I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
24
,
No
.
1
,
Octo
b
er
2
0
2
1
,
p
p
.
2
1
7
~
2
2
5
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
:
1
0
.
1
1
5
9
1
/ijeecs.v
24
.i
1
.
p
p
2
1
7
-
2
2
5
217
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Dete
c
tion o
f
ca
rd
i
a
c arrhy
thmia
usi
n
g
d
eep
C
NN and
optimized
SVM
M
o
hebb
a
na
a
z
,
Y
.
P
a
dm
a
Sa
i
,
L
.
V
.
Ra
j
a
ni K
um
a
ri
De
p
a
rtme
n
t
o
f
ECE
,
VN
R
Vig
n
a
n
a
Jy
o
t
h
i
I
n
stit
u
te o
f
En
g
in
e
e
rin
g
a
n
d
Tec
h
n
o
l
o
g
y
,
Tela
n
g
a
n
a
,
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
J
u
n
15
,
2
0
2
1
R
ev
is
ed
Au
g
19
,
2
0
2
1
Acc
ep
ted
Au
g
23
,
2
0
2
1
De
e
p
lea
rn
in
g
(DL)
h
a
s
b
e
c
o
m
e
a
to
p
ic
o
f
stu
d
y
in
v
a
ri
o
u
s
a
p
p
li
c
a
ti
o
n
s,
in
c
lu
d
in
g
h
e
a
lt
h
c
a
re
.
De
tec
ti
o
n
o
f
a
b
n
o
rm
a
li
ti
e
s
i
n
a
n
e
lec
tro
c
a
rd
io
g
ra
m
(ECG
)
p
lay
s
a
si
g
n
if
ica
n
t
r
o
le
in
p
a
ti
e
n
t
m
o
n
it
o
rin
g
.
It
is
n
o
te
d
t
h
a
t
a
d
e
e
p
n
e
u
ra
l
n
e
two
r
k
wh
e
n
train
e
d
o
n
h
u
g
e
d
a
ta,
c
a
n
e
a
sily
d
e
tec
t
c
a
rd
iac
a
rrh
y
th
m
ia.
T
h
is
m
a
y
h
e
lp
c
a
r
d
io
l
o
g
ists
t
o
sta
rt
trea
tme
n
t
a
s
e
a
rly
a
s
p
o
ss
ib
le.
Th
is
p
a
p
e
r
p
ro
p
o
se
s
a
n
e
w
d
e
e
p
lea
rn
i
n
g
m
o
d
e
l
a
d
a
p
ti
n
g
th
e
c
o
n
c
e
p
t
o
f
tran
sfe
r
lea
rn
i
n
g
to
e
x
trac
t
d
e
e
p
-
CNN
fe
a
tu
re
s
a
n
d
fa
c
il
it
a
tes
a
u
to
m
a
ted
c
las
sifica
t
io
n
o
f
e
lec
tro
c
a
rd
io
g
ra
m
(ECG
)
in
to
six
tee
n
ty
p
e
s
o
f
ECG
b
e
a
ts
u
sin
g
a
n
o
p
ti
m
ize
d
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(S
VM).
T
h
e
p
ro
p
o
se
d
stra
teg
y
b
e
g
i
n
s
wit
h
g
a
th
e
ri
n
g
E
CG
d
a
tas
e
ts,
re
m
o
v
a
l
o
f
n
o
ise
f
ro
m
ECG
sig
n
a
ls,
a
n
d
e
x
trac
ti
n
g
b
e
a
ts
fr
o
m
d
e
n
o
ise
d
ECG
sig
n
a
ls.
F
e
a
tu
r
e
e
x
trac
ti
o
n
is
d
o
n
e
u
si
n
g
Re
sN
e
t1
8
v
ia
c
o
n
c
e
p
t
o
f
tran
sfe
r
lea
rn
i
n
g
.
Th
e
s
e
e
x
trac
ted
fe
a
tu
re
s
a
re
c
las
sified
u
si
n
g
o
p
ti
m
ize
d
S
VM.
Th
e
se
m
e
th
o
d
s
a
re
e
v
a
lu
a
ted
a
n
d
tes
ted
o
n
t
h
e
M
I
T
-
BIH
a
rr
h
y
t
h
m
ia
d
a
tab
a
se
.
O
u
r
p
ro
p
o
se
d
m
o
d
e
l
is
e
ffe
c
ti
v
e
c
o
m
p
a
re
d
to
a
ll
S
tate
o
f
Art
Tec
h
n
iq
u
e
s
with
a
n
a
c
c
u
ra
c
y
o
f
9
8
.
7
0
%
.
K
ey
w
o
r
d
s
:
Dee
p
lear
n
in
g
Deep
C
NN
E
lectr
o
ca
r
d
io
g
r
am
Op
tim
ized
SVM
T
r
an
s
f
er
lear
n
i
n
g
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
:
Mo
h
eb
b
a
n
a
az
R
ese
ar
ch
Sch
o
lar
,
Dep
ar
tm
en
t
o
f
E
C
E
VNR Vig
n
an
a
J
y
o
th
i I
n
s
titu
t
e
o
f
E
n
g
i
n
ee
r
in
g
an
d
T
ec
h
n
o
lo
g
y
Hy
d
er
ab
ad
,
T
elan
g
a
n
a
-
5
0
0
0
9
0
,
I
n
d
ia
E
m
ail: m
o
h
eb
b
a
n
aa
z@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Hea
r
t
attac
k
s
ar
e
th
e
m
ajo
r
ca
u
s
e
o
f
d
ea
t
h
g
lo
b
ally
.
1
7
.
9
m
i
llio
n
p
eo
p
le
d
ie
ea
ch
y
ea
r
d
u
e
to
ca
r
d
iac
ab
n
o
r
m
alities
,
esti
m
atin
g
3
1
%
o
f
wo
r
ld
wid
e
d
ea
th
s
[
1
]
.
H
alf
o
f
th
e
ca
r
d
i
o
v
ascu
lar
d
ea
th
s
ar
e
d
u
e
to
ca
r
d
iac
ar
r
h
y
th
m
ias.
An
y
d
is
tu
r
b
a
n
c
e
o
r
ch
an
g
es
th
at
alter
th
e
n
o
r
m
al
f
u
n
ctio
n
in
g
o
f
th
e
h
ea
r
t
ca
u
s
es
ca
r
d
iac
ar
r
h
y
th
m
ia.
Do
cto
r
s
o
f
ten
r
ec
o
m
m
en
d
p
atien
ts
with
a
r
r
h
y
th
m
ia
to
wea
r
H
o
lter
f
o
r
co
n
tin
u
o
u
s
m
o
n
ito
r
in
g
o
f
E
C
G
f
o
r
2
4
h
o
u
r
s
.
As
th
is
r
ec
o
r
d
ed
d
ata
is
lar
g
e,
th
er
e
is
a
n
ec
ess
ity
to
ca
teg
o
r
ize
th
e
ty
p
e
o
f
ea
ch
h
ea
r
tb
ea
t
au
to
m
atica
lly
u
s
in
g
th
e
co
m
p
u
ter
-
aid
e
d
d
iag
n
o
s
is
to
o
l.
T
h
e
ad
v
an
ce
m
en
t
an
d
ad
o
p
tio
n
o
f
ar
tific
ial
in
tellig
en
c
e
h
av
e
s
u
r
g
ed
in
th
e
last
f
ew
y
ea
r
s
.
T
r
ad
itio
n
al
m
a
ch
in
e
lear
n
in
g
tech
n
iq
u
es
o
r
k
er
n
el
-
b
ased
n
eu
r
al
n
etwo
r
k
(
NN)
n
ee
d
s
d
o
m
ain
e
x
p
er
ts
to
id
en
tif
y
im
p
o
r
tan
t f
e
atu
r
es f
o
r
th
e
an
aly
s
is
o
f
E
C
G
wav
ef
o
r
m
[
2
]
.
R
esear
ch
er
s
h
av
e
p
r
o
p
o
s
ed
d
if
f
er
e
n
t
m
eth
o
d
o
lo
g
ies
f
o
r
th
e
class
if
icatio
n
o
f
a
r
r
h
y
th
m
ia
[
3
]
.
Mo
r
p
h
o
l
o
g
ical
an
d
tem
p
o
r
a
l
ch
ar
ac
ter
is
tics
ar
e
m
o
s
tly
co
n
s
id
er
ed
f
o
r
E
C
G
s
ig
n
al
an
aly
s
is
.
T
h
e
m
eth
o
d
o
l
o
g
ies
b
eg
i
n
with
t
h
e
co
llectio
n
o
f
d
ata
f
r
o
m
MI
T
-
B
I
H
ar
r
h
y
th
m
ia
d
atab
ase
[
4
]
,
p
r
e
-
p
r
o
ce
s
s
in
g
d
ata,
f
ea
tu
r
e
ex
tr
ac
ti
o
n
an
d
class
if
icatio
n
.
E
C
G
s
i
g
n
a
l
is
a
f
f
ec
t
e
d
b
y
v
a
r
i
o
u
s
t
y
p
e
s
o
f
n
o
i
s
es
l
i
k
e
p
o
w
e
r
li
n
e
i
n
t
e
r
f
e
r
e
n
c
e
,
b
as
e
l
i
n
e
w
a
n
d
e
r
,
M
o
t
i
o
n
a
r
t
i
f
a
ct
s
a
n
d
E
M
G
n
o
is
e
[
3
]
.
T
h
e
s
e
n
o
is
e
s
a
r
e
r
e
m
o
v
ed
b
y
p
r
e
-
p
r
o
c
e
s
s
i
n
g
t
h
e
d
a
t
a
u
s
i
n
g
D
W
T
[
5
]
,
[
6
]
,
F
il
t
e
r
i
n
g
[
7
]
,
[
8
]
.
T
h
e
D
e
n
o
is
e
d
E
C
G
s
i
g
n
al
is
f
u
r
t
h
e
r
c
o
n
s
i
d
e
r
e
d
f
o
r
f
e
at
u
r
e
e
x
t
r
a
c
t
i
o
n
.
T
i
m
e
d
o
m
a
i
n
a
n
a
ly
s
i
s
[
9
]
,
f
r
e
q
u
e
n
c
y
d
o
m
a
i
n
a
n
a
l
y
s
is
[
1
0
]
,
[
1
1
]
,
t
i
m
e
-
f
r
e
q
u
e
n
c
y
a
n
a
l
y
s
is
[
1
2
]
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
1
,
Octo
b
er
2
0
2
1
:
217
-
2
2
5
218
s
tatis
t
ics
-
b
ased
an
aly
s
is
[
1
3
]
an
d
h
y
b
r
id
f
ea
tu
r
e
-
b
ased
m
et
h
o
d
s
[
1
4
]
is
d
o
n
e
f
o
r
f
ea
tu
r
e
ex
tr
ac
tio
n
.
