I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
26
,
p
p
.
2398
~
2
4
0
9
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
3
.
p
p
2
3
9
8
-
2
4
0
9
2398
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Ada
ptive
sy
nthe
ti
c
-
ba
sed a
rrhy
th
m
ia
clas
sifica
tion u
sing
ma
chine learning
techniqu
es
M
d.
Ra
biu
l Is
la
m
1
,
T
a
pa
n K
um
a
r
G
o
dd
er
2
,
Ahs
a
n
-
Ul
-
Am
bia
3
,
F
er
dib
-
Al
-
I
s
la
m
4
,
B
ul
bu
l A
ha
m
ed
5
,
Anin
dy
a
Na
g
4
,
Arif
ul I
s
la
m
1
1
D
e
p
a
r
t
me
n
t
o
f
B
i
o
m
e
d
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
Te
c
h
n
o
l
o
g
y
,
I
sl
a
mi
c
U
n
i
v
e
r
si
t
y
,
K
u
sh
t
i
a
,
B
a
n
g
l
a
d
e
sh
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
Te
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
Te
c
h
n
o
l
o
g
y
,
I
sl
a
mi
c
U
n
i
v
e
r
si
t
y
,
K
u
sh
t
i
a
,
B
a
n
g
l
a
d
e
sh
3
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
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
I
sl
a
m
i
c
U
n
i
v
e
r
s
i
t
y
,
K
u
s
h
t
i
a
,
B
a
n
g
l
a
d
e
s
h
4
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
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
r
i
n
g
,
N
o
r
t
h
e
r
n
U
n
i
v
e
r
si
t
y
o
f
B
u
s
i
n
e
ss
a
n
d
Te
c
h
n
o
l
o
g
y
K
h
u
l
n
a
,
K
h
u
l
n
a
,
B
a
n
g
l
a
d
e
sh
5
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
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
r
i
n
g
,
S
o
n
a
r
g
a
o
n
U
n
i
v
e
r
s
i
t
y
,
D
h
a
k
a
,
B
a
n
g
l
a
d
e
s
h
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
14
,
2
0
2
5
R
ev
is
ed
Mar
31
,
2
0
2
6
Acc
ep
ted
Ap
r
22
,
2
0
2
6
Ca
rd
iac
a
rrh
y
t
h
m
ias
,
c
h
a
ra
c
teriz
e
d
b
y
irreg
u
lar
h
e
a
rt
r
h
y
t
h
m
s,
p
o
se
a
sig
n
ifi
c
a
n
t
c
h
a
ll
e
n
g
e
fo
r
ti
m
e
ly
a
n
d
a
c
c
u
ra
te
d
iag
n
o
sis.
Th
is
p
a
p
e
r
p
re
se
n
ts
a
n
a
d
v
a
n
c
e
d
fra
m
e
wo
rk
fo
r
a
rrh
y
th
m
ia
c
las
sifica
ti
o
n
th
a
t
a
d
d
re
ss
e
s
th
e
c
las
s
imb
a
lan
c
e
issu
e
u
sin
g
a
d
a
p
ti
v
e
sy
n
th
e
ti
c
(AD
ASYN)
sa
m
p
li
n
g
c
o
m
b
in
e
d
with
a
wid
e
-
ra
n
g
i
n
g
se
t
o
f
m
a
c
h
in
e
lea
rn
in
g
(
ML
)
tec
h
n
iq
u
e
s.
N
u
m
e
ro
u
s
c
las
sifiers
a
re
i
m
p
lem
e
n
ted
,
in
c
l
u
d
i
n
g
l
o
g
isti
c
re
g
re
ss
io
n
(LR)
,
n
a
iv
e
Ba
y
e
s
(NB),
d
e
c
isio
n
tree
(DT),
ra
n
d
o
m
fo
re
st
(RF
),
k
-
n
e
a
re
st
n
e
ig
h
b
o
rs
(KN
N),
m
u
lt
i
-
lay
e
r
p
e
rc
e
p
tro
n
(M
L
P
),
g
ra
d
ien
t
b
o
o
stin
g
(G
B),
a
d
a
p
ti
v
e
b
o
o
stin
g
(AB),
li
g
h
t
g
ra
d
ien
t
b
o
o
sti
n
g
m
a
c
h
in
e
(LG
BM
),
a
n
d
e
x
trem
e
g
ra
d
ien
t
b
o
o
st
in
g
(XG
B).
Th
e
e
x
p
e
rime
n
t
a
l
re
su
lt
s
d
e
m
o
n
stra
te
t
h
a
t
RF
,
G
B,
LG
BM
,
a
n
d
XG
B
a
c
h
iev
e
d
a
re
m
a
rk
a
b
le
a
c
c
u
ra
c
y
o
f
9
7
%
.
Ot
h
e
r
m
o
d
e
ls
,
su
c
h
a
s
DT
a
n
d
LR,
a
lso
a
c
c
o
m
p
li
sh
e
d
we
ll
,
a
c
h
iev
in
g
9
5
%
a
n
d
9
4
%
a
c
c
u
ra
c
y
,
re
sp
e
c
ti
v
e
ly
.
KN
N
a
n
d
NB
y
iel
d
e
d
9
3
%
a
n
d
8
1
%
a
c
c
u
ra
c
y
,
re
sp
e
c
ti
v
e
ly
,
wh
il
e
AB
u
n
d
e
rp
e
rfo
rm
e
d
wi
th
a
n
a
c
c
u
ra
c
y
o
f
2
4
%
.
P
re
c
isio
n
sc
o
re
s
a
c
ro
ss
th
e
m
o
d
e
ls
re
m
a
in
e
d
h
i
g
h
,
e
x
c
e
p
t
fo
r
NB
a
n
d
AB.
All
m
o
d
e
ls,
e
x
c
e
p
t
AB,
d
e
m
o
n
stra
ted
e
x
c
e
ll
e
n
t
re
c
a
ll
.
Th
e
p
ro
p
o
se
d
m
e
th
o
d
o
l
o
g
y
o
u
tp
e
rfo
rm
s
p
re
v
io
u
s
wo
r
k
s,
se
tt
i
n
g
a
n
e
w
b
e
n
c
h
m
a
rk
fo
r
a
rr
h
y
t
h
m
ia
c
las
sifica
ti
o
n
.
Th
e
se
o
u
tco
m
e
s
u
n
d
e
rsc
o
re
t
h
e
e
ffe
c
ti
v
e
n
e
ss
o
f
i
n
teg
ra
ti
n
g
AD
ASYN
wit
h
ML
tec
h
n
iq
u
e
s
to
e
n
h
a
n
c
e
a
rrh
y
th
m
ia
d
e
tec
ti
o
n
,
with
sig
n
ifi
c
a
n
t
p
o
ten
ti
a
l
fo
r
imp
r
o
v
i
n
g
c
li
n
ica
l
d
iag
n
o
stic
p
ro
c
e
ss
e
s a
n
d
p
a
ti
e
n
t
o
u
tco
m
e
s.
K
ey
w
o
r
d
s
:
Ad
ap
tiv
e
s
y
n
th
etic
Ar
r
h
y
th
m
ia
class
if
icatio
n
C
las
s
im
b
alan
ce
Gr
id
Sear
ch
C
V
Ma
ch
in
e
lear
n
in
g
Pair
ed
t
-
test
Statis
t
ical
s
ig
n
if
ican
ce
test
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
:
Fer
d
ib
-
Al
-
I
s
lam
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
E
n
g
in
ee
r
in
g
,
Facu
lty
o
f
Scien
ce
an
d
E
n
g
i
n
ee
r
in
g
No
r
th
er
n
Un
i
v
er
s
ity
o
f
B
u
s
in
e
s
s
an
d
T
ec
h
n
o
lo
g
y
Kh
u
ln
a
Kh
u
ln
a
–
9
1
0
0
,
B
an
g
lad
esh
E
m
ail: f
er
d
ib
.
b
s
m
r
s
tu
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Ar
r
h
y
th
m
ia
is
a
co
m
m
o
n
c
ar
d
iac
d
is
o
r
d
er
th
at
af
f
ec
ts
m
illi
o
n
s
o
f
p
eo
p
le
wo
r
ld
wid
e.
I
t
is
ch
ar
ac
ter
ized
b
y
an
ab
n
o
r
m
al
h
ea
r
t r
h
y
th
m
,
wh
ich
ca
n
b
e
lif
e
-
th
r
ea
ten
in
g
if
u
n
d
iag
n
o
s
ed
an
d
u
n
tr
ea
ted
.
E
ar
ly
d
etec
tio
n
an
d
class
if
icatio
n
o
f
ar
r
h
y
th
m
ia
ar
e
cr
itical
f
o
r
a
ctiv
e
tr
ea
tm
en
t
an
d
co
n
tr
o
llin
g
o
f
th
e
co
n
d
itio
n
.
T
h
e
tr
ad
itio
n
al
ar
r
h
y
th
m
ia
cla
s
s
if
icatio
n
m
eth
o
d
e
lectr
o
ca
r
d
io
g
r
am
(
E
C
G)
is
a
n
o
n
-
in
v
asi
v
e
tech
n
iq
u
e
wid
ely
u
s
ed
to
d
iag
n
o
s
e
an
d
class
if
y
ar
r
h
y
th
m
ia
[
1
]
.
T
h
ese
ty
p
es
o
f
m
an
u
al
a
p
p
r
o
ac
h
es
ar
e
ti
m
e
-
co
n
s
u
m
in
g
an
d
r
eq
u
ir
e
s
p
ec
ialized
m
ed
ical
k
n
o
wled
g
e.
H
o
wev
er
,
th
e
d
e
m
an
d
s
ar
e
ac
cu
r
ate
an
d
ef
f
i
cien
t
d
iag
n
o
s
is
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
d
a
p
tive
s
yn
th
etic
-
b
a
s
ed
a
r
r
h
yth
mia
cla
s
s
ifica
tio
n
u
s
in
g
ma
ch
in
e
lea
r
n
in
g
tec
h
n
iq
u
es (
Md
.
R
a
b
iu
l
I
s
la
m)
2399
class
if
icatio
n
o
f
ar
r
h
y
th
m
ias.
Ar
r
h
y
th
m
ia
is
a
ca
r
d
iac
co
m
p
lain
t
co
n
s
id
er
ed
b
y
ir
r
eg
u
l
ar
h
ea
r
tb
ea
ts
,
wh
ich
ca
n
lead
to
s
ev
er
e
im
p
ed
im
en
ts
s
u
ch
as
s
tr
o
k
e,
h
ea
r
t
f
ailu
r
e
,
o
r
s
u
d
d
en
ca
r
d
iac
ar
r
est.
Ar
r
h
y
th
m
ias
m
u
s
t
b
e
id
en
tifie
d
an
d
ca
teg
o
r
ize
d
ea
r
l
y
to
b
e
e
f
f
ec
tiv
ely
tr
ea
te
d
an
d
m
an
ag
ed
.
Ma
ch
in
e
lear
n
in
g
(
ML
)
alg
o
r
ith
m
s
s
h
o
w
g
r
ea
t
p
o
ten
tial
f
o
r
ar
r
h
y
th
m
ia
class
if
icatio
n
,
b
u
t
th
eir
p
er
f
o
r
m
an
ce
d
ep
e
n
d
s
o
n
t
h
e
ass
o
r
tm
en
t
o
f
in
f
o
r
m
ativ
e
f
ea
tu
r
es.
R
ec
u
r
s
iv
e
f
ea
t
u
r
e
elim
in
atio
n
(
R
FE)
is
a
tech
n
iq
u
e
th
at
ca
n
b
e
ap
p
lied
to
s
elec
t
ap
p
r
o
p
r
iate
f
ea
tu
r
es
f
o
r
class
if
icatio
n
,
a
n
d
t
h
is
s
tu
d
y
aim
s
to
ex
am
i
n
e
th
e
p
er
f
o
r
m
an
ce
o
f
s
ev
er
al
ML
m
o
d
els
u
s
in
g
th
is
tech
n
iq
u
e.
T
h
e
d
ev
elo
p
m
en
t
o
f
r
e
liab
le
an
d
ac
cu
r
ate
au
to
m
ated
ar
r
h
y
t
h
m
ia
class
if
icatio
n
s
y
s
tem
s
ca
n
aid
in
th
e
d
iag
n
o
s
is
an
d
m
a
n
ag
em
e
n
t
o
f
a
r
r
h
y
th
m
ias,
u
ltima
tely
im
p
r
o
v
in
g
p
atien
t
o
u
tco
m
es.
ML
m
o
d
els
h
av
e
s
h
o
wn
g
r
ea
t
p
o
ten
tial
in
th
e
class
if
icatio
n
o
f
ar
r
h
y
th
m
ia
[
2
]
.
H
o
wev
er
,
th
e
p
er
f
o
r
m
an
ce
o
f
th
ese
alg
o
r
ith
m
s
d
ep
en
d
s
o
n
th
e
s
elec
tio
n
o
f
in
f
o
r
m
ativ
e
f
ea
tu
r
es.
R
FE
is
a
p
r
o
ce
d
u
r
e
t
h
at
ca
n
b
e
u
s
ed
to
s
elec
t r
elev
an
t f
ea
tu
r
es f
o
r
class
if
icatio
n
[
3
]
,
[
4
]
.
T
h
e
r
esear
ch
e
r
s
p
r
o
p
o
s
e
an
ar
r
h
y
th
m
ia
class
if
icatio
n
p
r
o
ce
d
u
r
e
b
ased
o
n
s
y
n
th
etic
m
in
o
r
ity
o
v
er
-
s
am
p
lin
g
tech
n
iq
u
e
(
S
MO
T
E
)
an
d
f
ea
tu
r
e
s
elec
tio
n
[
5
]
.
T
h
e
s
tu
d
y
aim
s
to
ad
d
r
ess
th
e
co
n
ce
r
n
o
f
im
b
alan
ce
d
d
ata
in
E
C
G
ar
r
h
y
th
m
ia
class
if
icatio
n
,
wh
ich
ca
n
lead
to
lo
w
class
if
icatio
n
ac
cu
r
ac
y
f
o
r
m
in
o
r
ity
class
es.
T
h
e
p
r
o
p
o
s
ed
p
r
o
ce
d
u
r
e
ap
p
lies
SMOT
E
to
p
r
o
d
u
ce
s
y
n
th
etic
s
am
p
les
to
b
alan
ce
th
e
d
ata,
an
d
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
s
to
s
elec
t
th
e
m
o
s
t
in
f
o
r
m
ativ
e
f
e
atu
r
es
f
o
r
class
if
icatio
n
.
