
a
l
Ne
u
ra
l
Ne
tw
o
rk
a
n
d
Diff
e
r
e
n
ti
a
l
Ev
o
lu
ti
o
n
A
lg
o
rit
h
m
,
”
Co
mp
u
ta
ti
o
n
a
l
I
n
telli
g
e
n
c
e
a
n
d
Ne
u
ro
sc
ien
c
e
,
v
o
l.
2
0
1
9
,
p
p
.
1
-
1
6
,
2
0
1
9
.
DO
I:
h
tt
p
s:
//
d
o
i.
o
rg
/1
0
.
1
1
5
5
/
2
0
1
9
/
2
5
3
7
6
8
9
.
[2
5
]
Kim
H.
Je
o
n
g
,
Y.
-
S
.
,
“
S
e
n
ti
m
e
n
t
Clas
sif
ic
a
ti
o
n
Us
in
g
Co
n
v
o
lu
ti
o
n
a
l
Ne
u
ra
l
Ne
tw
o
rk
s,”
Ap
p
l.
S
c
i.
,
v
o
l.
9
,
2
0
1
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
10
,
No
.
4
,
A
u
g
u
s
t
2020
:
3
7
5
1
-
375
6
3756
[2
6
]
Ha
o
G
u
o
,
M
e
n
g
n
a
Qin
,
Ju
n
ji
e
Ch
e
n
,
Yo
n
g
Xu
,
a
n
d
Jie
X
ia
n
g
,
“
M
a
c
h
in
e
-
L
e
a
rn
in
g
Clas
sif
i
e
r
fo
r
P
a
ti
e
n
ts
w
it
h
M
a
jo
r
De
p
re
ss
iv
e
Diso
rd
e
r:
M
u
l
ti
f
e
a
tu
re
A
p
p
ro
a
c
h
Ba
se
d
o
n
a
Hig
h
-
Ord
e
r
M
in
im
u
m
S
p
a
n
n
in
g
T
re
e
F
u
n
c
ti
o
n
a
l
Bra
in
Ne
tw
o
rk
,
”
Co
m
p
u
tatio
n
a
l
a
n
d
M
a
th
e
m
a
ti
c
a
l
M
e
th
o
d
s i
n
M
e
d
icin
e
,
v
o
l.
2
0
1
7
,
p
p
.
1
-
14
,
2
0
1
7
.
[2
7
]
I
sh
it
a
Bh
a
k
ta
a
n
d
A
rk
a
p
ra
b
h
a
S
a
u
,
“
P
re
d
ictio
n
o
f
De
p
re
ss
io
n
a
m
o
n
g
S
e
n
io
r
Cit
ize
n
s
u
si
n
g
M
a
c
h
in
e
L
e
a
rn
in
g
Clas
sif
ier
s
,”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
C
o
mp
u
ter
Ap
p
li
c
a
t
io
n
s
,
v
o
l.
1
4
4
,
n
o
.
7
,
p
p
.
11
-
1
6
,
2
0
1
6
.
[2
8
]
A
n
k
it
Ba
sa
r
k
a
r,
“
Do
c
u
m
e
n
t
Clas
s
if
ica
ti
o
n
u
sin
g
M
a
c
h
i
n
e
L
e
a
rn
in
g
,”
S
a
n
J
o
se
S
t
a
te Un
ive
rs
it
y
.
Ca
li
f
o
rn
ia
,
2
0
1
7
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Nish
a
P.
S
h
e
tty
h
a
s p
u
b
li
s
h
e
d
in
th
e
a
re
a
s n
e
t
w
o
rk
se
c
u
rit
y
a
n
d
m
a
c
h
in
e
lea
rn
in
g
.
Cu
rre
n
t
ly
sh
e
is
w
o
rk
in
g
in
th
e
a
re
a
o
f
so
c
ial
n
e
two
rk
se
c
u
rit
y
.
Dr
.
B
a
la
c
h
a
n
d
r
a
’s
re
se
a
rc
h
a
re
a
in
c
lu
d
e
s
Ne
tw
o
rk
S
e
c
u
rit
y
,
A
lg
o
rit
h
m
s,
a
n
d
Op
e
ra
ti
n
g
sy
ste
m
s.
He
h
a
s
m
o
re
th
a
n
3
0
p
u
b
l
ica
ti
o
n
s
in
n
a
ti
o
n
a
l
a
n
d
in
ter
n
a
ti
o
n
a
l
c
o
n
f
e
re
n
c
e
s/jo
u
rn
a
ls.
Cu
rre
n
tl
y
h
e
is
w
o
rk
in
g
a
s
th
e
P
ro
f
e
ss
o
r
a
n
d
He
a
d
in
th
e
De
p
t.
o
f
In
f
o
rm
a
ti
o
n
&
Co
m
m
u
n
ica
ti
o
n
T
e
c
h
n
o
lo
g
y
,
M
a
n
ip
a
l
I
n
stit
u
te
o
f
T
e
c
h
n
o
lo
g
y
,
M
a
n
ip
a
l.
He
h
a
s
2
5
y
e
a
rs
o
f
tea
c
h
in
g
e
x
p
e
rien
c
e
in
v
a
rio
u
s In
stit
u
tes
.
Ar
sh
i
a
Ana
n
d
is
c
u
rre
n
tl
y
p
u
rsu
in
g
h
e
r
b
a
c
h
e
l
o
r’s
d
e
g
re
e
in
Co
m
p
u
ter
a
n
d
C
o
m
m
u
n
ica
ti
o
n
f
ro
m
M
IT
,
M
a
n
ip
a
l.
He
r
a
re
a
s o
f
in
tere
st i
n
c
l
u
d
e
Da
ta S
c
ien
c
e
a
n
d
F
u
ll
S
tac
k
De
v
e
lo
p
m
e
n
t.
S
u
sha
n
t
K
u
m
a
r
is
c
u
rre
n
tl
y
p
u
rsu
in
g
h
is
b
a
c
h
e
l
o
r'
s
d
e
g
re
e
in
Co
m
p
u
ter
a
n
d
Co
m
m
u
n
ica
ti
o
n
En
g
in
e
e
rin
g
f
ro
m
M
I
T
M
a
n
ip
a
l.
His
a
re
a
s
o
f
in
tere
sts
a
r
e
Da
ta
S
c
ien
c
e
a
n
d
f
u
ll
-
sta
c
k
d
e
v
e
lo
p
m
e
n
t.
S
u
sha
n
t
Pra
b
h
u
is
c
u
rre
n
tl
y
p
u
rsu
in
g
B.
T
e
c
h
in
Co
m
p
u
ter
a
n
d
C
o
m
m
u
n
ica
ti
o
n
T
e
c
h
n
o
l
o
g
y
(CCE)
a
t
M
IT
,
M
a
n
ip
a
l.
Cu
rre
n
tl
y
w
o
rk
in
g
in
De
e
p
Lea
rn
in
g
(G
r
a
p
h
Em
b
e
d
d
in
g
Ne
t
w
o
rk
s).
In
tere
ste
d
in
a
re
a
s o
f
De
e
p
L
e
a
rn
in
g
,
Da
ta S
c
ien
c
e
,
A
n
a
l
y
ti
c
s a
n
d
it
s a
p
p
li
c
a
ti
o
n
s.
Evaluation Warning : The document was created with Spire.PDF for Python.