T
h
ese
ex
tr
ac
ted
f
ea
tu
r
es
ca
n
b
e
cla
s
s
if
ied
u
s
in
g
ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
(
ANN)
[
1
5
]
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es
(
SVM)
[
1
6
]
-
[
2
1
]
,
Naïv
e
B
ay
es
[
2
2
]
,
d
ec
is
io
n
tr
ee
(
DT
)
[
2
3
]
,
K
-
n
ea
r
est
n
eig
h
b
o
r
(
KNN)
[
2
4
]
,
a
n
d
lin
ea
r
d
is
cr
im
in
an
t a
lg
o
r
ith
m
s
(
L
DA)
[
2
5
]
.
Ma
ch
in
e
lear
n
in
g
tec
h
n
iq
u
es
r
eq
u
ir
e
f
ea
tu
r
e
ex
tr
ac
tio
n
,
wh
ich
in
d
ee
d
n
ee
d
s
a
d
o
m
ai
n
ex
p
er
t
to
id
en
tify
ef
f
icien
t
f
ea
tu
r
es
[
2
6
]
,
[
27]
.
Dee
p
lear
n
i
n
g
m
o
d
els
in
cr
em
en
tally
lear
n
f
ea
tu
r
es
with
o
u
t
th
e
n
ee
d
f
o
r
d
o
m
ain
e
x
p
er
ts
[
2
8
]
,
[
29]
.
D
ee
p
lear
n
in
g
m
o
d
els
ar
e
lay
e
r
ed
ar
ch
itectu
r
es
th
at
lear
n
f
ea
tu
r
es
at
d
if
f
er
e
n
t
lay
er
s
.
I
n
p
u
t
lay
e
r
s
ex
tr
ac
t
f
ea
tu
r
es
u
s
in
g
th
e
co
n
v
o
lu
tio
n
al
lay
er
,
an
d
th
ese
f
ea
tu
r
es
ar
e
an
aly
z
ed
an
d
class
if
ied
u
s
in
g
o
u
tp
u
t la
y
e
r
s
.
I
t
is
o
b
s
er
v
e
d
f
r
o
m
th
e
liter
atu
r
e
th
at,
as
we
i
n
cr
ea
s
e
th
e
n
u
m
b
er
o
f
E
C
G
class
es
f
o
r
cla
s
s
if
icatio
n
,
th
e
ac
cu
r
ac
y
d
ec
r
ea
s
es.
T
h
e
m
ain
p
r
o
b
lem
o
b
s
er
v
ed
is
to
o
b
tain
b
esy
ac
cu
r
ac
y
wh
e
n
m
o
r
e
ab
n
o
r
m
al
class
es
ar
e
co
n
s
id
er
ed
.
Mo
s
t
o
f
th
e
m
eth
o
d
o
lo
g
ies
r
eq
u
ir
e
d
o
m
ain
e
x
p
er
ts
f
o
r
e
x
tr
ac
tin
g
f
ea
tu
r
es.
T
o
av
o
id
t
h
e
n
e
e
d
f
o
r
d
o
m
ain
ex
p
er
ts
,
o
u
r
wo
r
k
m
ain
ly
em
p
h
ases
ex
tr
ac
tin
g
f
ea
tu
r
es
u
s
in
g
a
tr
a
n
s
f
er
le
ar
n
in
g
-
b
ased
C
NN
m
o
d
el
-
R
esNet1
8
.
R
est
o
f
th
e
p
ap
er
is
s
y
s
tem
atize
d
as
f
o
llo
ws.
Sectio
n
2
d
escr
ib
es
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
,
in
clu
d
in
g
p
r
e
-
p
r
o
ce
s
s
in
g
(
r
e
m
o
v
al
o
f
n
o
is
e
an
d
E
C
G
b
e
at
s
eg
m
en
tatio
n
)
,
f
ea
t
u
r
e
ex
t
r
ac
tio
n
(
u
s
in
g
Dee
p
C
NN
-
R
e
s
Net1
8
m
o
d
el)
,
a
n
d
class
if
icatio
n
(
u
s
in
g
o
p
tim
ize
d
SVM)
.
I
n
s
ec
tio
n
3
,
in
v
esti
g
ativ
e
r
esu
lts
ar
e
p
r
esen
ted
an
d
co
m
p
ar
ed
with
t
h
e
ex
is
tin
g
s
tate
-
of
-
ar
t
m
o
d
el.
Sectio
n
4
co
n
cl
u
d
es o
u
r
wo
r
k
.
2.
M
E
T
H
O
DO
L
O
G
Y
T
h
e
p
u
r
p
o
s
e
o
f
t
h
i
s
s
t
u
d
y
i
s
t
o
d
e
t
e
c
t
a
b
n
o
r
m
a
l
i
t
i
e
s
i
n
E
C
G
s
i
g
n
a
l
s
.
T
r
a
n
s
f
e
r
r
i
n
g
t
h
e
l
e
a
r
n
e
d
k
n
o
w
l
e
d
g
e
u
t
i
l
i
z
i
n
g
p
r
e
-
t
r
a
i
n
e
d
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
s
e
m
p
o
w
e
r
s
u
s
t
o
g
a
t
h
e
r
i
n
f
o
r
m
a
t
i
o
n
r
e
q
u
i
r
e
d
t
o
s
t
u
d
y
n
e
w
d
a
t
a
.
W
e
h
e
r
e
e
x
p
l
o
r
e
d
t
h
e
d
e
e
p
l
e
a
r
n
i
n
g
m
o
d
e
l
-
R
e
s
N
e
t
1
8
a
n
d
e
x
t
r
a
c
t
ed
i
n
t
r
i
n
s
i
c
D
e
e
p
-
C
N
N
f
e
a
t
u
r
e
s
.
U
t
i
l
i
z
i
n
g
t
h
e
s
e
f
e
a
t
u
r
e
s
c
l
a
s
s
i
f
i
c
a
t
i
o
n
i
s
d
o
n
e
u
s
i
n
g
o
p
t
i
m
i
z
e
d
-
S
V
M
.
F
i
g
u
r
e
1
r
e
p
r
e
s
e
n
t
s
t
h
e
m
e
t
h
o
d
o
l
o
g
y
o
f
o
u
r
p
r
o
p
o
s
e
d
w
o
r
k
.
Fig
u
r
e
1
.
B
lo
ck
d
iag
r
am
o
f
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
2
.
1
.
P
re
-
pro
ce
s
s
ing
da
t
a
E
C
G
s
ig
n
als
ar
e
tak
en
f
r
o
m
MI
T
-
B
I
H
ar
r
h
y
th
m
ia
d
atab
as
e
[
5]
.
I
t
co
n
tain
s
4
8
r
ec
o
r
d
s
,
e
ac
h
with
3
0
m
in
u
tes
d
u
r
atio
n
.
All
r
ec
o
r
d
s
to
g
eth
e
r
c
o
n
tain
1
6
ty
p
es
o
f
b
ea
ts
,
o
f
wh
ich
o
n
e
is
n
o
r
m
al,
1
4
class
es
o
f
ar
r
h
y
th
m
ia
,
an
d
o
n
e
u
n
class
if
ied
b
e
at.
T
y
p
es
o
f
b
ea
ts
co
n
s
id
er
ed
ar
e
n
o
r
m
al
b
ea
ts
(
N)
,
lef
t
b
u
n
d
le
b
r
an
ch
b
lo
ck
b
ea
t
s
(
L
)
,
r
ig
h
t
b
u
n
d
le
b
r
an
c
h
b
l
o
ck
b
ea
ts
(
R
)
,
p
r
e
m
atu
r
e
v
e
n
tr
icu
lar
c
o
n
tr
ac
tio
n
b
ea
ts
(
V)
,
atr
ial
p
r
em
atu
r
e
b
e
ats
(
A)
,
p
ac
e
d
b
ea
ts
(
P),
f
u
s
io
n
o
f
n
o
r
m
al
an
d
p
ac
ed
b
ea
t
(
f
)
,
f
u
s
i
o
n
o
f
v
e
n
t
r
icu
lar
an
d
n
o
r
m
al
be
at
(
F),
v
en
tr
icu
lar
f
lu
tter
wav
e
(
!
)
,
n
o
d
al
(
ju
n
ctio
n
al)
es
ca
p
e
b
ea
t
(
j)
,
b
lo
ck
e
d
atr
ial
p
r
em
atu
r
e
b
ea
t
(
X)
,
ab
er
r
ated
atr
ial
p
r
em
atu
r
e
b
ea
t
(
a)
,
v
en
t
r
icu
lar
escap
e
b
ea
t
(
E
)
,
n
o
d
al
(
ju
n
ctio
n
al)
p
r
em
at
u
r
e
b
ea
t
(
J
)
,
atr
ial
escap
e
b
ea
t
(
e)
,
u
n
class
if
ied
b
ea
t
(
Q)
.
2
.
1
.
1
.
Rem
o
v
a
l o
f
n
o
is
e
E
C
G
s
ig
n
als
ar
e
m
o
s
t
f
r
eq
u
en
tly
af
f
ec
ted
b
y
v
ar
i
o
u
s
n
o
is
es
lik
e
p
o
wer
lin
e
in
ter
f
er
e
n
c
e,
b
aselin
e
d
r
if
ts
,
m
o
tio
n
ar
tifa
cts
an
d
elec
tr
o
m
y
o
g
r
ap
h
y
(
E
MG
)
n
o
is
e.
I
n
th
is
wo
r
k
,
n
o
is
e
is
r
em
o
v
ed
u
s
in
g
Sav
itzk
y
-
Go
lay
f
ilter
s
.
T
h
ese
f
ilter
s
s
m
o
o
th
o
u
t
n
o
is
y
s
ig
n
als
with
a
l
ar
g
e
f
r
e
q
u
en
c
y
s
p
an
.
I
n
g
e
n
er
al,
f
ilter
in
g
r
ep
lace
s
ea
ch
v
alu
e
o
f
a
s
ig
n
al
with
p
o
in
ts
th
at
m
ea
s
u
r
e
n
ea
r
ly
th
e
s
am
e
u
n
d
er
l
y
in
g
v
alu
e
.