T
h
e
s
tu
d
y
d
em
o
n
s
tr
ates
th
e
p
o
te
n
tial
o
f
u
s
in
g
SMOT
E
an
d
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
es
f
o
r
ac
cu
r
ate
an
d
ef
f
icien
t
E
C
G
ar
r
h
y
th
m
ia
class
if
icatio
n
[
5
]
.
Ar
r
h
y
th
m
ia
class
if
icatio
n
u
s
in
g
d
ata
f
r
o
m
th
e
UC
I
r
v
in
e
Ma
ch
in
e
L
ea
r
n
in
g
(
UC
I
ML
)
r
ep
o
s
ito
r
y
h
as
b
ee
n
ex
ten
s
iv
ely
s
tu
d
ied
with
v
ar
io
u
s
ML
tech
n
iq
u
es
to
en
h
a
n
ce
d
ia
g
n
o
s
tic
ac
cu
r
ac
y
a
n
d
ef
f
icien
cy
.
T
h
e
class
if
icatio
n
o
f
E
C
G
s
ig
n
als,
wh
ich
ar
e
cr
u
cial
f
o
r
d
iag
n
o
s
in
g
ar
r
h
y
th
m
ia,
h
as
b
ee
n
s
ig
n
if
ican
tly
en
h
an
ce
d
u
s
in
g
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NN
s
)
,
ac
h
iev
in
g
a
h
ig
h
a
cc
u
r
ac
y
o
f
9
9
.
8
8
%
with
an
en
s
em
b
le
o
f
d
ep
t
h
-
w
is
e
s
ep
ar
ab
le
co
n
v
o
l
u
tio
n
al
(
DSC
)
n
eu
r
al
n
etwo
r
k
s
[
6
]
.
R
an
d
o
m
f
o
r
est
(
RF
)
class
if
ier
s
h
av
e
also
s
h
o
wn
p
r
o
m
is
in
g
r
esu
lts
,
p
ar
ticu
lar
l
y
in
class
if
y
in
g
atr
ial
f
ib
r
illatio
n
(
AF)
an
d
o
th
er
h
ea
r
t r
h
y
th
m
s
,
with
an
ac
c
u
r
a
cy
o
f
8
9
%
[
7
]
.
C
o
m
p
ar
ativ
e
s
tu
d
ies
in
d
icate
th
at
R
F
p
er
f
o
r
m
s
b
est
f
o
r
a
r
r
h
y
th
m
ia
d
atasets
,
o
u
tp
er
f
o
r
m
in
g
o
th
er
class
if
ier
s
lik
e
n
aiv
e
B
ay
es
(
NB
)
,
m
u
lti
-
lay
er
p
e
r
ce
p
tr
o
n
(
MLP
)
,
an
d
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM
)
[
8
]
.
Ad
d
itio
n
ally
,
e
n
s
em
b
le
m
eth
o
d
s
co
m
b
in
in
g
class
if
ier
s
s
u
ch
as
d
ec
is
io
n
tr
ee
(
DT
)
,
S
VM
,
an
d
R
F
h
av
e
d
em
o
n
s
tr
ated
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
o
n
ce
r
tain
d
atasets
[
9
]
.
A
d
v
an
ce
d
tech
n
iq
u
es
lik
e
th
e
r
id
er
ch
a
o
tic
b
io
g
r
ap
h
y
o
p
tim
izatio
n
(
R
C
B
O)
alg
o
r
ith
m
in
teg
r
ated
with
d
ee
p
s
tack
ed
au
to
-
en
co
d
er
s
h
av
e
b
ee
n
p
r
o
p
o
s
ed
to
h
an
d
le
lar
g
e
d
atasets
ef
f
ec
tiv
e
ly
,
ac
h
iev
in
g
h
ig
h
ac
cu
r
ac
y
an
d
s
p
ec
if
icity
[
1
0
]
.
M
o
r
eo
v
e
r
,
t
h
e
u
s
e
o
f
p
r
in
ci
p
al
co
m
p
o
n
en
t
a
n
aly
s
is
(
PC
A)
f
o
r
r
e
d
u
cin
g
f
ea
tu
r
es
a
n
d
wav
elet
p
ac
k
et
d
ec
o
m
p
o
s
itio
n
(
W
PD)
f
o
r
f
ea
t
u
r
e
ex
tr
ac
tio
n
h
as
b
ee
n
b
e
n
ef
ici
al
in
im
p
r
o
v
i
n
g
class
if
icatio
n
ac
cu
r
ac
y
[
1
1
]
.
T
h
e
in
te
g
r
atio
n
o
f
ML
with
h
ar
d
war
e
-
s
o
f
twar
e
c
o
m
p
lex
es
,
s
u
ch
as
m
o
b
ile
elec
tr
o
n
ic
ca
r
d
io
g
r
a
p
h
s
,
h
as
en
ab
led
r
ea
l
-
t
im
e
E
C
G
an
aly
s
is
an
d
ar
r
h
y
th
m
ia
d
etec
tio
n
,
f
u
r
t
h
er
en
h
an
cin
g
d
ia
g
n
o
s
tic
ca
p
ab
ilit
ies
[
1
2
]
.
T
ec
h
n
iq
u
es
lik
e
g
en
etic
alg
o
r
ith
m
an
d
ar
tific
ial
f
lo
r
a
o
p
tim
izati
o
n
h
av
e
also
b
ee
n
em
p
l
o
y
ed
to
o
p
tim
ize
n
eu
r
al
n
etwo
r
k
to
p
o
lo
g
ies,
r
ed
u
cin
g
o
v
er
f
itti
n
g
a
n
d
im
p
r
o
v
i
n
g
g
en
er
aliza
tio
n
[
1
3
]
.
L
astl
y
,
teac
h
in
g
-
lear
n
in
g
-
b
ased
o
p
tim
izatio
n
(
T
L
B
O)
co
m
b
in
ed
with
k
-
m
ea
n
s
a
n
d
f
u
zz
y
c
-
m
ea
n
s
clu
s
ter
in
g
h
as
s
h
o
wn
h
ig
h
ac
cu
r
ac
y
in
class
if
y
in
g
h
ea
r
t
d
is
ea
s
e
d
atasets
,
in
d
icatin
g
th
e
r
o
b
u
s
t
n
ess
o
f
th
ese
m
eth
o
d
s
in
h
an
d
lin
g
n
o
n
-
cl
u
s
ter
ed
d
ata
[
1
4
]
.
T
h
ese
ad
v
an
ce
m
en
ts
co
llectiv
ely
co
n
tr
i
b
u
te
to
th
e
ef
f
ec
tiv
e
class
if
icatio
n
o
f
ar
r
h
y
th
m
ia
u
s
in
g
UC
I
M
L
r
ep
o
s
ito
r
y
d
ata,
h
ig
h
lig
h
tin
g
th
e
p
r
o
s
p
ec
tiv
e
o
f
ML
in
h
ea
lth
ca
r
e
d
iag
n
o
s
tics
.
Dee
p
lear
n
in
g
n
o
w
d
o
m
in
ate
s
E
C
G
ar
r
h
y
th
m
ia
class
if
icat
io
n
.
C
NNs
lear
n
h
ier
ar
ch
ical
tem
p
o
r
al
f
ea
tu
r
es
d
ir
ec
tly
f
r
o
m
r
aw
b
ea
ts
o
r
f
ix
ed
-
len
g
th
win
d
o
ws
an
d
s
et
ea
r
ly
s
tate
-
of
-
th
e
-
ar
t
(
SOTA
)
on
Ma
s
s
ac
h
u
s
etts
I
n
s
t
itu
te
o
f
T
ec
h
n
o
lo
g
y
–
B
eth
I
s
r
ae
l
Ho
s
p
i
tal
(
MI
T
-
B
I
H
)
an
d
r
elate
d
c
o
r
p
o
r
a
[
1
5
]
,
[
1
6
]
.
R
ec
u
r
r
en
t
m
o
d
els
(
e.
g
.
,
b
i
d
ir
e
ctio
n
al
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
B
iLST
M
)
)
ca
p
tu
r
e
l
o
n
g
e
r
-
r
an
g
e
d
ep
en
d
en
cies
ac
r
o
s
s
b
ea
ts
an
d
o
f
ten
b
o
o
s
t
m
in
o
r
ity
-
class
s
en
s
itiv
ity
wh
en
th
e
in
p
u
t
is
f
r
am
ed
as
a
s
e
q
u
en
ce
[
1
7
]
.
Mo
r
e
r
ec
en
tly
,
tr
a
n
s
f
o
r
m
er
ar
ch
itect
u
r
es
m
o
d
el
g
lo
b
al
co
n
tex
t
v
ia
s
elf
-
atten
tio
n
o
n
co
n
tin
u
o
u
s
E
C
G
s
eg
m
en
ts
,
with
s
tr
o
n
g
r
esu
lts
f
o
r
r
h
y
th
m
d
ete
ctio
n
(
in
clu
d
in
g
AF)
with
o
u
t
h
an
d
-
c
r
af
ted
s
eg
m
e
n
tatio
n
[
1
8
]
,
[
1
9
]
.
At
clin
ical
s
ca
le,
d
ee
p
n
etwo
r
k
s
h
av
e
m
atch
ed
o
r
ex
ce
ed
ed
ca
r
d
io
lo
g
is
t
p
er
f
o
r
m
an
ce
f
o
r
s
in
g
le
-
lead
r
h
y
th
m
class
if
icatio
n
o
n
am
b
u
lato
r
y
E
C
Gs
[
2
0
]
an
d
g
en
er
alize
d
t
o
1
2
-
lea
d
clin
ical
E
C
Gs
tr
ain
ed
o
n
m
illi
o
n
s
o
f
lab
eled
ex
am
s
,
ac
h
ie
v
in
g
h
ig
h
F1
an
d
>9
9
% sp
ec
if
icity
o
n
m
u
ltip
le
ab
n
o
r
m
alities
[
2
1
]
.
R
eg
ar
d
in
g
r
ea
l
-
w
o
r
ld
in
te
g
r
at
io
n
,
lar
g
e
p
r
ag
m
atic
s
tu
d
ies
s
h
o
w
f
ea
s
ib
ilit
y
an
d
u
tili
ty
.
T
h
e
Ap
p
le
Hea
r
t
Stu
d
y
(
n
≈
4
1
9
k
)
v
alid
ated
s
m
ar
twatch
ir
r
eg
u
lar
-
p
u
l
s
e
n
o
tific
atio
n
s
with
a
p
o
s
itiv
e
p
r
e
d
ictiv
e
v
alu
e
~8
4
%
f
o
r
AF
wh
en
co
n
f
ir
m
e
d
s
im
u
ltan
eo
u
s
ly
b
y
E
C
G
p
atch
m
o
n
ito
r
i
n
g
[
2
2
]
.
I
n
r
o
u
tin
e
p
r
im
ar
y
ca
r
e,
an
AI
-
E
C
G
d
ec
is
io
n
-
s
u
p
p
o
r
t
to
o
l
em
b
ed
d
ed
in
th
e
elec
tr
o
n
ic
h
ea
lth
r
ec
o
r
d
(
E
HR
)
in
cr
ea
s
e
d
n
ew
d
iag
n
o
s
es
o
f
lo
w
ejec
tio
n
f
r
ac
tio
n
in
a
clu
s
ter
-
r
an
d
o
m
ized
tr
ial
(
2
2
,
6
4
1
ad
u
lts
)
[
2
3
]
.
E
ar
lier
wo
r
k
also
d
em
o
n
s
tr
ated
AI
-
E
C
G
s
cr
ee
n
in
g
f
o
r
s
y
s
to
lic
d
y
s
f
u
n
ctio
n
u
s
in
g
s
tan
d
a
r
d
1
2
-
lea
d
E
C
Gs
in
clin
ical
wo
r
k
f
lo
ws
[
2
4
]
.
T
o
g
eth
er
,
th
ese
tr
e
n
d
s
m
o
tiv
at
e
th
e
f
o
cu
s
o
n
r
o
b
u
s
t
im
b
ala
n
ce
h
an
d
lin
g
(
ad
a
p
tiv
e
s
y
n
t
h
etic
(
ADASYN
)
)
an
d
en
s
em
b
le
lear
n
er
s
as
a
co
m
p
lem
en
tar
y
,
co
s
t
-
ef
f
icie
n
t
av
en
u
e
wh
er
e
d
ee
p
m
o
d
els
ar
e
n
o
t
y
et
d
ep
lo
y
ed
o
r
wh
er
e
d
ataset
s
ize,
lab
elin
g
co
s
t,
o
r
co
m
p
u
te
ar
e
co
n
s
tr
ain
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
26
:
2
3
9
8
-
2
4
0
9
2400
T
h
e
p
r
e
c
i
s
e
a
n
d
p
r
o
m
p
t
d
e
t
e
c
tio
n
o
f
a
r
r
h
y
t
h
m
i
a
s
i
s
f
u
n
d
a
m
e
n
t
a
l
f
o
r
e
f
f
e
c
t
i
v
e
cl
i
n
i
c
al
i
n
t
e
r
f
er
e
n
c
e
,
y
e
t
i
t
r
e
m
ai
n
s
a
s
t
i
m
u
la
t
i
n
g
t
a
s
k
d
u
e
t
o
t
h
e
c
o
m
p
l
e
x
a
n
d
v
a
r
i
ed
n
a
t
u
r
e
o
f
a
r
r
h
y
t
h
m
i
c
p
a
tt
e
r
n
s
i
n
E
C
G
s
i
g
n
a
ls
.
T
r
a
d
i
t
i
o
n
a
l
d
i
a
g
n
o
s
t
ic
m
e
t
h
o
d
s
a
r
e
n
o
t
o
n
l
y
ti
m
e
-
c
o
n
s
u
m
i
n
g
a
n
d
s
u
b
j
e
c
t
t
o
h
u
m
a
n
e
r
r
o
r
b
u
t
a
l
s
o
s
t
r
u
g
g
l
e
w
it
h
t
h
e
c
l
a
s
s
i
m
b
a
l
a
n
c
e
p
r
e
v
a
le
n
t
i
n
a
r
r
h
y
t
h
m
i
a
d
a
t
a
s
e
ts
,
w
h
e
r
e
c
e
r
t
a
i
n
t
y
p
e
s
o
f
a
r
r
h
y
t
h
m
i
as
a
r
e
s
i
g
n
i
f
ic
a
n
t
l
y
u
n
d
e
r
r
e
p
r
e
s
e
n
t
e
d
.