I
n
m
o
v
in
g
av
er
ag
e
f
ilter
s
,
to
tal
win
d
o
w
len
g
th
o
f
=
2
+
1
s
am
p
les
is
r
ep
lace
d
b
y
its
av
er
ag
e
v
alu
e
as
;
EC
G
B
e
a
t
S
e
g
me
n
t
a
t
i
o
n
P
e
ak
D
e
t
e
ct
i
o
n
S
e
g
me
n
t
e
d
B
e
a
t
ECG
Sig
n
al
M
I
T
-
B
I
H
D
a
t
a
b
a
se
Fe
at
u
r
e
E
x
t
r
ac
t
i
o
n
R
e
sN
e
t
1
8
O
p
t
i
m
i
z
e
d
S
V
M
C
l
a
ssi
f
i
c
a
t
i
o
n
Re
m
o
v
al
o
f
N
o
i
se
D
e
n
o
i
se
d
Sig
n
al
N
o
i
sy
S
i
g
n
a
l
E
C
G
2
D I
ma
g
e
D
a
t
a
b
a
se
S
V
M
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Dete
ctio
n
o
f c
a
r
d
ia
c
a
r
r
h
yth
m
ia
u
s
in
g
d
ee
p
C
N
N
a
n
d
o
p
timiz
ed
S
V
M
(
Mo
h
eb
b
a
n
a
a
z
)
219
→
̂
=
1
∑
+
=
−
(1
)
Sav
itzk
y
-
Go
lay
f
ilter
s
g
e
n
er
al
ize
th
is
id
ea
b
y
least
-
s
q
u
ar
es
f
itti
n
g
a
ℎ
-
o
r
d
er
p
o
ly
n
o
m
ial
t
h
r
o
u
g
h
th
e
s
ig
n
al
v
alu
es
in
t
h
e
win
d
o
w
an
d
tak
in
g
th
e
ca
lc
u
lated
c
en
tr
al
p
o
in
t
o
f
th
e
f
itted
p
o
ly
n
o
m
ial
cu
r
v
e
as
th
e
n
ew
s
m
o
o
th
ed
d
ata
p
o
i
n
t.
L
et
th
e
d
eg
r
ee
p
o
ly
n
o
m
ial
co
n
s
id
er
ed
as
(
+
)
=
∑
(
)
(
)
!
=
0
(2
)
Af
ter
co
n
v
o
lv
in
g
th
e
p
o
ly
n
o
m
ial
an
d
r
ea
r
r
a
n
g
in
g
th
e
s
u
m
m
a
tio
n
s
we
g
et
(
+
)
=
∑
∑
(
)
(
)
!
=
0
=
−
=
∑
(
)
(
)
!
=
0
∑
=
−
(3
)
Fo
llo
win
g
eq
u
atio
n
m
u
s
t b
e
s
a
tis
f
ied
f
o
r
r
ep
r
o
d
u
cin
g
t
h
e
v
al
u
es o
f
d
eg
r
ee
p
.
∑
=
−
=
{
1
=
0
0
=
1
,
2
,
…
.
≠
0
>
(4
)
2
.
1
.
2
.
E
CG
bea
t
s
eg
m
ent
a
t
io
n
C
o
n
s
id
er
in
g
n
o
is
e
-
f
r
ee
s
ig
n
al,
th
e
n
e
x
t
s
tep
is
to
ex
tr
ac
t
B
ea
ts
.
MI
T
-
B
I
H
ar
r
h
y
th
m
ia
d
atab
ase
also
p
r
o
v
id
e
d
a
n
n
o
tatio
n
f
i
les
wit
h
th
is
d
atab
ase.
T
h
ey
r
ec
o
r
d
ed
th
e
lo
ca
tio
n
o
f
R
-
p
ea
k
an
d
th
e
ty
p
e
o
f
ea
c
h
h
ea
r
tb
ea
t.
R
p
ea
k
s
ar
e
id
en
tif
ied
u
s
in
g
an
n
o
tatio
n
f
iles
.
T
h
e
a
v
er
ag
e
R
R
in
ter
v
al
is
ca
lc
u
lated
b
y
av
e
r
ag
in
g
p
r
eRR
in
ter
v
al
an
d
p
o
s
tR
R
in
ter
v
al.
An
o
n
s
et
d
u
r
atio
n
o
f
1
/3
r
d
o
f
a
v
er
ag
e
in
ter
v
al
an
d
o
f
f
s
et
d
u
r
atio
n
o
f
2
/3
r
d
o
f
av
e
r
ag
e
in
ter
v
al
is
s
elec
ted
to
ex
tr
ac
t
o
n
e
E
C
G
b
ea
t.
Simu
lated
r
esu
lts
o
f
n
o
is
e
r
em
o
v
al
a
n
d
b
ea
t
s
eg
m
en
tatio
n
p
r
o
ce
d
u
r
e
ar
e
s
h
o
wn
in
s
ec
tio
n
3
.
T
h
ese
b
ea
ts
ar
e
f
u
r
th
e
r
co
n
s
id
er
ed
f
o
r
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
2
.
2
.
F
ea
t
ure
ex
t
ra
ct
io
n t
hro
ug
h t
ra
ns
f
er
lea
rning
T
r
ain
in
g
a
d
ee
p
n
etwo
r
k
f
r
o
m
s
cr
atch
with
d
ata
s
ca
r
city
is
ti
m
e
-
co
n
s
u
m
in
g
.
An
a
lter
n
ativ
e
ap
p
r
o
ac
h
is
to
u
s
e
th
e
co
n
ce
p
t
o
f
tr
an
s
f
er
lear
n
in
g
to
ex
tr
ac
t
f
ea
tu
r
es.
Dee
p
m
o
d
els
ex
tr
ac
ts
f
ea
tu
r
e
s
b
y
u
tili
zin
g
co
n
v
o
l
u
tio
n
al
lay
er
s
.
T
h
ese
lay
er
s
h
av
e
a
g
r
o
u
p
o
f
f
ilter
s
th
at
co
n
v
o
lv
es
with
lay
er
k
e
r
n
els
to
g
en
er
ate
a
ten
s
o
r
o
f
f
ea
tu
r
es.
Dep
en
d
in
g
o
n
th
e
'
s
tr
id
e
'
f
ilter
wi
ll
m
o
v
e
f
r
o
m
o
n
e
p
o
in
t
to
th
e
n
e
x
t
in
ea
ch
p
r
o
g
r
ess
io
n
.
Z
er
o
p
ad
d
in
g
s
ar
e
d
o
n
e
to
t
h
e
f
ilter
s
if
th
e
co
n
v
o
lu
tio
n
al
lay
er
d
o
es
n
o
t
co
v
er
t
h
e
e
n
tire
E
C
G
b
ea
t.
E
ac
h
lay
er
o
f
a
C
NN
Mo
d
el
p
r
o
d
u
ce
s
a
r
esp
o
n
s
e,
o
r
ac
tiv
atio
n
,
f
o
r
a
g
iv
en
i
n
p
u
t
b
ea
t.
Ho
wev
e
r
,
o
n
ly
a
f
ew
lay
er
s
with
in
a
C
N
N
Netwo
r
k
ar
e
s
u
itab
le
f
o
r
ex
tr
ac
tin
g
f
ea
tu
r
e
s
.
T
h
e
in
itial
lay
er
s
o
f
th
e
n
et
wo
r
k
ca
p
tu
r
e
b
asic
E
C
G
f
ea
tu
r
es,
s
u
ch
as
ed
g
es,
s
lo
p
es
an
d
c
u
r
v
es.
As
m
o
r
e
lay
er
s
ar
e
ad
d
e
d
d
e
ep
f
ea
tu
r
e
s
lik
e
s
m
all
cu
r
v
es,
s
lo
p
es
g
et
ex
tr
ac
ted
.
T
h
is
b
u
i
ld
s
u
p
in
tu
itio
n
as
to
wh
y
th
e
f
ea
tu
r
es
ex
tr
ac
ted
f
r
o
m
C
NNs
wo
r
k
s
o
well
f
o
r
d
etec
tin
g
ab
n
o
r
m
al
E
C
G
s
ig
n
als.
T
h
e
ten
s
o
r
o
f
f
ea
tu
r
e
m
ap
s
g
en
er
ate
d
is
d
eter
m
i
n
e
d
th
r
o
u
g
h
a
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U)
as
an
ac
tiv
atio
n
f
u
n
ctio
n
.
C
o
n
v
o
lu
tio
n
o
p
er
atio
n
b
etwe
en
c
o
n
v
o
lu
tio
n
al
f
ilter
s
an
d
k
er
n
els
is
d
ef
in
ed
as
=
(
∑
−
1
∗
+
∈
)
(5
)
is
a
r
an
g
e
o
f
co
n
v
o
lu
tio
n
k
er
n
els
with
k
er
n
el
s
ize
5
o
r
3
,
g
iv
es
o
u
tp
u
t
o
f
n
e
u
r
o
n
at
lay
er
,
is
th
e
b
ias
o
f
n
eu
r
o
n
,
is
th
e
weig
h
t
k
er
n
el
n
eu
r
o
n
at
la
y
er
−
1
an
d
n
eu
r
o
n
at
lay
er
.
(
.
)
r
ep
r
esen
ts
ac
tiv
atio
n
f
u
n
ctio
n
R
eL
U.
W
e
h
av
e
ex
p
lo
r
ed
C
NN
ar
ch
itectu
r
e
r
esn
et1
8
f
o
r
ex
tr
ac
tin
g
f
ea
tu
r
es
.
T
h
e
lay
er
e
d
a
r
ch
itectu
r
e
o
f
R
esn
et1
8
an
d
in
ter
n
al
ar
ch
it
ec
tu
r
e
o
f
r
esid
u
al
b
lo
c
k
s
is
s
h
o
wn
in
Fig
u
r
e
2
.
R
esn
et1
8
h
as
o
n
e
co
n
v
o
lu
tio
n
lay
er
,
eig
h
t
r
esid
u
al
b
lo
ck
s
a
n
d
o
n
e
f
u
lly
co
n
n
ec
te
d
lay
er
.
E
ac
h
r
esid
u
al
b
lo
ck
h
as
two
co
n
v
o
lu
tio
n
al
lay
er
s
i.e
.
,
R
esNet1
8
h
as
a
to
tal
o
f
1
8
lay
er
s
.
R
esNet
-
1
8
h
as
b
ee
n
tr
ain
ed
u
s
in
g
t
h
e
I
m
ag
eNe
t
d
ataset
o
n
m
o
r
e
th
an
a
m
illi
o
n
im
a
g
es
an
d
ca
n
class
if
y
im
ag
es
in
to
1
0
0
0
o
b
j
ec
t
ca
teg
o
r
ies.
T
h
e
in
p
u
t
lay
e
r
r
e
q
u
ir
es
an
im
ag
e
o
f
s
ize
2
2
4
x
2
2
4
f
o
r
tr
ain
in
g
.
Hen
ce
,
d
ata
au
g
u
m
en
tatio
n
is
p
er
f
o
r
m
ed
.
I
n
th
is
s
tu
d
y
,
o
n
-
t
h
e
-
f
ly
d
ata
au
g
m
e
n
tatio
n
is
p
er
f
o
r
m
ed
o
n
ea
ch
b
ea
t
b
ef
o
r
e
e
x
tr
ac
tin
g
f
e
atu
r
es.
T
h
is
tech
n
iq
u
e
g
en
er
ates
m
o
r
e
ac
c
u
r
ate
an
d
r
o
b
u
s
t
b
ea
t
with
s
p
ec
if
ie
d
s
ize
an
d
av
o
id
s
lo
s
s
o
f
in
f
o
r
m
ati
o
n
.
A
u
g
u
m
e
n
tatio
n
r
ed
u
ce
s
th
e
r
is
k
o
f
o
v
er
f
itti
n
g
d
r
asti
ca
lly
.
T
h
is
au
g
m
en
te
d
d
atab
ase
a
d
ap
t
s
R
esNet1
8
co
n
f
ig
u
r
atio
n
.