T
h
is
i
m
b
a
la
n
c
e
c
a
n
s
e
v
e
r
e
l
y
i
m
p
a
i
r
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
ML
m
o
d
e
ls
,
l
e
a
d
i
n
g
t
o
b
i
a
s
ed
p
r
e
d
i
c
t
i
o
n
s
a
n
d
p
o
o
r
g
e
n
e
r
a
l
i
za
t
i
o
n
.
T
h
e
r
e
f
o
r
e
,
t
h
e
r
e
is
a
p
r
es
s
i
n
g
n
e
e
d
f
o
r
a
d
v
a
n
c
e
d
c
o
m
p
u
t
a
t
i
o
n
a
l
a
p
p
r
o
a
c
h
es
t
h
a
t
c
a
n
e
f
f
e
c
ti
v
e
l
y
h
a
n
d
l
e
i
m
b
a
l
a
n
c
e
d
d
a
t
a
a
n
d
l
e
v
e
r
a
g
e
t
h
e
p
o
w
e
r
o
f
M
L
f
o
r
r
o
b
u
s
t
a
r
r
h
y
t
h
m
i
a
c
l
as
s
i
f
i
c
a
ti
o
n
.
B
y
i
n
t
e
g
r
at
i
n
g
AD
A
SY
N
a
n
d
a
c
o
m
p
r
e
h
e
n
s
i
v
e
s
u
i
te
o
f
M
L
m
o
d
e
l
s
,
t
h
is
w
o
r
k
a
i
m
s
t
o
o
v
e
r
c
o
m
e
t
h
es
e
c
h
a
l
l
e
n
g
e
s
,
p
r
o
v
i
d
i
n
g
a
m
o
r
e
p
r
e
c
i
s
e
,
r
el
i
a
b
le
,
a
n
d
e
f
f
i
ci
e
n
t
s
o
l
u
t
i
o
n
f
o
r
a
r
r
h
y
t
h
m
ia
d
e
t
e
ct
i
o
n
.
T
h
i
s
n
o
t
o
n
l
y
b
o
o
s
t
s
i
n
d
i
c
a
ti
v
e
ac
c
u
r
a
c
y
b
u
t
a
l
s
o
h
as
t
h
e
p
o
t
e
n
t
i
a
l
t
o
c
o
n
s
i
d
e
r
a
b
l
y
i
m
p
r
o
v
e
p
a
ti
e
n
t
c
o
n
c
l
u
s
i
o
n
s
b
y
f
a
ci
l
it
a
ti
n
g
t
i
m
e
l
y
a
n
d
a
p
p
r
o
p
r
i
a
t
e
m
e
d
i
ca
l
i
n
t
e
r
v
e
n
t
i
o
n
.
T
h
e
m
ai
n
o
b
je
c
t
i
v
e
o
f
t
h
e
r
es
e
a
r
c
h
i
s
t
o
d
el
i
v
e
r
s
o
m
e
e
f
f
e
c
ti
v
e
t
e
c
h
n
i
q
u
e
s
f
o
r
t
h
e
cl
a
s
s
i
f
ic
a
t
i
o
n
o
f
a
r
r
h
y
t
h
m
i
a
u
s
i
n
g
s
e
v
e
r
a
l
M
L
m
o
d
e
l
s
,
i
n
t
e
r
m
s
o
f
a
c
c
u
r
ac
y
a
n
d
s
p
e
e
d
.
T
h
e
m
a
i
n
c
o
n
t
r
i
b
u
t
i
o
n
s
o
f
t
h
is
s
y
s
t
e
m
a
r
e
—
i
n
t
r
o
d
u
c
e
a
n
o
v
e
l
m
e
t
h
o
d
f
o
r
g
e
n
e
r
a
t
i
n
g
s
y
n
t
h
e
t
i
c
d
a
t
a
t
a
i
l
o
r
e
d
s
p
e
c
i
f
i
ca
l
l
y
f
o
r
a
r
r
h
y
t
h
m
i
a
c
l
as
s
i
f
i
c
at
i
o
n
t
as
k
s
.
B
y
l
e
v
e
r
a
g
i
n
g
a
d
a
p
t
i
v
e
t
e
c
h
n
i
q
u
es
,
t
h
e
a
p
p
r
o
a
c
h
e
n
s
u
r
e
s
t
h
a
t
t
h
e
s
y
n
t
h
et
i
c
d
a
t
a
c
l
o
s
e
l
y
r
e
s
e
m
b
l
es
r
e
al
-
w
o
r
l
d
a
r
r
h
y
t
h
m
i
a
p
a
tt
e
r
n
s
,
t
h
e
r
e
b
y
e
n
h
a
n
c
i
n
g
t
h
e
r
o
b
u
s
t
n
es
s
a
n
d
g
e
n
e
r
a
l
i
z
a
ti
o
n
p
r
o
f
i
c
i
e
n
c
y
o
f
th
e
c
l
a
s
s
i
f
i
c
a
ti
o
n
m
o
d
e
l
.
E
v
a
l
u
at
e
a
r
a
n
g
e
o
f
S
O
T
A
M
L
m
o
d
e
ls
.
T
h
r
o
u
g
h
r
i
g
o
r
o
u
s
e
x
p
e
r
i
m
e
n
t
a
t
i
o
n
a
n
d
c
o
m
p
a
r
a
ti
v
e
a
n
a
l
y
s
is
,
t
h
e
m
o
s
t
s
u
i
t
a
b
l
e
a
l
g
o
r
i
t
h
m
(
s
)
i
s
i
d
e
n
t
i
f
i
e
d
f
o
r
a
c
h
i
e
v
i
n
g
o
p
t
i
m
a
l
c
l
a
s
s
i
f
ic
a
t
i
o
n
a
c
c
u
r
a
c
y
a
n
d
p
e
r
f
o
r
m
a
n
c
e
.
C
o
n
d
u
c
t
c
o
m
p
r
e
h
e
n
s
i
v
e
e
x
p
e
r
i
m
e
n
ts
u
s
i
n
g
b
e
n
c
h
m
a
r
k
a
r
r
h
y
t
h
m
i
a
d
a
t
a
s
et
s
t
o
e
m
p
i
r
i
c
al
l
y
v
a
l
i
d
ate
t
h
e
e
f
f
e
c
t
i
v
e
n
e
s
s
o
f
t
h
e
p
r
o
p
o
s
e
d
a
p
p
r
o
a
c
h
.
B
y
c
o
m
p
a
r
i
n
g
t
h
e
p
e
r
f
o
r
m
a
n
c
e
m
e
t
r
i
c
s
s
u
c
h
a
s
a
c
c
u
r
a
c
y
,
p
r
e
c
i
s
i
o
n
,
a
n
d
r
e
c
a
l
l
w
i
t
h
e
x
is
ti
n
g
m
e
t
h
o
d
s
,
t
h
e
s
u
p
e
r
i
o
r
i
t
y
o
f
t
h
e
a
p
p
r
o
a
c
h
i
s
d
e
m
o
n
s
t
r
a
t
e
d
i
n
a
cc
u
r
a
t
e
l
y
cl
a
s
s
i
f
y
i
n
g
d
i
f
f
e
r
e
n
t
t
y
p
es
o
f
a
r
r
h
y
th
m
i
a
s
.
T
h
e
co
n
tin
u
in
g
p
o
r
tio
n
s
o
f
p
a
p
er
ar
e
s
tr
u
ctu
r
ed
i
n
f
o
llo
win
g
s
ty
l
e.
S
ec
tio
n
2
o
u
tlin
es
th
e
ap
p
r
o
ac
h
u
s
ed
in
th
is
p
ap
er
.
Sectio
n
3
p
r
esen
ts
th
e
f
in
d
in
g
s
o
f
th
is
r
esear
ch
.
Sectio
n
4
d
eliv
er
s
th
e
c
o
n
clu
s
io
n
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
e
p
r
o
p
o
s
ed
s
y
s
tem
id
en
tifie
s
wh
eth
er
a
p
er
s
o
n
'
s
h
ea
r
t
i
s
h
ea
lth
y
o
r
s
u
f
f
er
in
g
f
r
o
m
a
r
r
h
y
th
m
ia
d
is
ea
s
e.
T
h
e
im
p
lem
en
tatio
n
o
f
th
is
s
y
s
tem
h
as
b
ee
n
d
o
n
e
b
y
u
s
in
g
Nu
m
Py
,
Pan
d
as,
ADASYN,
Gr
id
Sear
ch
C
V,
an
d
Scik
it
-
lear
n
lib
r
ar
y
.
Fro
m
Fig
u
r
e
1
,
it
c
an
b
e
s
ee
n
th
at,
a
r
o
b
u
s
t
f
r
am
ewo
r
k
is
d
ev
elo
p
ed
f
o
r
ar
r
h
y
th
m
ia
class
if
icatio
n
th
at
in
co
r
p
o
r
ates
s
ev
er
al
k
ey
s
tag
es,
in
clu
d
in
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
class
im
b
alan
ce
h
an
d
lin
g
,
a
n
d
ML
m
o
d
el
ap
p
licatio
n
.
I
n
i
tially
,
th
e
d
ataset
u
n
d
er
we
n
t
co
m
p
r
e
h
en
s
iv
e
p
r
ep
r
o
ce
s
s
in
g
to
h
an
d
le
m
is
s
in
g
v
alu
es,
co
n
f
ir
m
in
g
th
e
in
teg
r
ity
an
d
co
m
p
r
eh
e
n
s
iv
en
ess
o
f
th
e
d
ata.
Fo
llo
win
g
th
is
,
th
e
d
ata
was sp
lit in
to
tr
ain
in
g
an
d
test
in
g
s
e
ts
to
s
im
p
lify
m
o
d
el
ev
al
u
atio
n
.
Fig
u
r
e
1
.
W
o
r
k
f
lo
w
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
d
a
p
tive
s
yn
th
etic
-
b
a
s
ed
a
r
r
h
yth
mia
cla
s
s
ifica
tio
n
u
s
in
g
ma
ch
in
e
lea
r
n
in
g
tec
h
n
iq
u
es (
Md
.
R
a
b
iu
l
I
s
la
m)
2401
T
o
ad
d
r
ess
th
e
in
h
er
e
n
t
class
im
b
alan
ce
in
ar
r
h
y
th
m
ia
d
atasets
,
AD
ASYN
is
em
p
lo
y
ed
,
wh
ich
g
e
n
e
r
a
te
d
s
y
n
th
eti
c
s
a
m
p
les
f
o
r
u
n
d
e
r
r
e
p
r
ese
n
t
e
d
cl
ass
es
w
it
h
i
n
t
h
e
tr
ai
n
i
n
g
s
et
,
t
h
u
s
c
r
ea
ti
n
g
a
m
o
r
e
b
ala
n
c
e
d
d
at
ase
t.
Su
b
s
e
q
u
e
n
t
ly
,
a
d
i
v
e
r
s
e
a
r
r
a
y
o
f
M
L
m
o
d
els
was
a
p
p
lie
d
t
o
t
h
e
b
al
an
ce
d
d
atas
et
,
i
n
c
lu
d
i
n
g
lo
g
is
tic
r
e
g
r
ess
i
o
n
(
LR
)
,
NB
,
DT
,
R
F,
k
-
n
ea
r
est
n
e
ig
h
b
o
r
s
(
KN
N
)
,
ML
P,
g
r
a
d
i
en
t
b
o
o
s
ti
n
g
(
GB
)
,
a
d
ap
ti
v
e
b
o
o
s
t
in
g
(
AB
)
,
li
g
h
t
g
r
a
d
i
e
n
t
b
o
o
s
ti
n
g
m
a
c
h
i
n
e
(
L
GB
M
)
,
a
n
d
e
x
tr
e
m
e
g
r
a
d
ie
n
t
b
o
o
s
t
in
g
(
X
GB
)
.
E
ac
h
m
o
d
el'
s
p
e
r
f
o
r
m
a
n
c
e
was
r
ig
o
r
o
u
s
l
y
a
s
s
ess
e
d
u
s
i
n
g
ac
c
u
r
ac
y
,
p
r
e
cis
io
n
,
r
ec
all
,
F
1
-
s
c
o
r
e,
C
o
h
e
n
’
s
Ka
p
p
a
s
c
o
r
e
,
an
d
Ma
th
ew’
s
co
r
r
elatio
n
co
ef
f
ici
en
t
(
MCC
)
to
d
eter
m
in
e
th
ei
r
ef
f
ec
tiv
en
ess
in
class
if
y
in
g
ar
r
h
y
th
m
ias.
T
h
is
co
m
p
r
eh
e
n
s
iv
e
m
eth
o
d
o
l
o
g
y
en
s
u
r
ed
a
th
o
r
o
u
g
h
ev
alu
atio
n
o
f
d
if
f
er
en
t
ML
tech
n
iq
u
es
in
co
n
j
u
n
ctio
n
with
ad
v
an
ce
d
d
ata
p
r
e
p
r
o
ce
s
s
in
g
a
n
d
b
alan
ci
n
g
s
tr
ateg
ies.
2
.
1
.
Da
t
a
s
et
des
cr
iptio
n
T
h
is
wo
r
k
u
s
es
th
e
ar
r
h
y
t
h
m
ia
d
ataset,
is
f
o
u
n
d
at
th
e
UC
I
ML
r
ep
o
s
ito
r
y
[
2
5
]
.
T
h
er
e
ar
e
2
7
9
f
ea
tu
r
es
an
d
4
5
2
r
ec
o
r
d
s
th
at
m
ak
e
u
p
th
e
d
ataset.
E
ac
h
r
ec
o
r
d
is
class
if
ied
in
to
o
n
e
o
f
th
e
s
ix
teen
class
es
; a
clas
s
lab
el
o
f
1
d
esi
g
n
ates r
eg
u
lar
E
C
G
p
atter
n
s
,
wh
ile
a
class
lab
el
r
an
g
in
g
f
r
o
m
2
to
1
6
d
esig
n
ates
v
ar
io
u
s
ar
r
h
y
th
m
ia
k
in
d
s
.
I
n
th
is
r
esear
ch
,
th
e
n
o
r
m
al
cl
ass
in
d
icate
s
a
r
eg
u
lar
E
C
G
p
atter
n
an
d
th
e
o
th
er
1
5
class
es
in
d
icate
ca
r
d
iac
ar
r
h
y
th
m
ia.
I
t sh
o
u
ld
b
e
n
o
ted
t
h
at
th
e
UC
I
ar
r
h
y
th
m
ia
d
ataset
d
o
es n
o
t p
r
o
v
id
e
r
aw
E
C
G
wav
ef
o
r
m
s
,
b
u
t
in
s
tead
2
7
9
p
r
e
-
ex
tr
ac
t
ed
n
u
m
er
ical
f
ea
tu
r
es
th
at
r
ep
r
esen
t
v
ar
io
u
s
clin
ical
an
d
s
ig
n
al
-
d
er
iv
e
d
ch
ar
ac
ter
is
tics
.
T
h
u
s
,
th
e
s
tu
d
y
is
lim
ited
to
wo
r
k
i
n
g
wit
h
th
ese
alr
ea
d
y
p
r
o
ce
s
s
ed
f
e
atu
r
es,
an
d
d
id
n
o
t
p
er
f
o
r
m
r
aw
-
s
ig
n
al
–
lev
el
p
r
e
p
r
o
ce
s
s
in
g
s
u
ch
as
f
ilter
in
g
o
r
b
ea
t
s
eg
m
en
tatio
n
.
T
h
is
en
s
u
r
es
r
ep
r
o
d
u
cib
ilit
y
an
d
co
m
p
ar
a
b
ilit
y
with
p
r
i
o
r
wo
r
k
s
u
s
in
g
th
e
s
am
e
d
ataset.
Fig
u
r
e
2
d
e
m
o
n
s
tr
ates
th
e
im
b
alan
ce
d
class
d
is
tr
ib
u
tio
n
o
f
th
e
d
ataset.
Fig
u
r
e
2
.
I
m
b
alan
ce
d
class
d
is
tr
ib
u
tio
n
o
f
th
e
d
ataset
2
.
2
.
Da
t
a
prepro
ce
s
s
ing
Sin
ce
th
e
d
ataset
co
m
p
r
is
es
ex
tr
ac
ted
f
ea
tu
r
es
r
at
h
er
th
a
n
r
aw
E
C
G
s
ig
n
als,
th
e
p
r
e
p
r
o
ce
s
s
in
g
f
o
cu
s
ed
o
n
e
n
s
u
r
in
g
d
ata
q
u
a
lity
at
th
e
f
ea
tu
r
e
lev
el.
Miss
in
g
v
alu
es
wer
e
id
en
tifie
d
an
d
m
an
ag
ed
,
wh
ile
co
lu
m
n
s
with
c
o
n
s
tan
t
v
alu
es (
all
0
s
o
r
all
1
s
)
wer
e
r
em
o
v
ed
as
n
o
n
-
in
f
o
r
m
ativ
e.
All r
em
ai
n
in
g
f
ea
t
u
r
es
wer
e
n
o
r
m
alize
d
u
s
in
g
Stan
d
ar
d
Scaler
to
ac
h
iev
e
a
co
m
p
a
r
ab
le
r
a
n
g
e.
2
.
2
.
1
.
No
rma
liza
t
io
n
No
r
m
alizin
g
o
r
s
ca
lin
g
f
ea
tu
r
e
v
ar
iab
les
in
s
tatis
tical
ev
alu
atio
n
an
d
ML
is
cr
u
cial
t
o
en
s
u
r
e
th
at
th
ey
p
o
s
s
ess
a
co
m
p
ar
ab
le
r
a
n
g
e
an
d
d
is
tr
ib
u
tio
n
.
Utilizin
g
th
is
tech
n
iq
u
e
m
a
y
en
h
an
c
e
th
e
ef
f
icien
cy
o
f
ce
r
tain
alg
o
r
ith
m
s
,
wh
ile
also
f
ac
ilit
atin
g
th
e
co
m
p
ar
is
o
n
a
n
d
in
ter
p
r
etatio
n
o
f
co
e
f
f
icie
n
ts
ass
o
ciate
d
with
v
ar
io
u
s
ch
a
r
ac
ter
is
tics
.
T
h
e
p
r
o
ce
s
s
o
f
d
ata
n
o
r
m
aliza
tio
n
is
ca
r
r
ied
o
u
t
o
n
th
e
f
ea
tu
r
e
v
alu
es
u
s
in
g
th
e
Stan
d
ar
d
Scaler
tech
n
iq
u
e
[
2
6
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
26
:
2
3
9
8
-
2
4
0
9
2402
2
.
2
.
2
.
ADASYN
Data
au
g
m
en
tatio
n
is
a
ML
-
b
ased
ap
p
r
o
ac
h
u
s
ed
to
t
ac
k
le
th
e
is
s
u
e
o
f
class
im
b
alan
ce
in
d
atasets
[
2
7
]
,
[
2
8
]
.
C
lass
im
b
alan
ce
r
ef
er
s
to
a
s
itu
atio
n
wh
er
e
o
n
e
class
in
a
class
if
icat
io
n
task
is
g
r
ea
tly
u
n
b
alan
ce
d
in
th
e
d
ataset.
ADASYN
u
tili
ze
s
a
d
en
s
ity
d
is
tr
ib
u
tio
n
f
u
n
ctio
n
t
o
cr
ea
te
ar
tifi
cial
r
ep
r
esen
tatio
n
s
o
f
th
e
m
in
o
r
ity
class
.
T
h
e
v
al
u
e
o
f
th
e
d
en
s
ity
d
is
tr
ib
u
tio
n
eq
u
atio
n
is
ad
ju
s
ted
ac
co
r
d
in
g
to
th
e
d
is
tr
ib
u
tio
n
o
f
th
e
d
ata
p
o
in
ts
,
en
ab
lin
g
m
o
r
e
ac
cu
r
ate
p
r
o
d
u
ctio
n
o
f
s
y
n
th
etic
s
am
p
les
th
at
r
esem
b
le
th
e
m
in
o
r
ity
class
.
Sy
n
th
etic
s
am
p
les
ar
e
cr
ea
ted
f
o
r
th
e
m
in
o
r
ity
class
d
ata
p
o
in
ts
th
at
p
r
o
v
id
e
ch
allen
g
es
f
o
r
th
e
ML
s
y
s
tem
t
o
lear
n
.
B
y
in
cr
ea
s
in
g
th
e
n
u
m
b
er
o
f
ar
tific
ial
s
am
p
les
f
o
r
th
e
u
n
d
er
r
e
p
r
esen
ted
class
,
th
e
d
is
p
ar
ity
with
in
th
e
class
es i
s
r
ed
u
ce
d
,
en
a
b
lin
g
th
e
ML
alg
o
r
ith
m
t
o
m
o
r
e
ef
f
ec
t
iv
ely
ca
teg
o
r
ize
t
h
e
u
n
d
er
r
e
p
r
esen
ted
class
.
T
h
e
ar
r
h
y
th
m
ia
d
ataset
s
h
o
wed
s
ev
er
e
class
im
b
alan
ce
,
with
th
e
lar
g
est
class
(
n
o
r
m
al)
h
av
in
g
2
4
5
s
am
p
les
wh
ile
s
o
m
e
m
i
n
o
r
ity
class
es
co
n
tain
ed
f
ew
er
th
an
1
0
.
T
h
is
im
b
alan
ce
(
≈
1
2
0
:
1
)
m
o
tiv
ated
th
e
u
s
e
o
f
ADASYN.
ADAS
YN
is
ap
p
lied
wi
th
k
=5
n
ea
r
est
n
eig
h
b
o
r
s
,
as
r
ec
o
m
m
en
d
ed
in
th
e
o
r
ig
in
al
p
ap
er
[
2
7
]
.
Sy
n
th
etic
s
am
p
les
wer
e
g
en
er
ated
o
n
ly
with
in
th
e
tr
ain
in
g
s
et,
an
d
th
e
o
v
er
s
am
p
lin
g
r
atio
was
ad
ap
tiv
ely
co
n
tr
o
lled
s
o
t
h
at
h
ar
d
er
-
to
-
lear
n
m
in
o
r
ity
s
am
p
le
s
co
n
tr
ib
u
ted
m
o
r
e
n
ew
in
s
tan
ce
s
.
2
.
2
.
3
.
G
ridSea
rc
hCV
I
n
th
is
s
tu
d
y
,
Gr
id
Sear
ch
C
V
p
lay
ed
a
p
iv
o
tal
r
o
le
in
o
p
tim
izin
g
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
ML
m
o
d
els
u
s
ed
f
o
r
ar
r
h
y
th
m
ia
class
if
icatio
n
.
Gr
id
Sear
ch
C
V
is
a
p
o
wer
f
u
l
h
y
p
er
p
ar
a
m
eter
tu
n
in
g
tech
n
i
q
u
e
t
h
at
s
y
s
tem
atica
lly
s
ea
r
ch
es
th
r
o
u
g
h
a
s
p
ec
if
ied
g
r
i
d
o
f
h
y
p
er
p
a
r
am
eter
s
to
id
en
tify
th
e
o
p
tim
al
co
m
b
in
atio
n
th
at
m
ax
im
izes
m
o
d
el
p
er
f
o
r
m
an
c
e.
B
y
ev
alu
atin
g
ea
ch
co
m
b
in
atio
n
o
f
h
y
p
er
p
a
r
am
eter
s
th
r
o
u
g
h
cr
o
s
s
-
v
alid
atio
n
(
C
V)
,
Gr
id
Sear
ch
C
V
en
s
u
r
es th
at
th
e
s
elec
ted
m
o
d
el
co
n
f
ig
u
r
atio
n
ac
h
iev
es th
e
b
est p
o
s
s
ib
le
p
er
f
o
r
m
an
ce
o
n
th
e
g
iv
en
d
ataset
[
2
9
]
,
[
3
0
]
.
F
o
r
th
is
wo
r
k
,
5
-
f
o
ld
Gr
id
Sear
ch
C
V
if
ap
p
lied
to
f
in
e
-
tu
n
e
t
h
e
h
y
p
er
p
ar
am
eter
s
o
f
v
ar
io
u
s
ML
class
if
ier
s
.
E
a
ch
m
o
d
el
h
as
its
o
wn
s
et
o
f
h
y
p
er
p
ar
am
eter
s
th
at
s
ig
n
if
ican
tly
in
f
lu
en
ce
its
p
er
f
o
r
m
an
ce
.
T
h
e
a
p
p
licatio
n
o
f
Gr
id
Sear
ch
C
V
in
v
o
lv
e
d
d
ef
i
n
in
g
a
p
a
r
am
eter
g
r
id
f
o
r
ea
ch
class
if
ier
,
wh
ich
in
clu
d
ed
a
r
an
g
e
o
f
p
o
s
s
ib
le
v
alu
es
f
o
r
ea
ch
h
y
p
e
r
p
ar
am
eter
.
T
h
e
to
o
l
th
en
co
n
d
u
cted
an
ex
h
au
s
tiv
e
s
ea
r
ch
o
v
e
r
th
e
g
r
i
d
b
y
f
itti
n
g
an
d
ev
alu
atin
g
th
e
m
o
d
el
f
o
r
ea
ch
co
m
b
in
atio
n
o
f
h
y
p
er
p
a
r
am
eter
s
u
s
in
g
C
V.
T
h
e
C
V
p
r
o
ce
s
s
in
v
o
lv
ed
s
p
litt
in
g
th
e
tr
ain
in
g
d
ata
in
to
s
ev
er
al
f
o
l
d
s
,
tr
ain
in
g
th
e
m
o
d
el
o
n
s
o
m
e
f
o
ld
s
wh
ile
v
alid
atin
g
it
o
n
th
e
r
e
m
ain
in
g
f
o
ld
,
an
d
th
en
av
e
r
a
g
in
g
t
h
e
r
esu
lts
to
en
s
u
r
e
r
o
b
u
s
t
an
d
u
n
b
iased
p
er
f
o
r
m
an
ce
esti
m
ates.
2
.
3
.
Use o
f
m
a
chine le
a
rnin
g
a
lg
o
rit
h
m
s
Ar
r
h
y
th
m
ia
class
if
icatio
n
h
as
b
ee
n
ex
p
lo
r
ed
u
s
in
g
v
a
r
io
u
s
ML
m
o
d
els,
s
u
ch
as
L
R
[
3
1
]
,
NB
[
3
2
]
,
DT
[
3
3
]
,
R
F
[
3
4
]
,
KNN
[
3
5
]
,
ML
P
[
3
6
]
,
GB
[
3
7
]
,
AB
[
3
8
]
,
L
GB
M
[
3
9
]
,
a
n
d
XGB
[
4
0
]
.
E
v
er
y
m
eth
o
d
h
as
d
is
tin
ct
ad
v
an
tag
es a
n
d
d
is
ad
v
an
tag
es,
an
d
th
e
s
elec
tio
n
o
f
a
n
alg
o
r
ith
m
[
4
1
]
is
co
n
tin
g
e
n
t
u
p
o
n
t
h
e
p
ar
ticu
lar
d
ataset
an
d
class
if
icatio
n
is
s
u
e.
T
o
o
p
tim
ize
m
o
d
el
p
er
f
o
r
m
an
ce
,
Gr
id
Sear
c
h
C
V
is
ap
p
l
ied
with
5
-
f
o
ld
C
V
ac
r
o
s
s
all
class
if
ier
s
.
Fo
r
ea
ch
m
o
d
el,
a
g
r
id
o
f
h
y
p
e
r
p
ar
am
eter
v
alu
es
was
s
y
s
tem
atica
ll
y
ex
p
lo
r
ed
,
a
n
d
th
e
co
n
f
ig
u
r
atio
n
y
ield
in
g
th
e
b
e
s
t
cr
o
s
s
-
v
alid
ated
ac
cu
r
ac
y
w
as
s
elec
ted
.
T
ab
le
1
s
u
m
m
ar
i
ze
s
b
o
th
th
e
test
ed
p
ar
am
eter
r
a
n
g
es a
n
d
th
e
f
in
al
ch
o
s
en
v
alu
es.
T
ab
le
1
.
Hy
p
er
-
p
a
r
am
eter
s
an
d
test
ed
r
an
g
es f
o
r
ML
m
o
d
els
M
o
d
e
l
H
y
p
e
r
-
p
a
r
a
met
e
r
(
s)
Te
st
e
d
r
a
n
g
e
S
e
l
e
c
t
e
d
v
a
l
u
e
LR
max
_
i
t
e
r
,
C
max
_
i
t
e
r
{
5
0
0
,
1
0
0
0
},
C
{
0
.