T
h
e
co
n
v
o
l
u
tio
n
k
er
n
el
s
ize
an
d
a
ctiv
atio
n
s
in
th
e
r
esid
u
al
b
lo
c
k
s
o
f
R
esNet1
8
r
em
ain
u
n
ch
a
n
g
ed
.
T
h
e
ex
is
tin
g
f
u
lly
-
co
n
n
ec
ted
lay
er
with
1
0
0
0
n
o
d
es
is
d
is
ca
r
d
ed
an
d
an
ac
tiv
atio
n
R
eL
U
is
u
s
ed
af
ter
th
e
p
o
o
lin
g
lay
er
.
Fo
r
a
wid
e
r
an
g
e
o
f
im
a
g
es,
th
e
R
esNet1
8
m
o
d
el
h
as
lear
n
ed
f
ea
tu
r
e
s
.
Ad
ap
tin
g
th
e
s
am
e
weig
h
ts
an
d
ac
tiv
atio
n
s
,
th
e
m
o
d
el
lear
n
s
r
eq
u
ir
ed
f
ea
tu
r
es.
T
h
e
f
in
al
p
o
o
lin
g
lay
er
p
o
o
ls
th
e
in
p
u
t f
ea
t
u
r
es o
v
er
all
s
p
atial
lo
ca
tio
n
s
,
g
iv
in
g
-
by
-
f
ea
tu
r
e
s
et
in
to
tal.
r
ep
r
esen
ts
th
e
to
ta
l
n
u
m
b
e
r
o
f
o
b
s
er
v
atio
n
s
an
d
r
ep
r
e
s
en
ts
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
1
,
Octo
b
er
2
0
2
1
:
217
-
2
2
5
220
n
u
m
b
er
o
f
f
ea
tu
r
es
ex
tr
ac
ted
f
o
r
ea
ch
o
b
s
er
v
atio
n
.
Po
o
lin
g
l
ay
er
h
as
5
1
2
n
o
d
es
wh
ich
af
te
r
ac
tiv
atio
n
s
g
iv
es
5
1
2
f
ea
t
u
r
es.
T
o
tal
o
b
s
er
v
atio
n
s
co
n
s
id
er
ed
i
n
th
is
s
tu
d
y
ar
e
4
5
9
4
3
.
Hen
ce
,
a
Featu
r
e
s
et
o
f
s
ize
4
5
9
4
3
-
by
-
5
1
2
is
e
x
tr
ac
ted
.
I
llu
s
tr
atio
n
o
f
Featu
r
e
ex
t
r
ac
tio
n
u
s
in
g
tr
a
n
s
f
er
lear
n
i
n
g
Via
R
esNet1
8
is
s
h
o
wn
in
Fig
u
r
e
3
.
T
h
e
Featu
r
es e
x
tr
ac
ted
u
s
in
g
t
h
is
m
o
d
el
ar
e
f
u
r
th
er
g
iv
en
f
o
r
class
if
icatio
n
.
(
a)
L
ay
er
e
d
ar
ch
itectu
r
e
(
b
)
R
esid
u
al
b
lo
ck
Fig
u
r
e
2
.
Ar
c
h
itectu
r
e
o
f
R
esn
et1
8
p
r
etr
ai
n
ed
m
o
d
el
Fig
u
r
e
3
.
I
ll
u
s
tr
atio
n
o
f
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
class
if
icatio
n
u
s
in
g
tr
an
s
f
er
lear
n
in
g
v
ia
R
esNet1
8
an
d
o
p
tim
ized
SVM
2
.
3
.
Cla
s
s
if
ica
t
io
n us
ing
o
pt
im
ized
-
SVM
T
h
e
f
ea
tu
r
e
s
et
e
x
tr
ac
ted
o
f
s
ize
4
5
9
4
3
-
by
-
5
1
2
h
as
in
p
u
t
d
ata
=
{
1
,
2
,
…
.
}
,
=
{
1
,
2
,
…
.
}
with
lear
n
in
g
tar
g
ets
=
{
1
,
2
,
…
.
}
.
T
h
is
f
ea
tu
r
e
s
et
is
s
p
lit
in
to
tr
ain
in
g
an
d
test
in
g
f
ea
tu
r
e
s
ets.
Usi
n
g
th
e
tr
ain
in
g
f
ea
tu
r
e
s
et
th
e
SVM
is
tr
ain
ed
to
o
n
tain
h
y
p
e
r
p
l
an
e
e
x
p
r
ess
ed
as
+
=
0
.
Her
e
is
th
e
n
u
m
b
er
o
f
la
b
els/
clas
s
es,
is
n
o
r
m
al
v
ec
to
r
a
n
d
is
th
e
in
ter
ce
p
t
o
f
h
y
p
er
p
lan
e.
W
h
en
m
in
im
u
m
o
f
‖
‖
is
ac
h
iev
ed
th
en
th
is
p
lan
e
b
ec
o
m
es
o
p
tim
u
m
h
y
p
e
r
p
lan
e.
T
o
av
o
id
o
v
er
f
itti
n
g
s
to
ch
asti
c
g
r
ad
ien
t
d
escen
t
(
SGD)
o
p
tim
izer
is
im
p
lem
en
ted
to
m
in
im
ize
th
e
o
b
ject
iv
e
f
u
n
ctio
n
.
T
h
e
o
b
jectiv
e
f
u
n
ctio
n
o
f
SVM
class
if
ier
is
g
iv
en
as
=
1
2
+
∑
∈
=
1
(6
)
wh
er
e
∈
is
r
elax
atio
n
v
a
r
iab
le.
T
h
e
lo
s
s
f
u
n
ctio
n
m
ea
s
u
r
es
t
h
e
f
it
b
etwe
en
th
e
ac
tu
al
an
d
p
r
ed
icted
v
al
u
es.
As
th
e
lo
s
s
f
u
n
ctio
n
d
ec
r
ea
s
es,
r
o
b
u
s
tn
ess
in
cr
ea
s
es.
T
h
e
lo
s
s
f
u
n
ctio
n
is
t
h
e
cr
u
cial
Me
asu
r
e
f
o
r
a
m
o
d
el
to
co
n
v
er
g
e.
C
o
n
s
id
er
in
g
th
is
co
n
s
tr
ain
,
t
h
e
o
b
jectiv
e
f
u
n
ctio
n
o
f
SVM
b
e
co
m
es
;
=
1
2
+
∑
[
0
,
1
−
]
=
1
(7
)
M
a
x
p
o
o
l
3
X3
c
o
n
v
,
6
4
,
s
t
r
i
d
e
2
M
a
x
p
o
o
l
F
u
l
l
y
c
o
n
n
e
c
t
e
d
l
a
y
e
r
R
e
si
d
u
a
l
B
l
o
c
k
s
D
a
t
a
F
l
o
w
(a
)
L
a
y
e
re
d
ar
c
h
itec
tu
re
3
X
3
c
o
n
v
,
1
2
8
,
3
X
3
c
o
n
v
,
1
2
8
R
e
LU
R
e
LU
R
e
LU
(b
)
Re
s
id
u
a
l
Blo
c
k
M
a
x
p
o
o
l
3
X3
c
o
n
v
,
6
4
,
s
t
r
i
d
e
2
M
a
x
p
o
o
l
F
u
l
l
y
c
o
n
n
e
c
t
e
d
l
a
y
e
r
R
e
si
d
u
a
l
B
l
o
c
k
s
D
a
t
a
F
l
o
w
(a
)
L
a
y
e
re
d
ar
c
h
itec
tu
re
3
X
3
c
o
n
v
,
1
2
8
,
3
X
3
c
o
n
v
,
1
2
8
R
e
LU
R
e
LU
R
e
LU
(b
)
Re
s
id
u
a
l
Blo
c
k
Re
siz
e
F
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
u
si
n
g
R
e
sN
e
t
1
8
F
e
a
t
u
r
e
S
e
t
T
e
st
S
e
t
T
r
a
i
n
i
n
g
S
e
t
L
e
a
r
n
i
n
g
A
l
g
o
r
i
t
h
m
G
e
n
e
r
at
e
d
M
o
d
e
l
C
l
a
ssi
f
i
c
a
t
i
o
n
o
u
t
p
u
t
ECG
2
D
I
m
ag
e
D
a
t
a
b
a
se
A
u
g
m
e
n
t
e
d
D
a
t
a
b
a
se
T
r
an
sf
e
r
L
e
a
r
n
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Dete
ctio
n
o
f c
a
r
d
ia
c
a
r
r
h
yth
m
ia
u
s
in
g
d
ee
p
C
N
N
a
n
d
o
p
timiz
ed
S
V
M
(
Mo
h
eb
b
a
n
a
a
z
)
221
L
o
ca
l
o
p
tim
u
m
o
f
SVM
is
f
o
u
n
d
o
u
t
u
s
in
g
SGD
alg
o
r
ith
m
.
T
o
u
p
d
ate
th
e
p
ar
am
eter
s
s
a
m
p
les
ar
e
r
an
d
o
m
l
y
s
elec
ted
.
Acc
o
r
d
i
n
g
ly
,
th
e
o
b
jectiv
e
f
u
n
cti
o
n
is
f
o
r
m
u
lated
as
;
=
1
2
+
m
a
x
[
0
,
1
−
]
(8
)
Pro
ce
s
s
o
f
d
ataset
lear
n
in
g
b
e
co
m
es
v
er
y
f
ast
wh
en
s
am
p
le
g
r
ad
ien
t
is
u
s
ed
to
u
p
d
ate
g
r
a
d
ien
t
.
T
h
e
g
r
ad
ien
t is n
o
te
d
as f
o
llo
ws.
=
−
[
>
1
]
(9
)
Her
e
s
ec
o
n
d
ter
m
is
an
in
d
icat
or
f
u
n
ctio
n
,
wh
ich
b
ec
o
m
es
1
wh
en
>
1
o
r
else it
b
ec
o
m
es
0
.
I
ter
atio
n
s
ca
n
b
e
d
ef
in
e
d
as
;
+
1
=
−
(1
0
)
=
0
1
+
0
,
0
=
1
(1
1
)
wh
er
e
is
th
e
s
tr
id
e,
th
at
d
eter
m
in
es
th
e
tim
e
elap
s
ed
b
y
m
o
d
el
to
r
ea
ch
th
e
o
p
tim
al
v
alu
e.
Af
ter
tr
ai
n
in
g
th
e
SVM
class
if
ier
,
it
is
test
ed
u
s
in
g
test
in
g
f
ea
tu
r
e
s
et.
T
o
ju
s
tify
th
e
ef
f
ec
tiv
en
ess
o
f
o
p
ti
m
izatio
n
,
we
h
a
v
e
class
if
ied
E
C
G
b
ea
ts
u
s
in
g
SV
M
an
d
th
e
c
o
m
p
ar
e
d
its
p
er
f
o
r
m
an
ce
with
o
p
tim
ized
SVM.
3.
RE
SU
L
T
S
A
ND
D
I
SCU
SS
I
O
NS
I
n
th
is
s
tu
d
y
,
MI
T
-
B
I
H
ar
r
h
y
th
m
ia
d
atab
ase
is
co
n
s
id
er
ed
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
is
s
im
u
lated
an
d
test
ed
u
s
in
g
MA
T
L
AB
s
o
f
t
war
e
(
2
0
2
0
b
)
p
ac
k
a
g
e
in
s
t
alled
o
n
W
in
d
o
ws
1
0
p
lat
f
o
r
m
with
s
y
s
tem
co
n
f
ig
u
r
atio
n
I
n
tel
c
o
r
e
i7
-
9
t
h
Gen
,
3
.
6
GHz
C
PU,
1
6
G
B
R
AM
,
an
d
NVI
DI
A
GeFo
r
ce
GT
X
1
0
6
0
4
GB
GPU.
T
h
e
s
im
u
lated
r
esu
lts
a
r
e
s
h
o
wn
an
d
d
is
cu
s
s
ed
in
th
e
f
o
llo
win
g
s
ec
tio
n
s
.
No
is
e
i
s
r
em
o
v
ed
u
s
in
g
S
-
Go
lay
Fil
ter
in
g
m
en
tio
n
ed
in
s
ec
tio
n
2
.
1
.
1
.
E
C
G
Sig
n
al
af
ter
r
em
o
v
al
o
f
No
is
e
f
r
o
m
No
i
s
y
E
C
G
s
ig
n
al
is
as
s
h
o
wn
in
Fig
u
r
e
4
.
Fig
u
r
e
4
.
R
em
o
v
al
o
f
n
o
is
e
f
r
o
m
E
C
G
s
ig
n
al
As
ea
ch
r
ec
o
r
d
h
as
d
if
f
er
e
n
t
t
y
p
es
o
f
E
C
G
b
ea
ts
,
we
n
ee
d
to
s
eg
m
en
t
th
e
b
ea
ts
an
d
lab
el
as
p
er
th
e
ab
n
o
r
m
alities
.
E
C
G
b
ea
t
s
eg
m
en
tatio
n
is
d
o
n
e
as
d
is
cu
s
s
ed
in
s
ec
tio
n
2
.
2
.
1
.
Fig
u
r
e
5
s
h
o
ws
E
C
G
p
ea
k
s
d
etec
ted
an
d
Fig
u
r
e
6
s
h
o
ws
s
eg
m
en
ted
E
C
G
b
ea
t.
T
h
ese
s
eg
m
en
ted
E
C
G
b
ea
ts
ar
e
s
to
r
ed
as
E
C
G
im
ag
es
an
d
co
n
s
id
er
ed
f
o
r
f
u
r
th
er
p
r
o
c
ess
in
g
.
A
to
tal
o
f
4
5
,
9
4
3
E
C
G
b
ea
ts
ar
e
ex
tr
ac
ted
o
f
wh
ich
3
6
,
7
5
3
b
e
ats
ar
e
u
s
ed
f
o
r
tr
ain
in
g
an
d
9
,
1
9
0
b
ea
ts
ar
e
u
s
ed
f
o
r
test
in
g
.
T
ab
le
1
s
u
m
m
ar
i
ze
s
th
e
to
tal
b
ea
ts
co
n
s
id
er
ed
af
ter
s
eg
m
en
tatio
n
.
Featu
r
e
ex
tr
ac
tio
n
is
d
o
n
e
u
s
in
g
T
r
an
s
f
e
r
le
ar
n
i
n
g
b
ased
R
esNet1
8
m
o
d
el
as
d
is
cu
s
s
ed
in
s
ec
tio
n
2
.
2
an
d
class
if
ied
u
s
in
g
o
p
tim
ized
-
SVM
as d
is
cu
s
s
ed
in
s
ec
tio
n
2
.
3
.
T
h
e
c
o
n
f
u
s
io
n
ch
a
r
t
d
e
p
icts
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
m
o
d
e
l.
I
t
d
ep
icts
h
o
w
th
e
p
a
r
ticu
lar
tr
ain
e
d
m
o
d
el
is
p
er
f
o
r
m
in
g
o
n
a
g
iv
e
n
test
d
at
aset
f
o
r
wh
ich
th
e
ac
tu
al
v
alu
es
ar
e
alr
ea
d
y
k
n
o
wn
.
Pre
d
icted
c
lass
o
f
E
C
G
b
ea
t
is
co
n
s
id
er
ed
o
n
X
-
ax
is
an
d
t
r
u
e
c
lass
o
f
E
C
G
b
ea
t
is
co
n
s
id
er
ed
o
n
Y
-
ax
is
.
Fig
u
r
e
7
s
h
o
ws
th
e
co
n
f
u
s
io
n
ch
ar
t
o
b
tain
ed
u
s
in
g
u
s
in
g
R
esn
et1
8
+SVM
an
d
Fig
u
r
e
8
s
h
o
ws
t
h
e
co
n
f
u
s
io
n
c
h
ar
t
o
b
tain
ed
u
s
in
g
u
s
in
g
R
esn
et1
8
+o
p
tim
ized
SV
M.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
1
,
Octo
b
er
2
0
2
1
:
217
-
2
2
5
222
Fig
u
r
e
5
.
E
E
G
p
ea
k
d
etec
io
n
Fig
u
r
e
6
.
Seg
m
e
n
ted
E
E
G
b
ea
r
T
ab
le
1
.
Su
m
m
a
r
y
o
f
E
C
G
b
e
ats
N
L
R
V
P
A
f
F
!
j
X
a
E
J
e
Q
To
t
a
l
B
e
a
t
s
Tr
a
i
n
i
n
g
B
e
a
t
s
8
6
7
0
6
4
6
0
5
8
0
7
5
7
0
4
5
6
2
2
2
0
3
7
7
8
6
6
4
2
3
7
8
1
8
3
1
5
4
1
2
0
85
66
13
26
3
6
7
5
3
Te
st
i
n
g
B
e
a
t
s
2
1
6
8
1
6
1
5
1
4
5
2
1
4
2
6
1
4
0
6
5
0
9
1
9
6
1
6
1
94
46
39
30
21
17
3
7
9
1
9
0
To
t
a
l
B
e
a
t
s
1
0
3
8
3
8
0
7
5
7
2
5
9
7
1
3
0
7
0
2
8
2
5
4
6
9
8
2
8
0
3
4
7
2
2
2
9
1
9
3
1
5
0
1
0
6
83
16
33
4
5
9
4
3
Fig
u
r
e
7
.
C
o
n
f
u
s
io
n
ch
a
r
t u
s
i
n
g
R
esn
et1
8
+SVM
Fig
u
r
e
8
.
C
o
n
f
u
s
io
n
ch
a
r
t
u
s
in
g
R
esn
et1
8
+o
p
tim
ized
SVM
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Dete
ctio
n
o
f c
a
r
d
ia
c
a
r
r
h
yth
m
ia
u
s
in
g
d
ee
p
C
N
N
a
n
d
o
p
timiz
ed
S
V
M
(
Mo
h
eb
b
a
n
a
a
z
)
223
T
o
ch
ec
k
th
e
ef
f
icien
cy
o
f
d
ev
elo
p
ed
class
if
ier
s
,
we
ev
alu
ate
th
e
p
r
ec
is
io
n
,
s
p
ec
if
icity
,
r
ec
all
,
F
-
s
co
r
e,
m
ath
ews c
o
r
r
elatio
n
co
ef
f
icien
t
(
MCC
)
an
d
ac
cu
r
ac
y
.
Fo
r
m
u
lae
ar
e
lis
ted
b
elo
w
(
)
=
+
(
1
2
)
(
)
=
+
(
1
3
)
(
)
=
+
(
1
4
)
−
(
)
=
+
0
.
5
(
+
)
(
1
5
)
=
(
∗
)
−
(
∗
)
√
(
+
)
(
+
)
(
+
)
(
+
)
(
1
6
)
=
+
+
+
+
(
1
7
)
W
h
er
e
g
iv
es
T
r
u
e
p
o
s
itiv
es,
g
iv
es
T
r
u
e
n
eg
ativ
es,
g
iv
es
Fals
e
p
o
s
itiv
es
an
d
g
iv
es
Fals
e
n
eg
ativ
es.
Per
f
o
r
m
an
ce
Me
asu
r
es
o
f
ea
ch
E
C
G
class
ar
e
s
u
m
m
ar
ized
in
T
a
b
le
2
.
Av
er
ag
e
p
r
ec
is
io
n
,
R
ec
all,
s
p
ec
if
icity
,
F
s
co
r
e
an
d
MCC
ac
h
iev
ed
b
y
SVM
is
9
8
.
4
7
%,
9
4
.
2
5
%,
9
9
.
8
7
%,
9
8
.
8
0
%,
9
5
.
9
5
%
.
Av
er
ag
e
p
r
ec
is
io
n
,
R
ec
all,
s
p
ec
if
icity
,
F
s
co
r
e
an
d
MCC
ac
h
iev
ed
b
y
Op
tim
ized
SVM
is
9
9
.
5
1
%,
9
6
.
4
2
%,
9
9
.
9
2
%,
9
7
.
7
8
%,
9
7
.
8
0
%
.
Ov
er
all
ac
cu
r
ac
y
ac
h
iev
ed
b
y
SVM
is
9
8
.
0
6
%
an
d
ac
c
u
r
ac
y
ac
h
ie
v
ed
b
y
Op
tim
ized
-
SVM
is
9
8
.
7
%.
T
ab
le
2
.
Per
f
o
r
m
an
ce
m
ea
s
u
r
e
s
u
s
in
g
R
esn
et1
8
f
ea
tu
r
es+
SVM
La
b
e
l
s
/
N
o
t
a
t
i
o
n
R
e
s
n
e
t
1
8
+
S
V
M
La
b
e
l
s
/
N
o
t
a
t
i
o
n
R
e
s
n
e
t
1
8
+
o
p
t
i
mi
z
e
d
-
S
V
M
Pr
(
%)
Re
(
%)
S
p
(
%)
Fs
(
%)
M
C
C
(
%)
Pr
(
%)
Re
(
%)
S
p
(
%)
Fs
(
%)
M
C
C
(
%)
N
(
‘
1
’
)
9
8
.
5
4
9
9
.
9
0
9
9
.
9
7
9
9
.
2
2
9
8
.
9
8
N
(
‘
1
’
)
9
9
.
5
4
9
9
.
7
6
9
9
.
9
2
9
9
.
6
5
9
9
.
5
4
L
(
‘
2
’
)
9
7
.
6
9
9
9
.
7
5
9
9
.
9
4
9
8
.
7
1
9
8
.
4
4
L
(
‘
2
’
)
9
8
.
8
9
9
9
.
2
5
9
9
.
8
4
9
9
.
0
7
9
8
.
8
7
R
(
‘
3
’
)
9
9
.
5
6
9
4
.
9
7
9
9
.
0
6
9
7
.
1
2
9
6
.
7
4
R
(
‘
3
’
)
9
9
.
1
1
9
9
.