1
,
1
,
1
0
}
max
_
i
t
e
r
=
1
0
0
0
,
C
=1
NB
v
a
r
_
smo
o
t
h
i
n
g
{1
e
-
9
,
1
e
-
8
,
1
e
-
7}
1e
-
9
DT
max
_
d
e
p
t
h
{5
,
1
0
,
1
5
}
10
RF
n
_
e
st
i
ma
t
o
r
s
,
ma
x
_
d
e
p
t
h
n
_
e
st
i
ma
t
o
r
s
{1
0
0
,
2
0
0
,
3
0
0
},
ma
x
_
d
e
p
t
h
{N
o
n
e
,
1
0
,
2
0
}
n
_
e
st
i
ma
t
o
r
s
=
2
0
0
,
ma
x
_
d
e
p
t
h
=
N
o
n
e
K
N
N
n
_
n
e
i
g
h
b
o
r
s
{3
,
5
,
7
,
9
}
5
M
LP
h
i
d
d
e
n
_
l
a
y
e
r
_
si
z
e
s,
max
_
i
t
e
r
h
i
d
d
e
n
_
l
a
y
e
r
_
si
z
e
s {(
5
0
,
)
,
(
1
0
0
,
)
}
,
max
_
i
t
e
r
{
5
0
0
,
1
0
0
0
}
h
i
d
d
e
n
_
l
a
y
e
r
_
si
z
e
s
=
(
5
0
,
)
,
ma
x
_
i
t
e
r
=
1
0
0
0
GB
n
_
e
st
i
ma
t
o
r
s
,
l
e
a
r
n
i
n
g
_
r
a
t
e
n
_
e
st
i
ma
t
o
r
s
{5
0
,
1
0
0
,
2
0
0
}
,
l
e
a
r
n
i
n
g
_
r
a
t
e
{0
.
0
5
,
0
.
1
,
0
.
2
}
n
_
e
st
i
ma
t
o
r
s
=
2
0
0
,
l
e
a
r
n
i
n
g
_
r
a
t
e
=
0
.
1
AB
n
_
e
st
i
ma
t
o
r
s
,
l
e
a
r
n
i
n
g
_
r
a
t
e
n
_
e
st
i
ma
t
o
r
s
{5
0
,
1
0
0
,
2
0
0
}
,
l
e
a
r
n
i
n
g
_
r
a
t
e
{0
.
0
5
,
0
.
1
,
0
.
2
}
n
_
e
st
i
ma
t
o
r
s
=
2
0
0
,
l
e
a
r
n
i
n
g
_
r
a
t
e
=
0
.
1
LG
B
M
n
_
e
st
i
ma
t
o
r
s
,
n
u
m
_
l
e
a
v
e
s
n
_
e
st
i
ma
t
o
r
s
{5
0
,
1
0
0
,
2
0
0
}
,
n
u
m
_
l
e
a
v
e
s
{3
1
,
6
3
}
n
_
e
st
i
ma
t
o
r
s
=
5
0
,
n
u
m
_
l
e
a
v
e
s
=
3
1
XGB
n
_
e
st
i
ma
t
o
r
s
,
ma
x
_
d
e
p
t
h
,
l
e
a
r
n
i
n
g
_
r
a
t
e
n
_
e
st
i
ma
t
o
r
s
{5
0
,
1
0
0
,
2
0
0
}
,
ma
x
_
d
e
p
t
h
{3
,
6
},
l
e
a
r
n
i
n
g
_
r
a
t
e
{
0
.
0
5
,
0
.
1
}
n
_
e
st
i
ma
t
o
r
s
=
5
0
,
m
a
x
_
d
e
p
t
h
=
3
,
l
e
a
r
n
i
n
g
_
r
a
t
e
=
0
.
1
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
Du
r
in
g
th
e
tr
ain
in
g
p
r
o
ce
s
s
,
th
e
m
o
d
els
wer
e
tr
ain
ed
o
n
th
e
tr
ain
in
g
d
ata
u
s
in
g
CV
t
o
p
r
e
v
en
t
o
v
er
f
itti
n
g
.
T
h
e
tr
ai
n
in
g
o
f
th
e
ML
m
o
d
els
h
as
b
ee
n
d
o
n
e
u
s
in
g
CV
,
an
d
e
v
alu
atin
g
t
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
d
a
p
tive
s
yn
th
etic
-
b
a
s
ed
a
r
r
h
yth
mia
cla
s
s
ifica
tio
n
u
s
in
g
ma
ch
in
e
lea
r
n
in
g
tec
h
n
iq
u
es (
Md
.
R
a
b
iu
l
I
s
la
m)
2403
m
o
d
els
u
s
in
g
test
in
g
d
ata.
I
n
th
is
s
tu
d
y
,
th
e
d
ataset
was
s
p
lit
in
to
8
0
-
2
0
%
f
o
r
tr
ain
in
g
an
d
test
in
g
,
r
esp
ec
tiv
ely
,
an
d
th
e
tr
ain
e
d
m
o
d
els we
r
e
ev
alu
ated
u
s
in
g
t
h
e
test
d
ata.
Fig
u
r
e
3
s
u
m
m
ar
izes
th
e
co
m
p
ar
ativ
e
p
er
f
o
r
m
an
ce
o
f
th
e
ML
m
o
d
els
ac
r
o
s
s
s
ix
ev
alu
at
io
n
m
etr
ics.
E
n
s
em
b
le
class
if
ier
s
s
u
ch
as
XGB,
G
B
,
L
GB
M,
an
d
R
F
ac
h
iev
ed
th
e
h
ig
h
est
s
co
r
es,
ea
c
h
r
ec
o
r
d
in
g
9
7
%
f
o
r
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
a
n
d
F1
-
s
co
r
e
,
with
s
tr
o
n
g
a
g
r
ee
m
en
t
in
d
ices
(
MCC
=0
.
9
0
-
0
.
9
6
an
d
C
o
h
en
’
s
Kap
p
a
=0
.
9
1
-
0
.
9
5
)
,
c
o
n
f
ir
m
i
n
g
th
eir
r
o
b
u
s
tn
ess
.
DT
f
o
llo
wed
clo
s
ely
with
9
5
%
ac
r
o
s
s
all
f
o
u
r
p
r
im
ar
y
m
etr
ics
an
d
s
o
lid
ag
r
ee
m
en
t
(
MCC
=0
.
9
4
,
Kap
p
a
=
0
.
9
2
)
.
L
R
an
d
ML
P
m
ain
tain
ed
b
alan
ce
d
p
e
r
f
o
r
m
an
ce
at
9
4
%
with
MCC
v
alu
es
o
f
0
.
9
3
a
n
d
0
.
8
8
,
an
d
Kap
p
a
s
co
r
es
o
f
0
.
9
0
an
d
0
.
8
6
,
r
esp
e
ctiv
ely
.
KNN
also
p
er
f
o
r
m
ed
r
eliab
ly
with
9
3
-
9
4
%
an
d
h
ig
h
ag
r
ee
m
e
n
t
(
MCC
=0
.
9
0
,
Kap
p
a
=0
.
9
2
)
.
I
n
co
n
t
r
ast,
NB
p
r
o
d
u
ce
d
o
n
ly
m
o
d
er
ate
r
esu
lts
(
8
1
-
8
2
%,
MC
C
=0
.
7
8
,
Kap
p
a
=0
.
7
5
)
,
wh
ile
AB
u
n
d
er
p
er
f
o
r
m
ed
s
ig
n
if
ican
tly
,
r
ea
c
h
in
g
ju
s
t
2
4
%
ac
cu
r
ac
y
an
d
a
n
MCC
o
f
0
.
0
2
with
a
Kap
p
a
o
f
0
.
0
8
.
T
h
ese
f
in
d
in
g
s
clea
r
ly
h
ig
h
l
ig
h
t
th
e
s
u
p
er
i
o
r
ity
o
f
en
s
em
b
le
-
b
ased
m
o
d
els in
ar
r
h
y
th
m
ia
class
if
icatio
n
co
m
p
ar
ed
to
t
r
ad
itio
n
al
class
if
ier
s
.
T
o
v
alid
ate
th
e
s
ig
n
if
ican
ce
o
f
m
o
d
el
p
er
f
o
r
m
an
ce
d
if
f
er
en
ce
s
,
p
air
ed
t
-
test
s
wer
e
p
er
f
o
r
m
ed
co
m
p
ar
in
g
AB
with
th
e
to
p
-
p
er
f
o
r
m
in
g
m
o
d
els.
T
h
e
r
esu
lts
co
n
f
ir
m
e
d
th
at
AB
’
s
ac
cu
r
ac
y
was
s
ig
n
if
ican
tly
lo
wer
(
p
<0
.
0
1
in
all
ca
s
es).
Sp
ec
if
ically
,
co
m
p
ar
is
o
n
s
y
ield
ed
:
R
F
v
s
AB
(
t
=
-
1
4
8
.
6
0
,
p
=1
.
2
3
×1
0
⁻⁸)
,
GB
v
s
AB
(
t
=
-
1
4
8
.
6
0
,
p
=1
.
2
3
×1
0
-
8
)
,
L
GB
M
v
s
AB
(
t
=
-
1
4
1
.
9
9
,
p
=
1
.
4
8
×
1
0
-
8
)
,
an
d
XGB
v
s
AB
(
t
=
-
1
9
4
.
5
7
,
p
=4
.
1
9
×
1
0
-
9
)
.
T
h
e
test
s
r
ev
ea
led
s
tati
s
tical
ly
s
ig
n
if
ican
t
d
if
f
er
en
ce
s
(
p
<0
.
0
1
)
,
ap
p
r
o
v
i
n
g
th
at
AB
’
s
m
ar
k
ed
ly
lo
wer
ac
cu
r
ac
y
(
2
4
%)
is
n
o
t
d
u
e
to
r
an
d
o
m
v
ar
iatio
n
b
u
t
r
ep
r
esen
ts
a
g
e
n
u
in
e
p
er
f
o
r
m
an
ce
g
ap
.
T
ab
le
2
s
h
o
ws
th
e
an
al
y
s
is
b
y
r
ep
o
r
tin
g
9
5
%
co
n
f
id
en
ce
in
ter
v
als
(
C
I
)
f
o
r
th
e
ac
cu
r
ac
y
d
if
f
e
r
en
ce
s
b
etwe
en
AB
an
d
th
e
t
o
p
-
p
e
r
f
o
r
m
in
g
m
o
d
els
(
R
F,
GB
,
L
GB
M,
an
d
XGB).
All
CI
ex
clu
d
e
ze
r
o
.
T
h
is
r
ein
f
o
r
ce
s
o
u
r
c
o
n
clu
s
io
n
t
h
a
t
AB
’
s
p
o
o
r
p
e
r
f
o
r
m
an
ce
is
i
n
h
er
en
t
t
o
its
s
en
s
itiv
ity
u
n
d
er
ADASYN
-
b
ased
o
v
er
s
am
p
lin
g
,
wh
ile
en
s
em
b
le
m
eth
o
d
s
m
ain
tain
s
tab
le
a
n
d
s
u
p
er
io
r
ac
cu
r
ac
y
.
Fig
u
r
e
3
.
Per
f
o
r
m
an
c
es o
f
ML
m
o
d
els
T
ab
le
2
.
Pair
ed
t
-
test
r
esu
lts
with
9
5
%
CI
(
AB
v
s
.
to
p
m
o
d
el
s
)
C
o
m
p
a
r
i
so
n
t
-
v
a
l
u
e
p
-
v
a
l
u
e
9
5
%
C
I
o
f
a
c
c
u
r
a
c
y
d
i
f
f
e
r
e
n
c
e
R
F
v
s AB
-
1
4
8
.
6
0
1
.
2
3
×
1
0
-
8
[
0
.
7
1
,
0
.
7
5
]
G
B
v
s
A
B
-
1
4
8
.
6
0
1
.
2
3
×
1
0
-
8
[
0
.
7
1
,
0
.
7
5
]
LG
B
M
v
s
A
B
-
1
4
1
.
9
9
1
.
4
8
×
1
0
-
8
[
0
.
7
0
,
0
.
7
4
]
X
G
B
v
s
A
B
-
1
9
4
.
5
7
4
.
1
9
×
1
0
-
9
[
0
.
7
2
,
0
.
7
6
]
Fig
u
r
e
4
p
r
es
en
ts
th
e
a
r
e
a
u
n
d
er
t
h
e
cu
r
v
e
(
AUC
)
-
r
ec
e
iv
er
o
p
e
r
a
ti
n
g
ch
ar
ac
te
r
is
ti
c
(
R
OC
)
o
f
th
e
ML
m
o
d
els,
wh
er
e
e
n
s
e
m
b
l
e
m
et
h
o
d
s
ac
h
i
ev
ed
t
h
e
b
est
p
er
f
o
r
m
an
ce
.
XGB
att
ai
n
e
d
t
h
e
h
i
g
h
est
A
UC
(
0
.
9
9
)
,
f
o
ll
o
we
d
c
lo
s
el
y
b
y
L
GB
M
a
n
d
GB
(
0
.
9
8
e
ac
h
)
,
w
h
il
e
R
F
al
s
o
p
er
f
o
r
m
ed
s
t
r
o
n
g
ly
w
it
h
0
.
9
7
.
M
o
d
e
r
ate
r
es
u
lts
wer
e
o
b
s
e
r
v
ed
f
o
r
D
T
(
0
.
9
4
)
,
KNN
(
0
.
9
2
)
,
ML
P
(
0
.
9
1
)
,
an
d
L
R
(
0
.
8
8
)
,
w
h
e
r
e
as
N
B
s
h
o
we
d
w
ea
k
e
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
26
:
2
3
9
8
-
2
4
0
9
2404
d
is
c
r
i
m
i
n
a
ti
o
n
(
0
.
8
2
)
.
I
n
c
o
n
t
r
ast,
AB
p
e
r
f
o
r
m
e
d
p
o
o
r
ly
w
it
h
a
n
AUC
o
f
o
n
ly
0
.
3
0
,
co
n
f
i
r
m
in
g
its
i
n
s
t
a
b
ili
ty
wit
h
ADAS
YN
b
a
la
n
ci
n
g
.
T
h
ese
f
i
n
d
i
n
g
s
e
m
p
h
asiz
e
t
h
e
r
o
b
u
s
tn
ess
o
f
e
n
s
e
m
b
le
-
b
ase
d
cl
ass
i
f
i
er
s
f
o
r
ar
r
h
y
th
m
i
a
c
lass
i
f
i
ca
t
io
n
.
Fi
g
u
r
e
5
ill
u
s
t
r
a
tes
th
e
co
n
f
u
s
i
o
n
m
at
r
i
x
f
o
r
t
h
e
R
F
m
o
d
e
l
w
h
i
c
h
p
e
r
f
o
r
m
e
d
b
es
t
i
n
th
e
a
f
o
r
e
m
e
n
ti
o
n
e
d
m
e
tr
ics
i
n
class
i
f
y
i
n
g
v
a
r
i
o
u
s
h
e
a
r
t
c
o
n
d
i
tio
n
s
.
T
h
e
m
a
tr
ix
d
e
m
o
n
s
t
r
ate
s
t
h
e
t
r
u
e
p
o
s
it
iv
es,
f
als
e
p
o
s
i
ti
v
es
,
f
alse
n
e
g
ati
v
es
,
a
n
d
tr
u
e
n
e
g
ati
v
es
f
o
r
ea
c
h
h
ea
r
t
c
o
n
d
iti
o
n
,
w
it
h
c
o
l
o
r
i
n
t
e
n
s
it
y
r
e
f
le
cti
n
g
t
h
e
n
u
m
b
e
r
o
f
i
n
s
t
an
ce
s
f
o
r
e
ac
h
c
lass
i
f
i
ca
ti
o
n
.
T
h
e
d
ia
g
o
n
al
ele
m
e
n
ts
,
s
u
ch
as
6
7
f
o
r
“
n
o
r
m
al
”
an
d
8
2
f
o
r
“
r
i
g
h
t
b
u
n
d
l
e
b
r
an
ch
b
l
o
ck
,
”
h
i
g
h
li
g
h
t
t
h
e
R
F
m
o
d
e
l’
s
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
i
n
c
o
r
r
ec
t
ly
c
lass
i
f
y
in
g
t
h
e
m
o
s
t
co
m
m
o
n
c
o
n
d
iti
o
n
s
.
H
o
we
v
er
,
s
o
m
e
m
is
cl
ass
i
f
ic
ati
o
n
s
a
r
e
v
is
ib
le
,
p
a
r
ti
cu
la
r
l
y
f
o
r
“
le
f
t
b
u
n
d
le
b
r
a
n
c
h
b
l
o
ck
”
(
5
6
m
is
class
if
i
e
d
i
n
s
ta
n
ce
s
)
a
n
d
“
AF
o
r
f
l
u
t
te
r
”
(
3
m
is
cl
ass
i
f
icat
io
n
s
)
.
Des
p
it
e
t
h
es
e
m
is
c
la
s
s
if
i
ca
t
io
n
s
,
t
h
e
R
F
m
o
d
el
ac
h
i
e
v
es
h
i
g
h
ac
cu
r
a
c
y
i
n
p
r
ed
ict
in
g
t
h
e
m
aj
o
r
it
y
o
f
c
o
n
d
iti
o
n
s
,
w
it
h
ce
r
t
ai
n
less
f
r
e
q
u
en
t
ca
te
g
o
r
ies
s
h
o
wi
n
g
r
o
o
m
f
o
r
f
u
r
t
h
er
i
m
p
r
o
v
e
m
e
n
t
.
Fig
u
r
e
4
.
AUC
-
R
OC
cu
r
v
e
o
f
th
e
ML
m
o
d
els
Fig
u
r
e
5
.
C
o
n
f
u
s
io
n
m
atr
i
x
f
o
r
th
e
b
est (
R
F)
m
o
d
el
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
d
a
p
tive
s
yn
th
etic
-
b
a
s
ed
a
r
r
h
yth
mia
cla
s
s
ifica
tio
n
u
s
in
g
ma
ch
in
e
lea
r
n
in
g
tec
h
n
iq
u
es (
Md
.
R
a
b
iu
l
I
s
la
m)
2405
Fig
u
r
e
6
s
h
o
ws
th
e
p
r
e
d
ictio
n
p
r
o
b
ab
ilit
ies
f
o
r
a
s
p
ec
if
ic
in
s
tan
ce
,
h
ig
h
lig
h
tin
g
th
e
m
o
d
e
l’
s
o
u
tp
u
t
f
o
r
th
e
p
r
e
d
icted
class
(
4
)
with
a
p
r
o
b
ab
ilit
y
o
f
0
.
9
9
.
T
h
e
tab
le
o
n
th
e
r
ig
h
t lis
ts
th
e
v
alu
es o
f
k
ey
f
ea
tu
r
es th
at
m
o
s
t
in
f
lu
en
ce
th
e
p
r
ed
ictio
n
.
Fo
r
in
s
tan
ce
,
f
ea
tu
r
e
1
2
4
(
with
a
v
alu
e
o
f
3
6
.
0
0
)
h
as
t
h
e
m
o
s
t
s
ig
n
if
ican
t
im
p
ac
t
o
n
th
e
p
r
e
d
ictio
n
,
wh
il
e
f
ea
tu
r
es
s
u
ch
as
2
7
6
an
d
1
9
6
also
co
n
tr
ib
u
te
s
u
b
s
tan
tially
.
T
h
is
v
is
u
aliza
tio
n
d
em
o
n
s
tr
ates
th
e
in
ter
p
r
etab
il
ity
o
f
th
e
m
o
d
el,
o
f
f
e
r
in
g
in
s
ig
h
ts
in
to
wh
ich
f
e
atu
r
es
d
r
iv
e
th
e
class
if
icatio
n
d
ec
is
io
n
f
o
r
t
h
is
s
p
ec
if
ic
ca
s
e.
T
ab
le
3
d
em
o
n
s
tr
ates
th
e
co
m
p
ar
ativ
e
a
n
aly
s
is
o
f
th
e
p
r
o
p
o
s
ed
wo
r
k
with
th
e
p
r
ev
io
u
s
wo
r
k
s
.
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
d
em
o
n
s
tr
ate
a
s
ig
n
if
ican
t
im
p
r
o
v
em
en
t in
th
e
f
ield
o
f
ar
r
h
y
th
m
ia
class
if
icatio
n
co
m
p
ar
ed
to
p
r
ev
i
o
u
s
wo
r
k
s
.
B
y
ad
d
r
ess
in
g
th
e
class
im
b
alan
ce
p
r
o
b
lem
u
s
in
g
ADASYN
an
d
lev
e
r
ag
in
g
a
co
m
p
r
eh
e
n
s
iv
e
s
u
ite
o
f
ML
m
o
d
els,
s
u
p
er
io
r
class
if
icatio
n
p
er
f
o
r
m
an
ce
was
ac
h
ie
v
ed
.
I
n
p
r
e
v
io
u
s
wo
r
k
s
,
class
if
ier
s
s
u
ch
as
PC
A
with
SVM
an
d
R
F
ac
h
iev
ed
ac
cu
r
ac
i
es
o
f
9
1
.
2
%
[
2
]
an
d
8
9
%
[
7
]
,
r
esp
ec
tiv
ely
,
wh
ile
NB
r
ep
o
r
ted
a
lo
wer
ac
cu
r
ac
y
o
f
8
5
%
[
3
2
]
.
GB
m
o
d
els
l
ik
e
XGB
an
d
AB
y
ield
ed
h
ig
h
er
ac
cu
r
a
cies
o
f
9
5
.
6
5
%
[
3
7
]
an
d
9
4
.
1
5
%
[
3
8
]
,
r
esp
ec
tiv
ely
,
d
em
o
n
s
tr
atin
g
th
eir
s
tr
o
n
g
p
r
e
d
ictiv
e
ca
p
ab
ilit
y
.
Ho
wev
er
,
R
F
alo
n
e
in
p
r
e
v
io
u
s
wo
r
k
p
er
f
o
r
m
ed
at
a
s
u
b
o
p
tim
al
lev
el
wi
th
8
4
%
[
4
2
]
ac
cu
r
ac
y
.
I
n
c
o
n
tr
ast,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
l
o
g
y
u
s
in
g
ADASYN
s
ig
n
if
ican
tly
b
o
o
s
ts
th
e
p
er
f
o
r
m
an
ce
o
f
s
ev
er
al
m
o
d
els.
Fo
r
in
s
tan
ce
,
R
F,
GB
,
L
GB
M,
an
d
XGB
all
ac
h
iev
ed
an
im
p
r
ess
iv
e
ac
cu
r
ac
y
o
f
9
7
%.
T
h
is
m
ar
k
s
a
s
u
b
s
tan
tia
l
im
p
r
o
v
em
en
t
o
v
er
th
eir
co
u
n
ter
p
ar
ts
in
p
r
ev
io
u
s
wo
r
k
s
,
p
ar
ticu
lar
ly
f
o
r
R
F,
wh
ich
im
p
r
o
v
e
d
f
r
o
m
8
4
%
to
9
7
%,
h
ig
h
lig
h
tin
g
th
e
im
p
ac
t o
f
ad
d
r
ess
in
g
th
e
class
im
b
alan
ce
is
s
u
e.
Fig
u
r
e
6
.
Pre
d
ictio
n
an
aly
s
is
u
s
in
g
lo
ca
l in
ter
p
r
etab
le
m
o
d
el
-
ag
n
o
s
tic
ex
p
lan
atio
n
s
(
LIME
)
f
o
r
R
F m
o
d
el
T
ab
le
3
.
Per
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
wo
r
k
with
ex
is
tin
g
wo
r
k
s
R
e
f
e
r
e
n
c
e
M
e
t
h
o
d
u
se
d
A
c
c
u
r
a
c
y
(
%)
[
2
]
P
C
A
+
S
V
M
9
1
.
2
[
7
]
P
C
A
+
R
F
89
[
3
2
]
NB
85
[
3
7
]
XGB
9
5
.
6
5
[
3
8
]
AB
9
4
.
1
5
[
4
2
]
RF
84
P
r
o
p
o
se
d
w
o
r
k
A
D
A
S
Y
N
+
R
F
97
A
D
A
S
Y
N
+
G
B
97
A
D
A
S
Y
N
+
LB
G
M
97
A
D
A
S
Y
N
+
X
G
B
97
Ad
d
itio
n
ally
,
DT
an
d
L
R
m
o
d
els
in
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
also
p
er
f
o
r
m
e
d
well,
ac
h
iev
in
g
9
5
%
an
d
9
4
%
ac
cu
r
ac
y
,
r
esp
ec
tiv
ely
,
wh
ich
is
s
u
p
er
io
r
to
m
o
s
t
p
r
ev
io
u
s
ap
p
r
o
ac
h
es.
I
n
te
r
e
s
tin
g
ly
,
wh
ile
m
o
s
t
class
if
ier
s
in
th
e
p
r
o
p
o
s
ed
wo
r
k
p
er
f
o
r
m
ed
ex
ce
p
tio
n
ally
,
NB
ex
p
er
ien
ce
d
a
d
r
o
p
in
p
e
r
f
o
r
m
a
n
ce
,
with
an
ac
cu
r
ac
y
o
f
8
1
%.
T
h
is
is
lik
el
y
d
u
e
to
NB
’
s
ass
u
m
p
tio
n
o
f
f
ea
tu
r
e
in
d
ep
e
n
d
en
ce
,
wh
ic
h
m
ay
n
o
t h
o
ld
tr
u
e
f
o
r
th
e
co
m
p
lex
p
atter
n
s
p
r
esen
t
in
ar
r
h
y
th
m
ia
d
ata.
Fu
r
th
er
m
o
r
e,
KNN
p
er
f
o
r
m
e
d
r
ea
s
o
n
a
b
ly
well,
with
9
3
%
ac
cu
r
ac
y
,
alth
o
u
g
h
it
d
i
d
n
o
t
r
ea
ch
th
e
s
am
e
h
ig
h
lev
el
as
o
th
er
m
o
d
els
lik
e
R
F
o
r
GB
.
On
e
n
o
tab
le
o
b
s
er
v
atio
n
is
th
e
p
o
o
r
p
er
f
o
r
m
an
ce
o
f
AB
in
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
,
wh
ich
r
esu
lted
in
a
s
ig
n
if
ican
tly
lo
w
ac
cu
r
ac
y
o
f
2
4
%.
T
h
is
s
h
ar
p
d
ec
lin
e
co
m
p
ar
e
d
to
p
r
i
o
r
wo
r
k
(
9
4
.
1
5
%)
s
u
g
g
ests
th
at
AB
m
ay
b
e
p
ar
ticu
lar
l
y
s
en
s
itiv
e
to
th
e
s
y
n
th
etic
s
am
p
les g
en
er
ated
b
y
ADASYN,
lead
in
g
to
o
v
er
f
itti
n
g
o
r
i
n
ef
f
ec
t
iv
e
lear
n
in
g
.
A
d
r
awb
ac
k
o
f
th
is
wo
r
k
is
th
e
r
estricte
d
v
ar
iab
ilit
y
o
f
th
e
UC
I
ar
r
h
y
th
m
ia
d
ataset.
R
ea
l
-
wo
r
ld
E
C
G
r
ec
o
r
d
in
g
s
ar
e
in
f
lu
en
ce
d
b
y
n
o
is
e,
b
aselin
e
d
r
if
t,
elec
tr
o
d
e
p
lace
m
en
t,
an
d
d
e
v
ice
-
s
p
ec
if
ic
ar
tifa
cts.
I
n
ad
d
itio
n
,
p
atien
t
p
o
p
u
latio
n
s
d
if
f
er
b
y
ag
e,
c
o
m
o
r
b
id
ities
,
an
d
eth
n
icity
,
wh
ic
h
m
ay
af
f
ec
t
E
C
G
m
o
r
p
h
o
lo
g
y
an
d
ar
r
h
y
th
m
ia
ex
p
r
ess
io
n
.
B
ec
au
s
e
th
is
d
ataset
lack
s
s
u
ch
d
iv
er
s
ity
,
t
h
e
r
esu
lts
m
ay
n
o
t
f
u
lly
g
en
er
alize
t
o
h
eter
o
g
en
e
o
u
s
clin
ical
p
o
p
u
latio
n
s
.
Ov
er
all,
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
h
as
d
em
o
n
s
tr
ated
th
at
th
e
in
teg
r
atio
n
o
f
ADASYN
with
ML
m
o
d
els,
p
ar
ticu
lar
ly
en
s
em
b
le
m
eth
o
d
s
lik
e
R
F,
G
B
,
an
d
XG
B
,
is
h
ig
h
ly
ef
f
ec
tiv
e
f
o
r
ar
r
h
y
th
m
ia
class
if
icatio
n
.
T
h
e
ab
ilit
y
to
h
a
n
d
le
im
b
alan
ce
d
d
atasets
with
ADASY
N
co
n
tr
ib
u
ted
to
s
u
p
er
io
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
26
:
2
3
9
8
-
2
4
0
9
2406
class
if
icatio
n
p
er
f
o
r
m
an
ce
,
m
ak
in
g
th
is
m
eth
o
d
o
lo
g
y
m
o
r
e
r
eliab
le
f
o
r
r
ea
l
-
w
o
r
ld
m
ed
ical
ap
p
licatio
n
s
co
m
p
ar
ed
to
ea
r
lier
a
p
p
r
o
ac
h
e
s
.