6
5
9
9
.
9
3
9
9
.
3
8
9
9
.
2
6
V
(
‘
4
’
)
9
8
.
1
5
9
7
.
1
2
9
9
.
6
1
9
7
.
6
3
9
7
.
3
2
V
(
‘
4
’
)
9
9
.
3
6
9
8
.
9
4
9
9
.
8
5
9
9
.
1
5
9
9
.
0
4
P
(
‘
5
’
)
9
9
.
7
8
9
8
.
5
0
9
9
.
8
0
9
9
.
1
4
9
9
.
0
3
P
(
‘
5
’
)
9
6
.
2
3
9
9
.
9
2
9
9
.
9
9
9
8
.
0
4
9
7
.
8
0
A
(
‘
6
’
)
8
7
.
8
3
9
6
.
4
6
9
9
.
8
1
9
1
.
9
4
9
1
.
6
1
A
(
‘
6
’
)
9
9
.
1
3
8
9
.
7
8
9
9
.
4
6
9
4
.
2
2
9
4
.
0
6
f
(
‘
7
’
)
9
9
.
4
8
9
8
.
4
6
9
9
.
9
6
9
8
.
9
7
9
8
.
9
5
f
(
‘
7
’
)
1
0
0
9
3
.
8
7
9
9
.
8
7
9
6
.
8
4
9
6
.
8
2
F
(
‘
8
’
)
9
8
.
7
7
1
0
0
1
0
0
9
9
.
3
8
9
9
.
3
7
F
(
‘
8
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
!
(
‘
9
’
)
1
0
0
9
5
.
7
4
9
9
.
9
5
9
7
.
8
2
9
7
.
8
2
!
(
‘
9
’
)
1
0
0
9
6
.
8
0
9
9
.
9
6
9
8
.
3
7
9
8
.
3
7
j
(
’
1
0
)
9
5
.
8
3
1
0
0
1
0
0
9
7
.
8
7
9
7
.
8
8
j
(
’
1
0
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
X
(
‘
1
1
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
X
(
‘
1
1
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
a
(
‘
1
2
’
)
1
0
0
70
9
9
.
9
0
8
2
.
3
5
8
3
.
6
2
a
(
‘
1
2
’
)
1
0
0
9
3
.
3
3
9
9
.
9
7
9
6
.
5
5
9
6
.
5
9
E
(
‘
1
3
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
E
(
‘
1
3
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
J (‘
1
4
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
J (‘
1
4
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
e
(
‘
1
5
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
e
(
‘
1
5
’
)
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
Q
(
‘
1
6
’
)
1
0
0
5
7
.
1
4
9
9
.
9
6
7
2
.
7
2
7
5
.
5
8
Q
(
‘
1
6
’
)
1
0
0
7
1
.
4
2
9
9
.
9
7
8
3
.
3
3
8
4
.
5
0
A
v
e
r
a
g
e
9
8
.
4
7
9
4
.
2
5
9
9
.
8
7
9
8
.
8
0
9
5
.
9
5
A
v
e
r
a
g
e
9
9
.
5
1
9
6
.
4
2
9
9
.
9
2
9
7
.
7
8
9
7
.
8
0
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
=
9
8
.
0
6
%
O
v
e
r
a
l
l
A
c
c
u
r
a
c
y
=
9
8
.
7
0
%
Sev
er
al
s
tate
o
f
a
r
t
s
tu
d
ies
atte
m
p
ted
b
y
m
an
y
r
esear
ch
er
s
is
s
h
o
wn
in
T
ab
le
3
.
Ou
r
p
r
o
p
o
s
ed
m
o
d
el,
o
u
tp
er
f
o
r
m
ed
all
State
o
f
ar
t T
ec
h
n
iq
u
es a
n
d
class
if
ied
1
6
ty
p
e
s
o
f
E
C
G
b
ea
ts
with
an
ac
cu
r
ac
y
o
f
9
8
.
7
0
%.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
1
,
Octo
b
er
2
0
2
1
:
217
-
2
2
5
224
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
s
with
s
tate
-
of
-
ar
t m
o
d
el
s
S
t
u
d
y
B
e
a
t
s
C
l
a
s
ses
M
e
t
h
o
d
P
e
r
f
o
r
ma
n
c
e
M
e
l
g
a
n
i
e
t
a
l
.
[
1
6
]
4
0
4
3
8
6
P
S
O
+
S
V
M
A
c
c
8
9
.
7
2
%
Li
e
t
a
l
.
[
1
7
]
7
2
0
5
k
-
I
C
A
+
S
V
M
A
c
c
9
7
.
7
8
%
O
so
w
sk
i
e
t
a
l
.
[
1
8
]
1
2
7
8
5
13
H
O
S
+
S
V
M
A
c
c
9
5
.
9
1
%
V
e
n
k
a
t
e
sa
n
e
t
al
.
[
1
9
]
-
2
H
R
V
+
S
V
M
A
c
c
9
6
.
0
0
%
Y
a
n
g
e
t
al
.
[
2
0
]
1
0
4
9
8
6
15
A
K
M
C
+
K
N
N
A
c
c
9
7
.
7
0
%
R
o
d
r
i
g
u
e
z
e
t
a
l
.
[
2
6
]
6
4
2
6
0
16
M
o
r
p
h
o
l
o
g
y
+
D
e
c
i
s
i
o
n
A
c
c
9
6
.
1
3
%
O
h
e
t
al
.
[
2
7
]
1
6
4
9
9
5
C
N
N
+
LST
M
A
c
c
9
8
.
1
0
%
M
o
u
s
a
v
i
e
t
a
l
.
[
2
8
]
7
5
0
5
C
N
N
-
LSTM
A
c
c
9
2
.
5
0
%
P
r
o
p
o
se
d
M
o
d
e
l
4
5
9
4
3
16
R
e
sN
e
t
1
8
+
S
V
M
R
e
sN
e
t
1
8
+
O
p
t
i
m
i
z
e
d
S
V
M
A
c
c
9
8
.
0
6
%
A
c
c
9
8
.
7
0
%
4.
CO
NCLU
SI
O
N
Dee
p
l
ea
r
n
in
g
-
b
ased
C
NN
m
o
d
els
h
av
e
b
ee
n
p
r
o
p
o
s
ed
to
cl
ass
if
y
1
6
t
y
p
es
o
f
E
C
G
b
ea
ts
.
MI
T
-
B
I
H
ar
r
h
y
th
m
ia
d
atab
ase
h
as
b
ee
n
co
n
s
id
er
ed
in
t
h
is
s
tu
d
y
.
T
o
r
em
o
v
e
th
e
n
o
is
e
f
r
o
m
s
ig
n
al
s
S
-
Go
lay
f
ilter
in
g
h
as
b
ee
n
p
er
f
o
r
m
e
d
.
C
o
n
s
id
er
in
g
n
o
is
e
-
f
r
ee
s
ig
n
al,
a
to
tal
o
f
4
5
,
9
4
3
E
C
G
b
ea
ts
ar
e
ex
tr
ac
ted
u
s
in
g
E
C
G
b
ea
t
s
eg
m
en
tatio
n
tech
n
iq
u
e
.
Featu
r
e
ex
tr
ac
tio
n
is
d
o
n
e
u
s
in
g
d
ee
p
lear
n
in
g
m
o
d
el
an
d
c
lass
if
ied
u
s
in
g
th
e
p
r
o
p
o
s
ed
o
p
tim
ized
-
SVM.
M
u
lticlas
s
SVM
with
er
r
o
r
-
c
o
r
r
ec
tin
g
o
u
tp
u
t
co
d
e
was
u
s
ed
an
d
th
e
n
o
p
tim
ized
u
s
in
g
SGD
o
p
tim
izer
.
Op
tim
i
za
tio
n
im
p
r
o
v
ed
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
o
u
r
m
o
d
el.
O
u
r
m
o
d
el
r
ec
o
r
d
e
d
th
e
b
est
ac
cu
r
ac
y
o
f
9
8
.
7
0
% c
o
m
p
ar
e
d
to
all
th
e
r
elate
d
wo
r
k
s
.
RE
F
E
R
E
NC
E
S
[
1
]
A.
S
G
o
,
D.
M
o
z
a
ffa
rian
,
E.
J
Be
n
jam
in
,
D.
K
Arn
e
tt
,
a
n
d
M
.
J.
Blah
a
,
“
Am
e
rica
n
He
a
rt
As
so
c
iat
io
n
C
o
u
n
c
il
o
n
Ep
id
e
m
io
lo
g
y
a
n
d
P
re
v
e
n
ti
o
n
S
ta
ti
stics
Co
m
m
it
tee
a
n
d
S
tro
k
e
S
tat
isti
c
s
S
u
b
c
o
m
m
it
tee
,
”
He
a
rt
Dise
a
se
a
n
d
S
tr
o
k
e
S
ta
ti
st
ics
-
2
0
1
8
Up
d
a
te:
a
re
p
o
rt
f
ro
m
th
e
Ame
ric
a
n
He
a
rt
Asso
c
ia
t
io
n
,
Circ
u
l
a
ti
o
n
,
v
o
l.
1
3
7
,
n
o
.
1
2
,
p
p
.
e
6
7
-
e
4
9
2
,
2
0
1
8
.
[
2
]
R.
J.
M
a
rt
is,
U.
R.
Ac
h
a
ry
a
,
a
n
d
L.
C.
M
i
n
,
‘‘E
CG
b
e
a
t
c
las
sifica
ti
o
n
u
si
n
g
P
CA,
LDA,
ICA
a
n
d
d
isc
re
te
wa
v
e
let
tran
sfo
rm
,
”
Bi
o
me
d
.
S
i
g
n
a
l
Pro
c
e
ss
.
Co
n
tro
l
,
v
o
l.
8
,
n
o
.
5
,
p
p
.
4
3
7
-
4
4
8
,
S
e
p
.
2
0
1
3
,
d
o
i:
1
0
.
1
0
1
6
/
j.
b
s
p
c
.
2
0
1
3
.
0
1
.
0
0
5
.
[
3
]
M
o
h
e
b
b
a
n
a
a
z
,
Y.
P
a
d
m
a
S
a
i
a
n
d
L.
Ra
jan
i
Ku
m
a
ri
,
"
A
re
v
iew
o
n
a
rrh
y
t
h
m
ia
c
las
sifica
ti
o
n
u
si
n
g
ECG
sig
n
a
ls
,
"
2
0
2
0
IEE
E
I
n
ter
n
a
ti
o
n
a
l
S
tu
d
e
n
t
s'
Co
n
fer
e
n
c
e
o
n
El
e
c
trica
l
El
e
c
tro
n
ics
a
n
d
Co
mp
u
ter
S
c
ien
c
e
(S
CEE
CS
)
,
2
0
2
0
,
p
p
.
1
-
6
,
d
o
i:
1
0
.
1
1
0
9
/S
CE
ECS
4
8
3
9
4
.
2
0
2
0
.
9
.
[
4
]
G
.
B.
M
o
o
d
y
a
n
d
R
.