B
ey
o
n
d
alg
o
r
ith
m
ic
ev
al
u
atio
n
,
an
im
p
o
r
tan
t
co
n
s
id
er
atio
n
is
clin
ical
ap
p
licab
ilit
y
.
W
h
ile
th
is
s
tu
d
y
em
p
lo
y
ed
th
e
UC
I
ar
r
h
y
th
m
i
a
d
ataset
o
f
p
r
e
-
ex
tr
ac
te
d
f
e
atu
r
es,
th
e
p
r
o
p
o
s
ed
f
r
am
ew
o
r
k
ca
n
b
e
r
ea
d
il
y
ad
ap
ted
f
o
r
r
ea
l
-
tim
e
u
s
e.
E
C
G
d
ev
ices
s
u
ch
as
wea
r
a
b
le
p
atch
es,
Ho
lter
m
o
n
ito
r
s
,
a
n
d
s
m
ar
twatch
es
ca
n
co
n
tin
u
o
u
s
ly
r
ec
o
r
d
r
aw
E
C
G
s
ig
n
als,
wh
ich
ar
e
th
en
t
r
an
s
f
o
r
m
ed
i
n
to
n
u
m
er
ical
f
ea
t
u
r
es
th
r
o
u
g
h
e
m
b
ed
d
ed
s
ig
n
al
p
r
o
ce
s
s
in
g
m
o
d
u
les.
T
h
ese
f
ea
tu
r
es
co
u
ld
b
e
d
ir
ec
tl
y
f
ed
in
to
th
e
ADASYN+
ML
p
ip
elin
e,
en
a
b
lin
g
r
ea
l
-
tim
e
class
if
icatio
n
o
f
ar
r
h
y
th
m
ias
at
th
e
p
o
in
t
o
f
ca
r
e.
I
n
h
o
s
p
ital
s
ettin
g
s
,
in
teg
r
atio
n
with
E
HR
s
y
s
tem
s
co
u
ld
p
r
o
v
i
d
e
clin
ician
s
with
au
to
m
ated
aler
ts
d
u
r
i
n
g
p
atien
t
m
o
n
ito
r
in
g
,
im
p
r
o
v
in
g
e
ar
ly
d
etec
tio
n
an
d
d
iag
n
o
s
is
.
Fu
r
th
er
m
o
r
e,
th
is
f
r
am
ewo
r
k
is
co
m
p
u
tatio
n
ally
lig
h
tweig
h
t
co
m
p
ar
e
d
to
d
ee
p
lear
n
in
g
m
eth
o
d
s
,
m
ak
in
g
it
f
it
f
o
r
d
ep
lo
y
m
en
t
o
n
p
o
r
tab
le
o
r
lo
w
-
p
o
wer
d
e
v
ices.
C
h
allen
g
es
f
o
r
clin
ical
tr
an
s
latio
n
r
em
ain
,
in
clu
d
in
g
t
h
e
n
ee
d
f
o
r
p
r
o
s
p
ec
tiv
e
v
alid
atio
n
o
n
r
aw
E
C
G
s
ig
n
als,
en
s
u
r
in
g
in
ter
p
r
eta
b
ilit
y
o
f
m
o
d
el
o
u
tp
u
ts
,
an
d
ad
d
r
ess
in
g
r
eg
u
lato
r
y
an
d
eth
ical
is
s
u
es
r
elate
d
to
d
ata
p
r
iv
ac
y
a
n
d
s
af
ety
.
Nev
e
r
th
el
ess
,
th
ese
s
ce
n
ar
io
s
h
ig
h
lig
h
t th
e
r
ea
l
-
wo
r
ld
p
o
ten
t
ial
o
f
th
is
ap
p
r
o
ac
h
to
c
o
m
p
le
m
en
t e
x
is
tin
g
clin
ical
wo
r
k
f
lo
ws.
An
ad
d
itio
n
al
co
n
s
id
er
atio
n
i
s
th
e
f
ea
s
ib
ilit
y
o
f
d
ep
lo
y
m
e
n
t
o
n
ed
g
e
d
e
v
ices.
B
ec
au
s
e
th
e
m
o
d
els
ar
e
b
ased
o
n
lig
h
tweig
h
t M
L
a
lg
o
r
ith
m
s
,
th
ey
ca
n
b
e
in
te
g
r
at
ed
in
to
m
o
b
ile
E
C
G
m
o
n
ito
r
s
,
wea
r
ab
le
d
ev
ices,
o
r
b
e
d
s
id
e
h
o
s
p
ital
m
o
n
ito
r
s
with
o
u
t
r
e
q
u
ir
in
g
h
i
g
h
-
p
e
r
f
o
r
m
a
n
ce
g
r
ap
h
ics
p
r
o
ce
s
s
in
g
u
n
it
s
(
GPUs
)
.
E
m
b
ed
d
e
d
p
r
o
ce
s
s
o
r
s
(
e.
g
.
,
ad
v
an
ce
d
R
I
SC
m
ac
h
in
es
(
AR
M
)
-
b
ased
ch
ip
s
in
s
m
ar
t
watc
h
es
o
r
Ho
lter
m
o
n
ito
r
s
)
ca
n
e
x
ec
u
te
s
u
ch
m
o
d
els
in
r
ea
l
tim
e,
e
n
ab
lin
g
co
n
tin
u
o
u
s
m
o
n
ito
r
in
g
an
d
ea
r
ly
a
r
r
h
y
th
m
ia
d
etec
tio
n
at
th
e
p
atien
t’
s
s
id
e.
I
n
h
o
s
p
ital
wo
r
k
f
l
o
ws,
in
teg
r
atio
n
with
m
o
n
ito
r
in
g
s
y
s
tem
s
o
r
E
HR
-
lin
k
e
d
aler
ts
co
u
ld
s
u
p
p
o
r
t
clin
ician
s
b
y
p
r
o
v
id
in
g
a
u
to
m
ated
r
is
k
s
tr
atif
icatio
n
alo
n
g
s
id
e
r
o
u
t
in
e
E
C
G
r
ec
o
r
d
in
g
.
Mo
r
eo
v
er
,
wh
ile
t
h
e
e
x
p
er
im
en
ts
u
s
ed
th
e
UC
I
ar
r
h
y
th
m
i
a
d
ataset,
f
u
t
u
r
e
w
o
r
k
s
h
o
u
ld
ex
ten
d
to
ex
ter
n
a
l
v
alid
atio
n
o
n
wid
ely
u
s
ed
b
e
n
ch
m
ar
k
d
atasets
s
u
ch
as
MI
T
-
B
I
H,
p
u
lm
o
n
ar
y
tu
b
er
c
u
lo
s
is
(
PTB
)
d
iag
n
o
s
tic
E
C
G,
Ph
y
s
io
Net
ch
allen
g
e
d
atasets
o
r
o
n
p
r
o
s
p
ec
tiv
ely
co
llected
h
o
s
p
ital
E
C
G
d
ata.
S
u
ch
ex
ter
n
al
test
in
g
will b
e
cr
itical
to
en
s
u
r
e
g
e
n
er
aliza
b
ilit
y
ac
r
o
s
s
p
o
p
u
latio
n
s
,
d
ev
ice
s
ettin
g
s
,
an
d
n
o
is
e
en
v
i
r
o
n
m
en
ts
.
4.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
p
r
o
p
o
s
ed
a
n
en
h
a
n
ce
d
f
r
am
ewo
r
k
f
o
r
ar
r
h
y
th
m
ia
class
if
icatio
n
th
at
ad
d
r
ess
es
th
e
cr
itical
is
s
u
e
o
f
clas
s
im
b
alan
ce
u
s
in
g
ADASYN
co
m
b
in
ed
with
a
co
m
p
r
eh
e
n
s
iv
e
s
et
o
f
ML
a
lg
o
r
ith
m
s
.
T
h
r
o
u
g
h
r
ig
o
r
o
u
s
ex
p
er
im
e
n
tatio
n
,
t
h
e
ef
f
ec
tiv
en
ess
o
f
t
h
is
ap
p
r
o
a
c
h
was
d
em
o
n
s
tr
ated
ac
r
o
s
s
m
u
ltip
le
class
if
ier
s
,
ac
h
iev
in
g
s
u
b
s
tan
tial
im
p
r
o
v
em
en
ts
in
ac
cu
r
ac
y
co
m
p
a
r
e
d
to
p
r
e
v
io
u
s
wo
r
k
s
.
No
tab
l
y
,
en
s
em
b
le
m
o
d
els
s
u
ch
as
R
F,
GB
,
L
GB
M,
an
d
XGB
ac
h
iev
ed
an
o
u
ts
ta
n
d
in
g
ac
cu
r
ac
y
o
f
9
7
%,
m
a
r
k
in
g
a
s
ig
n
if
ican
t
en
h
an
ce
m
e
n
t
o
v
er
p
ast
r
esu
lts
.
Oth
er
m
o
d
els,
in
clu
d
in
g
DT
(
9
5
%),
L
R
(
9
4
%),
an
d
ML
P
(
9
4
%),
also
s
h
o
wed
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
,
u
n
d
er
s
co
r
in
g
th
e
r
o
b
u
s
tn
ess
o
f
th
e
m
eth
o
d
o
l
o
g
y
.
Desp
ite
th
e
o
v
er
all
s
u
cc
ess
,
s
o
m
e
m
o
d
els
lik
e
NB
an
d
KNN
u
n
d
er
p
er
f
o
r
m
ed
r
elativ
e
to
o
th
er
s
,
ac
h
iev
in
g
8
1
%
an
d
9
3
%
ac
c
u
r
ac
y
,
r
esp
ec
tiv
ely
.
Mo
r
eo
v
er
,
AB
ex
h
ib
ited
a
n
u
n
ex
p
ec
ted
ly
lo
w
ac
c
u
r
ac
y
o
f
2
4
%,
s
u
g
g
esti
n
g
th
at
it
m
ay
n
o
t
b
e
well
-
s
u
ited
f
o
r
d
atasets
with
s
y
n
th
etic
s
am
p
les
g
en
er
ated
b
y
ADASYN.
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
d
em
o
n
s
tr
ate
th
e
im
p
o
r
tan
ce
o
f
ad
d
r
ess
in
g
class
im
b
alan
c
e
in
a
r
r
h
y
t
h
m
ia
class
if
icatio
n
task
s
,
an
d
th
e
in
teg
r
atio
n
o
f
ADASYN
with
ad
v
an
ce
d
ML
tec
h
n
iq
u
es
p
r
o
v
ed
to
b
e
a
h
ig
h
ly
e
f
f
ec
tiv
e
s
o
lu
tio
n
.
T
h
is
wo
r
k
o
u
t
p
er
f
o
r
m
s
p
r
io
r
s
tu
d
ies
an
d
s
ets a
n
ew
b
en
ch
m
ar
k
f
o
r
ar
r
h
y
th
m
ia
d
etec
tio
n
an
d
class
if
icatio
n
.
Fu
tu
r
e
r
esear
ch
will f
o
c
u
s
o
n
v
alid
atin
g
th
e
f
r
am
ewo
r
k
u
s
in
g
r
aw
E
C
G
s
i
g
n
als
f
r
o
m
wea
r
ab
le
an
d
clin
i
ca
l
m
o
n
ito
r
i
n
g
d
ev
ices,
e
x
p
lo
r
in
g
a
d
d
itio
n
al
d
ata
r
esam
p
lin
g
tech
n
iq
u
es;
f
u
r
th
e
r
r
ef
in
in
g
m
o
d
el
o
p
tim
izatio
n
,
an
d
test
in
g
th
e
p
r
o
p
o
s
ed
f
r
a
m
ewo
r
k
in
r
ea
l
-
tim
e
clin
ical
s
ettin
g
s
to
en
h
an
ce
its
ap
p
licab
ilit
y
in
ca
r
d
iac
ca
r
e
d
iag
n
o
s
tics
.
Fu
tu
r
e
r
esear
ch
will
also
f
o
cu
s
o
n
d
ep
lo
y
in
g
th
e
f
r
a
m
ewo
r
k
o
n
e
d
g
e
d
ev
ices
s
u
ch
as
wea
r
ab
le
E
C
G
m
o
n
ito
r
s
an
d
h
o
s
p
ital
b
e
d
s
id
e
u
n
its
.
Giv
en
th
e
r
elativ
ely
lo
w
c
o
m
p
u
tatio
n
al
f
o
o
tp
r
in
t
o
f
t
h
is
ap
p
r
o
ac
h
,
r
ea
l
-
tim
e
in
f
er
en
ce
o
n
p
o
r
tab
le
d
ev
ices
is
f
ea
s
ib
le.
I
n
a
d
d
itio
n
,
v
alid
atio
n
o
n
ex
ter
n
al
b
en
c
h
m
ar
k
d
atas
ets
an
d
u
n
s
ee
n
h
o
s
p
ital
d
ata
will
b
e
p
r
io
r
itized
to
co
n
f
ir
m
r
o
b
u
s
tn
ess
an
d
clin
ical
u
tili
ty
.
I
t
is
also
ac
k
n
o
wled
g
ed
th
at
th
e
UC
I
d
ataset
u
s
ed
h
er
e
lack
s
th
e
v
ar
iab
ilit
y
an
d
p
atien
t
d
iv
er
s
ity
en
co
u
n
ter
ed
in
r
ea
l
-
wo
r
l
d
p
r
ac
tice.
E
x
ten
d
in
g
t
h
e
f
r
am
ewo
r
k
t
o
lar
g
e
r
,
h
eter
o
g
en
e
o
u
s
d
atasets
will b
e
ess
en
tial f
o
r
en
s
u
r
in
g
clin
ical
r
eliab
ilit
y
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
e
au
th
o
r
s
r
ec
eiv
ed
n
o
f
u
n
d
i
n
g
f
o
r
th
is
r
esear
ch
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
d
a
p
tive
s
yn
th
etic
-
b
a
s
ed
a
r
r
h
yth
mia
cla
s
s
ifica
tio
n
u
s
in
g
ma
ch
in
e
lea
r
n
in
g
tec
h
n
iq
u
es (
Md
.
R
a
b
iu
l
I
s
la
m)
2407
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Md
.