G
.
M
a
rk
,
"
Th
e
im
p
a
c
t
o
f
t
h
e
M
IT
-
BIH
Arrh
y
t
h
m
ia
Da
tab
a
se
,
"
in
IE
EE
En
g
in
e
e
rin
g
in
M
e
d
icin
e
a
n
d
Bi
o
lo
g
y
M
a
g
a
zin
e
,
v
o
l.
2
0
,
n
o
.
3
,
p
p
.
4
5
-
5
0
,
M
a
y
-
Ju
n
e
2
0
0
1
,
d
o
i:
1
0
.
1
1
0
9
/5
1
.
9
3
2
7
2
4
.
L.
V.
[
5
]
L.
V.
Ra
jan
i
K
u
m
a
ri,
Y.
P
a
d
m
a
S
a
i,
a
n
d
N.
Ba
laji,
“
R
-
P
e
a
k
I
d
e
n
ti
fica
ti
o
n
in
ECG
S
ig
n
a
ls
u
sin
g
P
a
tt
e
rn
-
Ad
a
p
ted
Wav
e
let
Tec
h
n
i
q
u
e
,
”
IE
T
E
J
o
u
rn
a
l
o
f
Res
e
a
rc
h
,
2
0
2
1
,
d
o
i:
1
0
.
1
0
8
0
/0
3
7
7
2
0
6
3
.
2
0
2
1
.
1
8
9
3
2
2
9
.
[
6
]
P
.
S
ri
Lak
sh
m
i
a
n
d
V.
L
o
k
e
sh
Ra
ju
,
“
ECG
De
-
n
o
isi
n
g
u
si
n
g
Hy
b
ri
d
Li
n
e
a
riza
ti
o
n
M
e
th
o
d
,
”
I
n
d
o
n
e
s
ia
n
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
Co
m
p
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l.
1
5
,
n
o
.
3
,
p
p
.
5
0
4
-
5
0
8
,
2
0
1
5
.
[
7
]
P
.
S
a
rt
h
a
k
,
a
n
d
N.
M
.
M
ih
ir,
“
I
m
p
u
lsiv
e
No
ise
Ca
n
c
e
ll
a
ti
o
n
fro
m
ECG
S
ig
n
a
l
u
si
n
g
Ad
a
p
ti
v
e
F
il
ters
a
n
d
t
h
e
ir
Co
m
p
a
riso
n
,
”
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
i
n
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
(IJ
EE
C
S
)
,
v
o
l
.
3
,
n
o
.
2
,
p
p
.
369
-
3
7
6
,
2
0
1
6
,
d
o
i:
1
0
.
1
1
5
9
1
/
ij
e
e
c
s.v
3
.
i
2
.
p
p
3
6
9
-
3
7
6
.
[
8
]
H.
Ra
h
m
a
d
,
L.
Nin
i
k
S
ri
a
n
d
He
r
a
wa
ti
,
“
An
a
p
p
ro
a
c
h
o
f
a
d
a
p
ti
v
e
n
o
tch
fi
lt
e
rin
g
d
e
sig
n
fo
r
e
lec
tr
o
c
a
rd
io
g
ra
m
n
o
ise
c
a
n
c
e
ll
a
ti
o
n
,
”
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
E
n
g
i
n
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l
.
2
2
,
n
o
.
3
,
p
p
.
1
3
0
3
-
1
3
1
1
,
2
0
2
1
,
d
o
i:
1
0
.
1
1
5
9
1
/
i
jee
c
s.v
2
2
.
i
3
.
p
p
1
3
0
3
-
1
3
1
1
.
[
9
]
S
.
Ra
j
,
K.
M
a
u
ry
a
,
a
n
d
K
.
C.
Ra
y
,
“
A
k
n
o
wle
d
g
e
-
b
a
se
d
re
a
l
ti
m
e
e
m
b
e
d
d
e
d
p
latfo
rm
f
o
r
a
rrh
y
t
h
m
ia
b
e
a
t
c
las
sifica
t
io
n
.
Bio
m
e
d
,
”
En
g
.
L
e
tt
.
v
o
l
.
5
,
n
o
.
4
,
p
p
.
2
7
1
-
2
8
0
,
2
0
1
5
,
d
o
i:
1
0
.
1
0
0
7
/s1
3
5
3
4
-
0
1
5
-
0
1
9
6
-
9
.
[
1
0
]
Q.
Li
,
C.
Ra
ja
g
o
p
a
lan
a
n
d
G
.
D.
Cli
ffo
rd
,
"
Ve
n
tri
c
u
lar
F
ib
ril
lati
o
n
a
n
d
Tac
h
y
c
a
rd
ia
Clas
sifica
ti
o
n
U
sin
g
a
M
a
c
h
i
n
e
Lea
rn
in
g
Ap
p
ro
a
c
h
,
"
in
IE
EE
T
r
a
n
sa
c
ti
o
n
s
o
n
Bi
o
me
d
ica
l
E
n
g
i
n
e
e
rin
g
,
v
o
l.
6
1
,
n
o
.
6
,
p
p
.
1
6
0
7
-
1
6
1
3
,
Ju
n
e
2
0
1
4
,
d
o
i:
1
0
.
1
1
0
9
/T
BM
E.
2
0
1
3
.
2
2
7
5
0
0
0
.
[
1
1
]
L.
V.
Ra
jan
i
Ku
m
a
ri,
Y.
P
a
d
m
a
S
a
i
a
n
d
N.
Ba
laji,
“
ECG
S
i
g
n
a
l
P
re
p
r
o
c
e
ss
in
g
Ba
se
d
o
n
E
m
p
iri
c
a
l
M
o
d
e
De
c
o
m
p
o
siti
o
n
,
”
M
icr
o
e
lec
tro
n
ic
s,
El
e
c
tro
ma
g
n
e
ti
c
s
a
n
d
T
e
lec
o
m
mu
n
ica
t
io
n
s.
S
p
rin
g
e
r,
Ne
w
De
lh
i
,
p
p
.
6
7
3
-
6
7
9
,
2
0
1
6
.
[
1
2
]
L.
D
.
S
h
a
rm
a
a
n
d
R
.
K.
S
u
n
k
a
ria
,
“
In
fe
rio
r
m
y
o
c
a
rd
ial
in
fa
rc
ti
o
n
d
e
tec
ti
o
n
u
sin
g
sta
ti
o
n
a
ry
wa
v
e
let
tran
sfo
r
m
a
n
d
m
a
c
h
in
e
lea
rn
in
g
a
p
p
r
o
a
c
h
,
”
S
ig
n
a
l
Ima
g
e
Vi
d
e
o
Pro
c
e
ss
,
v
o
l
.
1
2
,
n
o
.
2
,
p
p
.
1
9
9
-
2
0
6
,
2
0
1
7
,
d
o
i:
1
0
.
1
0
0
7
/s1
1
7
6
0
-
017
-
1
1
4
6
-
z
.
[
1
3
]
M
.
M
o
h
a
n
t
y
,
S
.
S
a
h
o
o
,
P
.
Bisw
a
l
a
n
d
S
.
S
a
b
u
t,
“
Eff
ic
ien
t
c
las
sifica
ti
o
n
o
f
v
e
n
tri
c
u
lar
a
rrh
y
th
m
i
a
s
u
sin
g
fe
a
tu
re
se
lec
ti
o
n
a
n
d
C4
.
5
c
las
sifier,
”
Bi
o
me
d
ica
l
S
i
g
n
a
l
Pro
c
e
ss
in
g
a
n
d
C
o
n
tr
o
l
,
v
o
l.
4
4
,
p
p
.
2
0
0
-
2
0
8
,
2
0
1
8
,
d
o
i:
1
0
.
1
0
1
6
/
j.
b
s
p
c
.
2
0
1
8
.
0
4
.
0
0
5
.
[
1
4
]
M
.
S
h
a
rm
a
,
R
.
S
.
Tan
,
a
n
d
U.
R
.
Ac
h
a
ry
a
,
“
Au
to
m
a
ted
h
e
a
rtb
e
a
t
c
las
sifica
ti
o
n
a
n
d
d
e
tec
ti
o
n
o
f
a
rr
h
y
t
h
m
ia
u
si
n
g
o
p
ti
m
a
l
o
rth
o
g
o
n
a
l
wa
v
e
let
fil
ters
,
”
In
fo
rm
a
ti
c
s
i
n
M
e
d
icin
e
Un
lo
c
k
e
d
,
v
o
l
.
1
6
,
p
.
1
0
0
2
2
1
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j.
imu
.
2
0
1
9
.
1
0
0
2
2
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Dete
ctio
n
o
f c
a
r
d
ia
c
a
r
r
h
yth
m
ia
u
s
in
g
d
ee
p
C
N
N
a
n
d
o
p
timiz
ed
S
V
M
(
Mo
h
eb
b
a
n
a
a
z
)
225
[
1
5
]
S.
C.
M
a
tt
a
,
Z.
S
a
n
k
a
ri
,
a
n
d
S
.
R
ih
a
n
a
,
“
He
a
rt
ra
te
v
a
riab
il
it
y
a
n
a
l
y
sis
u
sin
g
n
e
u
ra
l
n
e
two
rk
m
o
d
e
ls
fo
r
a
u
to
m
a
ti
c
d
e
tec
ti
o
n
o
f
li
fe
st
y
le
a
c
ti
v
i
ti
e
s,
”
Bi
o
me
d
.
S
ig
n
a
l
Pr
o
c
e
ss
Co
n
tro
l
,
v
o
l.
42
,
p
p
.
1
4
5
-
1
5
7
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j.
b
s
p
c
.
2
0
1
8
.
0
1
.
0
1
6
.
[
1
6
]
L.
V.
Ra
jan
i
Ku
m
a
ri,
Y.
P
a
d
m
a
S
a
i
a
n
d
N.
Ba
laji,
“
P
e
rfo
rm
a
n
c
e
Ev
a
lu
a
t
io
n
o
f
Ne
u
ra
l
Ne
two
r
k
s
a
n
d
Ad
a
p
ti
v
e
Ne
u
ro
F
u
z
z
y
In
fe
re
n
c
e
S
y
ste
m
fo
r
Clas
sifica
ti
o
n
o
f
Ca
r
d
iac
Arrh
y
th
m
ia,”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
En
g
in
e
e
rin
g
&
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
7
,
p
p
.
2
5
0
-
2
5
3
,
2
0
1
8
.
[
1
7
]
F
.
M
e
l
g
a
n
i
a
n
d
Y.
Ba
z
i,
"
Clas
sif
ica
ti
o
n
o
f
El
e
c
tro
c
a
rd
io
g
ra
m
S
ig
n
a
ls
Wi
t
h
S
u
p
p
o
rt
Ve
c
to
r
M
a
c
h
i
n
e
s
a
n
d
P
a
rti
c
le
S
wa
rm
Op
ti
m
iza
ti
o
n
,
"
in
IEE
E
Tran
sa
c
ti
o
n
s
o
n
In
f
o
rm
a
ti
o
n
Tec
h
n
o
lo
g
y
in
Bi
o
m
e
d
icin
e
,
v
o
l.