R
ab
iu
l I
s
lam
✓
✓
✓
✓
✓
✓
✓
✓
✓
T
ap
an
Ku
m
a
r
Go
d
d
e
r
✓
✓
✓
✓
✓
✓
✓
Ah
s
an
-
Ul
-
Am
b
ia
✓
✓
✓
✓
✓
✓
Fer
d
ib
-
Al
-
I
s
lam
✓
✓
✓
✓
✓
✓
✓
✓
✓
B
u
lb
u
l A
h
am
ed
✓
✓
✓
✓
✓
✓
An
in
d
y
a
Nag
✓
✓
✓
✓
✓
✓
Ar
if
u
l I
s
lam
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
T
h
e
au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
ter
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
u
s
ed
in
th
is
s
tu
d
y
ar
e
av
ailab
le
in
UC
I
ML
r
ep
o
s
ito
r
y
at
h
ttp
s
://d
o
i.o
r
g
/1
0
.
2
4
4
3
2
/C
5
B
S3
2
,
r
ef
er
en
ce
[
2
5
]
.
RE
F
E
R
E
NC
E
S
[
1
]
S
.
K
.
B
e
r
k
a
y
a
,
A
.
K
.
U
y
sa
l
,
E
.
S
.
G
u
n
a
l
,
S
.
Er
g
i
n
,
S
.
G
u
n
a
l
,
a
n
d
M
.
B
.
G
u
l
m
e
z
o
g
l
u
,
“
A
s
u
r
v
e
y
o
n
E
C
G
a
n
a
l
y
si
s
,
”
B
i
o
m
e
d
i
c
a
l
S
i
g
n
a
l
Pr
o
c
e
ssi
n
g
a
n
d
C
o
n
t
ro
l
,
v
o
l
.
4
3
,
p
p
.
2
1
6
–
2
3
5
,
M
a
y
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
s
p
c
.
2
0
1
8
.
0
3
.
0
0
3
.
[
2
]
P
.
S
h
i
mp
i
,
S
.
S
h
a
h
,
M
.
S
h
r
o
f
f
,
a
n
d
A
.
G
o
d
b
o
l
e
,
“
A
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
f
o
r
t
h
e
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
c
a
r
d
i
a
c
a
r
r
h
y
t
h
mi
a
,
”
i
n
2
0
1
7
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
i
n
g
M
e
t
h
o
d
o
l
o
g
i
e
s
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
(
I
C
C
M
C
)
,
Ju
l
.
2
0
1
7
,
p
p
.
6
0
3
–
607
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
C
M
C
.
2
0
1
7
.
8
2
8
2
5
3
7
.
[
3
]
A
.
B
a
h
l
e
t
a
l
.
,
“
R
e
c
u
r
s
i
v
e
f
e
a
t
u
r
e
e
l
i
m
i
n
a
t
i
o
n
i
n
r
a
n
d
o
m
f
o
r
e
s
t
c
l
a
ssi
f
i
c
a
t
i
o
n
s
u
p
p
o
r
t
s
n
a
n
o
ma
t
e
r
i
a
l
g
r
o
u
p
i
n
g
,
”
N
a
n
o
I
m
p
a
c
t
,
v
o
l
.
1
5
,
M
a
r
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
m
p
a
c
t
.
2
0
1
9
.
1
0
0
1
7
9
.
[
4
]
A
.
Jo
v
i
c
,
K
.
B
r
k
i
c
,
a
n
d
N
.
B
o
g
u
n
o
v
i
c
,
“
A
r
e
v
i
e
w
o
f
f
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
m
e
t
h
o
d
s
w
i
t
h
a
p
p
l
i
c
a
t
i
o
n
s
,
”
i
n
2
0
1
5
3
8
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
v
e
n
t
i
o
n
o
n
I
n
f
o
rm
a
t
i
o
n
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
El
e
c
t
r
o
n
i
c
s
a
n
d
Mi
c
ro
e
l
e
c
t
ro
n
i
c
s
(
MIP
RO
)
,
M
a
y
2
0
1
5
,
p
p
.
1
2
0
0
–
1
2
0
5
,
d
o
i
:
1
0
.
1
1
0
9
/
M
I
P
R
O
.
2
0
1
5
.
7
1
6
0
4
5
8
.
[
5
]
T.
W
a
n
g
,
P
.
C
h
e
n
,
T
.
B
a
o
,
J.
Li
,
a
n
d
X
.
Y
u
,
“
A
r
r
h
y
t
h
mi
a
c
l
a
s
si
f
i
c
a
t
i
o
n
a
l
g
o
r
i
t
h
m
b
a
se
d
o
n
S
M
O
TE
a
n
d
f
e
a
t
u
r
e
se
l
e
c
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
P
e
rf
o
rm
a
b
i
l
i
t
y
E
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
1
7
,
n
o
.
3
,
p
p
.
2
6
3
–
2
7
5
,
2
0
2
1
,
d
o
i
:
1
0
.
2
3
9
4
0
/
i
j
p
e
.
2
1
.
0
3
.
p
2
.
2
6
3
2
7
5
.
[
6
]
P
.
G
i
u
d
i
c
i
,
A
.
G
r
a
me
g
n
a
,
a
n
d
E.
R
a
f
f
i
n
e
t
t
i
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
c
l
a
ss
i
f
i
c
a
t
i
o
n
mo
d
e
l
c
o
m
p
a
r
i
s
o
n
,
”
S
o
c
i
o
-
E
c
o
n
o
m
i
c
P
l
a
n
n
i
n
g
S
c
i
e
n
c
e
s
,
v
o
l
.
8
7
,
Ju
n
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
s
e
p
s
.
2
0
2
3
.
1
0
1
5
6
0
.
[
7
]
S
.
G
.
C
h
i
c
k
a
r
a
m
a
n
n
a
e
t
a
l
.
,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
o
f
a
r
r
h
y
t
h
mi
a
u
s
i
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m,
”
Re
v
u
e
d
’
I
n
t
e
l
l
i
g
e
n
c
e
Art
i
f
i
c
i
e
l
l
e
,
v
o
l
.
3
6
,
n
o
.
4
,
p
p
.
5
2
9
–
5
3
4
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
1
8
2
8
0
/
r
i
a
.
3
6
0
4
0
3
.
[
8
]
E.
H
a
e
r
a
n
i
,
F
.
S
y
a
f
r
i
a
,
F
.
Le
st
a
r
i
,
N
.
N
o
v
r
i
y
a
n
t
o
,
a
n
d
I
.
M
a
r
z
u
k
i
,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
a
c
a
d
e
mi
c
d
a
t
a
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
d
e
c
i
si
o
n
ma
k
i
n
g
p
r
o
c
e
ss,
”
J
o
u
rn
a
l
o
f
Ap
p
l
i
e
d
En
g
i
n
e
e
ri
n
g
a
n
d
T
e
c
h
n
o
l
o
g
i
c
a
l
S
c
i
e
n
c
e
,
v
o
l
.
4
,
n
o
.
2
,
p
p
.
9
5
5
–
9
6
8
,
J
u
n
.
2
0
2
3
,
d
o
i
:
1
0
.
3
7
3
8
5
/
j
a
e
t
s.
v
4
i
2
.
1
9
8
3
.
[
9
]
S
.
S
.
,
K
.
H
.
N
e
h
e
m
i
a
h
,
a
n
d
K
.
A
.
,
“
A
f
r
a
mew
o
r
k
t
o
c
l
a
ssi
f
y
c
l
i
n
i
c
a
l
d
a
t
a
u
si
n
g
a
g
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
a
n
d
a
r
t
i
f
i
c
i
a
l
f
l
o
r
a
-
o
p
t
i
m
i
z
e
d
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
S
w
a
rm
I
n
t
e
l
l
i
g
e
n
c
e
Re
sea
r
c
h
,
v
o
l
.
1
3
,
n
o
.
1
,
p
p
.
1
–
2
2
,
J
u
l
.
2
0
2
2
,
d
o
i
:
1
0
.
4
0
1
8
/
I
JS
I
R
.
3
0
4
7
1
9
.
[
1
0
]
J.
-
G
.
C
h
o
i
,
I
.
K
o
,
a
n
d
S
.
H
a
n
,
“
D
e
p
r
e
ssi
o
n
l
e
v
e
l
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
si
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
c
l
a
ssi
f
i
e
r
s
b
a
se
d
o
n
a
c
t
i
g
r
a
p
h
y
d
a
t
a
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
9
,
p
p
.
1
1
6
6
2
2
–
1
1
6
6
4
6
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
1
.
3
1
0
5
3
9
3
.
[
1
1
]
H
.
Ja
v
a
i
d
,
E.
K
u
mar
n
si
t
,
a
n
d
S
.
C
h
a
t
p
u
n
,
“
C
l
a
ss
i
f
i
c
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
i
n
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
a
g
e
-
r
e
l
a
t
e
d
e
l
e
c
t
r
o
e
n
c
e
p
h
a
l
o
g
r
a
p
h
y
d
a
t
a
a
n
a
l
y
si
s
,
”
P
S
U
M
e
d
i
c
a
l
J
o
u
r
n
a
l
,
v
o
l
.
3
,
n
o
.
2
,
p
p
.
1
0
3
–
1
1
3
,
J
u
l
.
2
0
2
3
,
d
o
i
:
1
0
.
3
1
5
8
4
/
p
su
mj
.
2
0
2
3
2
5
5
2
0
8
.
[
1
2
]
A
.
F
a
r
o
o
q
,
M
.
S
e
y
e
d
m
a
h
m
o
u
d
i
a
n
,
a
n
d
A
.
S
t
o
j
c
e
v
s
k
i
,
“
A
w
e
a
r
a
b
l
e
w
i
r
e
l
e
ss
sen
s
o
r
s
y
st
e
m u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
c
l
a
ss
i
f
i
c
a
t
i
o
n
t
o
d
e
t
e
c
t
a
r
r
h
y
t
h
mi
a
,
”
I
EEE
S
e
n
s
o
rs
J
o
u
rn
a
l
,
v
o
l
.
2
1
,
n
o
.
9
,
p
p
.
1
1
1
0
9
–
1
1
1
1
6
,
M
a
y
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
JS
EN
.
2
0
2
1
.
3
0
6
2
3
9
5
.
[
1
3
]
P
.
G
a
u
t
a
m
,
M
.
S
i
n
g
h
,
a
n
d
B
.
R
.
M
u
k
i
n
d
r
a
o
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
se
d
a
r
r
h
y
t
h
mi
a
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
i
n
I
n
t
e
g
ra
t
i
n
g
D
i
g
i
t
a
l
H
e
a
l
t
h
S
t
r
a
t
e
g
i
e
s
f
o
r
Ef
f
e
c
t
i
v
e
Ad
m
i
n
i
st
r
a
t
i
o
n
,
I
G
I
G
l
o
b
a
l
S
c
i
e
n
t
i
f
i
c
P
u
b
l
i
s
h
i
n
g
,
2
0
2
3
,
p
p
.
3
4
5
–
377
,
d
o
i
:
1
0
.
4
0
1
8
/
9
7
8
-
1
-
6
6
8
4
-
8
3
3
7
-
4
.
c
h
0
1
6
.
[
1
4
]
K
.
P
e
r
u
m
a
l
l
a
,
A
.
S
o
n
i
,
R
.
D
e
y
,
a
n
d
S
.
R
i
c
h
,
“
Ze
r
o
I
n
:
c
h
a
r
a
c
t
e
r
i
z
i
n
g
t
h
e
d
a
t
a
d
i
s
t
r
i
b
u
t
i
o
n
s
o
f
c
o
mm
i
t
s
i
n
s
o
f
t
w
a
r
e
r
e
p
o
s
i
t
o
r
i
e
s,”
A
p
r
.
2
0
2
2
,
a
rX
i
v
:
2
2
0
4
.
0
7
8
6
3
.
[
1
5
]
U
.
R
.
A
c
h
a
r
y
a
e
t
a
l
.
,
“
A
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
m
o
d
e
l
t
o
c
l
a
s
si
f
y
h
e
a
r
t
b
e
a
t
s,
”
C
o
m
p
u
t
e
rs
i
n
Bi
o
l
o
g
y
a
n
d
M
e
d
i
c
i
n
e
,
v
o
l
.
8
9
,
p
p
.
3
8
9
–
3
9
6
,
O
c
t
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
b
i
o
m
e
d
.
2
0
1
7
.
0
8
.
0
2
2
.
[
1
6
]
T.
F
.
R
o
md
h
a
n
e
,
H
.
A
l
h
i
c
h
r
i
,
R
.
O
u
n
i
,
a
n
d
M
.
A
t
r
i
,
“
El
e
c
t
r
o
c
a
r
d
i
o
g
r
a
m
h
e
a
r
t
b
e
a
t
c
l
a
ssi
f
i
c
a
t
i
o
n
b
a
se
d
o
n
a
d
e
e
p
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
n
d
f
o
c
a
l
l
o
ss
,
”
C
o
m
p
u
t
e
rs
i
n
Bi
o
l
o
g
y
a
n
d
Me
d
i
c
i
n
e
,
v
o
l
.
1
2
3
,
A
u
g
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
b
i
o
me
d
.
2
0
2
0
.
1
0
3
8
6
6
.
[
1
7
]
Ö
.
Y
i
l
d
i
r
i
m
,
“
A
n
o
v
e
l
w
a
v
e
l
e
t
se
q
u
e
n
c
e
b
a
se
d
o
n
d
e
e
p
b
i
d
i
r
e
c
t
i
o
n
a
l
L
S
TM
n
e
t
w
o
r
k
m
o
d
e
l
f
o
r
E
C
G
si
g
n
a
l
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
i
n
B
i
o
l
o
g
y
a
n
d
Me
d
i
c
i
n
e
,
v
o
l
.
9
6
,
p
p
.
1
8
9
–
2
0
2
,
M
a
y
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
b
i
o
m
e
d
.
2
0
1
8
.
0
3
.
0
1
6
.
[
1
8
]
R
.
H
u
,
J.
C
h
e
n
,
a
n
d
L.
Z
h
o
u
,
“
A
t
r
a
n
s
f
o
r
mer
-
b
a
se
d
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
f
o
r
a
r
r
h
y
t
h
mi
a
d
e
t
e
c
t
i
o
n
u
si
n
g
c
o
n
t
i
n
u
o
u
s
EC
G
si
g
n
a
l
s
,
”
C
o
m
p
u
t
e
rs
i
n
B
i
o
l
o
g
y
a
n
d
Me
d
i
c
i
n
e
,
v
o
l
.
1
4
4
,
M
a
y
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
b
i
o
me
d
.
2
0
2
2
.
1
0
5
3
2
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.