1
2
,
n
o
.
5
,
p
p
.
6
6
7
-
6
7
7
,
S
e
p
t.
2
0
0
8
,
d
o
i:
1
0
.
1
1
0
9
/T
I
TB.
2
0
0
8
.
9
2
3
1
4
7
.
[
1
8
]
H.
Li
,
X.
F
e
n
g
,
L
.
Ca
o
,
H.
Li
a
n
g
,
a
n
d
C
.
M
iao
.,
“
No
v
e
l
ECG
sig
n
a
l
c
las
sifica
ti
o
n
b
a
se
d
o
n
KIC
A
n
o
n
li
n
e
a
r
fe
a
tu
re
e
x
trac
ti
o
n
,
”
Circ
u
it
s,
S
y
st.,
S
ig
n
a
l
Pro
c
e
ss
,
v
o
l
.
3
5
,
n
o
.
4
,
p
p
.
1
1
8
7
-
1
1
9
7
,
2
0
1
6
,
d
o
i:
1
0
.
1
0
0
7
/s0
0
0
3
4
-
015
-
0
1
0
8
-
3
.
[
1
9
]
S
.
Os
o
ws
k
i
,
L.
T.
Ho
a
i
,
a
n
d
T.
M
a
rk
iew
icz
“
S
u
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
b
a
se
d
e
x
p
e
rt
sy
ste
m
f
o
r
r
e
li
a
b
le
h
e
a
rt
b
e
a
t
re
c
o
g
n
it
i
o
n
,
”
IEE
E
T
ra
n
s.B
io
me
d
.
En
g
,
v
o
l
.
5
1
,
n
o
.
4
,
p
p
.
5
8
2
-
5
8
9
,
2
0
0
4
,
d
o
i:
1
0
.
1
1
0
9
/T
BM
E
.
2
0
0
4
.
8
2
4
1
3
8
.
[
2
0
]
C.
Ve
n
k
a
tes
a
n
,
P
.
Ka
rth
ig
a
ik
u
m
a
r,
A.
P
a
u
l,
S
.
S
a
th
e
e
sk
u
m
a
ra
n
a
n
d
R
.
K
u
m
a
r,
"
ECG
S
ig
n
a
l
P
re
p
ro
c
e
ss
in
g
a
n
d
S
VM
Clas
sifier
-
Ba
se
d
Ab
n
o
rm
a
li
ty
De
tec
ti
o
n
i
n
Re
m
o
te
He
a
lt
h
c
a
re
Ap
p
li
c
a
ti
o
n
s,"
in
I
EE
E
Ac
c
e
ss
,
v
o
l
.
6
,
p
p
.
9
7
6
7
-
9
7
7
3
,
2
0
1
8
,
d
o
i:
1
0
.
1
1
0
9
/A
CCES
S
.
2
0
1
8
.
2
7
9
4
3
4
6
.
[
2
1
]
H.
Ya
n
g
a
n
d
Z
.
Wei,
"
Arrh
y
t
h
m
ia
Re
c
o
g
n
it
i
o
n
a
n
d
Clas
sifica
ti
o
n
Us
in
g
C
o
m
b
in
e
d
P
a
ra
m
e
tri
c
a
n
d
Visu
a
l
P
a
tt
e
r
n
F
e
a
tu
re
s
o
f
ECG
M
o
r
p
h
o
lo
g
y
,
"
i
n
IEE
E
Acc
e
ss
,
v
o
l.
8
,
p
p
.
4
7
1
0
3
-
4
7
1
1
7
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/ACCE
S
S
.
2
0
2
0
.
2
9
7
9
2
5
6
.
[
2
2
]
L.
B.
M
a
rin
h
o
,
N.
D.
M
.
M
.
Na
sc
ime
n
to
,
J.
W.
M
.
S
o
u
z
a
,
M
.
V.
G
u
rg
e
l,
P
.
P
.
R.
F
il
h
o
a
n
d
V.
H.
C.
d
e
Alb
u
q
u
e
rq
u
e
,
"
A n
o
v
e
l
e
lec
tro
c
a
r
d
io
g
ra
m
fe
a
tu
re
e
x
trac
ti
o
n
a
p
p
ro
a
c
h
fo
r
c
a
rd
iac
a
rr
h
y
t
h
m
ia
c
las
sifica
ti
o
n
,
"
Fu
t
u
re
Ge
n
e
r.
Co
mp
u
t.
S
y
st
,
v
o
l.
9
7
,
p
p
.
5
6
4
-
5
7
7
,
A
u
g
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/j
.
fu
t
u
re
.
2
0
1
9
.
0
3
.
0
2
5
.
[
2
3
]
M
.
S
h
a
rm
a
,
R
.
S
.
Ta
n
,
a
n
d
U.
R
a
jen
d
ra
Ac
h
a
ry
a
,
”
Au
t
o
m
a
ted
h
e
a
rtb
e
a
t
c
las
sifica
ti
o
n
a
n
d
d
e
tec
ti
o
n
o
f
a
rr
h
y
t
h
m
ia
u
sin
g
o
p
ti
m
a
l
o
rt
h
o
g
o
n
a
l
wa
v
e
l
e
t
fil
ters
,
”
In
fo
rm
a
t
ics
in
M
e
d
icin
e
Un
l
o
c
k
e
d
,
v
o
l
.
1
6
,
p
.
1
0
0
2
2
1
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j.
imu
.
2
0
1
9
.
1
0
0
2
2
1
.
[
2
4
]
M
o
h
e
b
b
a
n
a
a
z
,
L
.
V
.
Ra
jan
i
K
u
m
a
ri,
a
n
d
Y.
P
a
d
m
a
S
a
i,
“
Clas
sifica
ti
o
n
o
f
Arrh
y
t
h
m
ia
Be
a
ts
Us
in
g
Op
ti
m
ize
d
K
-
Ne
a
re
st
N
e
ig
h
b
o
r
Clas
sifier,”
In
:
Ud
g
a
t
a
S
.
K
.
,
S
e
t
h
i
S
.
,
S
rira
ma
S
.
N.
(e
d
s)
In
tell
ig
e
n
t
S
y
ste
ms
.
L
e
c
tu
r
e
No
tes
in
Ne
two
rk
s a
n
d
S
y
ste
ms
,
S
p
r
in
g
e
r,
S
in
g
a
p
o
re
,
v
ol
.
1
8
5
,
2
0
2
1
,
d
o
i:
1
0
.
1
0
0
7
/
9
7
8
-
9
8
1
-
33
-
6
0
8
1
-
5
_
3
1
.
[
2
5
]
R.
Va
ra
th
a
ra
jan
,
G
.
M
a
n
o
g
a
ra
n
,
a
n
d
M.
K.
P
riy
a
n
,
“
A
b
ig
d
a
t
a
c
las
sific
a
ti
o
n
a
p
p
r
o
a
c
h
u
sin
g
LDA
with
a
n
e
n
h
a
n
c
e
d
S
VM
m
e
th
o
d
f
o
r
ECG
sig
n
a
ls
in
c
lo
u
d
c
o
m
p
u
t
in
g
,
”
M
u
lt
ime
d
i
a
T
o
o
ls
Ap
p
l
,
v
o
l
.
7
7
,
p
p
.
1
0
1
9
5
-
1
0
2
1
5
,
2
0
1
8
,
d
o
i:
1
0
.
1
0
0
7
/s
1
1
0
4
2
-
0
1
7
-
5
3
1
8
-
1
.
[
2
6
]
S.
Sa
b
a
r
,
S
.
Dja
n
g
g
a
n
,
P
.
S
.
S
e
t
y
a
wa
n
,
a
n
d
S
.
An
d
a
rin
i,
“
T
h
e
P
e
a
k
o
f
t
h
e
P
QRST
a
n
d
th
e
Traje
c
to
ry
P
a
th
o
f
Eac
h
Cy
c
le
o
f
t
h
e
ECG
1
2
-
Lea
d
Wav
e
,
”
In
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
E
lec
trica
l
E
n
g
i
n
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
i
e
n
c
e
(IJ
EE
CS
)
,
v
o
l.
4
,
n
o
.
1
,
p
p
.
1
6
9
-
1
7
5
,
2
0
1
6
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
4
.
i1
.
p
p
1
6
9
-
1
7
5
.
[
2
7
]
J.
Ro
d
ri
g
u
e
z
,
A.
G
o
n
i
,
a
n
d
A.
Ill
a
rra
m
e
n
d
i,
“
Re
a
l
-
ti
m
e
c
las
sifi
c
a
ti
o
n
o
f
ECG
s
o
n
a
P
DA
,
”
IE
EE
T
ra
n
s.
In
f
.
T
e
c
h
n
o
l
.
Bi
o
me
d
,
v
o
l
.
9
,
n
o
.
1
,
p
p
.
23
-
34
,
2
0
0
5
,
d
o
i:
1
0
.
1
1
0
9
/
TIT
B.
2
0
0
4
.
8
3
8
3
6
9
.
[
2
8
]
S
.
L.
Oh
,
E
.
Y.
K.
Ng
,
R.
S
.
Ta
n
,
a
n
d
U.
R.
Ac
h
a
ry
a
,
‘‘Au
t
o
m
a
ted
d
iag
n
o
sis
o
f
a
rr
h
y
t
h
m
ia
u
sin
g
c
o
m
b
in
a
ti
o
n
o
f
CNN
a
n
d
LS
T
M
tec
h
n
iq
u
e
s
wit
h
v
a
riab
le
le
n
g
t
h
h
e
a
rt
b
e
a
ts,’
’
Co
mp
u
t
.
Bi
o
l
.
M
e
d
,
v
o
l.
1
0
2
,
p
p
.
2
7
8
-
2
8
7
,
No
v
.
2
0
1
8
,
d
o
i:
1
0
.
1
0
1
6
/
j.
c
o
m
p
b
io
m
e
d
.
2
0
1
8
.
0
6
.
0
0
2
.
[
2
9
]
S
.
M
o
u
sa
v
i,
A.
F
o
to
o
h
i
n
a
sa
b
,
a
n
d
F
.
Afg
h
a
h
,
“
S
i
n
g
le
-
m
o
d
a
l
a
n
d
m
u
lt
i
-
m
o
d
a
l
fa
lse
a
rrh
y
th
m
ia
a
larm
re
d
u
c
ti
o
n
u
sin
g
a
tt
e
n
ti
o
n
-
b
a
se
d
c
o
n
v
o
lu
ti
o
n
a
l
a
n
d
re
c
u
rre
n
t
n
e
u
ra
l
n
e
two
rk
s,
”
PL
o
S
On
e
,
v
o
l
.
1
5
,
n
o
.
1
,
p
.
e
0
2
2
6
9
9
0
,
2
0
2
0
,
d
o
i:
1
0
.
1
3
7
1
/j
o
u
r
n
a
l.
p
o
n
e
.
0
2
2
6
9
9
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.