I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
rm
a
t
ics a
nd
Co
m
m
un
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
,
p
p
.
1
2
9
0
~
1
3
0
2
I
SS
N:
2252
-
8
7
7
6
,
DOI
:
1
0
.
1
1
5
9
1
/iji
ct
.
v
1
5
i
3
.
pp
1
2
9
0
-
1
3
0
2
1290
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ict.
ia
esco
r
e.
co
m
Intellig
ent
eng
ine
ering
f
ra
mewo
rk
for ma
na
g
ing
hos
pital
ca
rdia
c arrest r
es
o
urces
Cha
m
s
E
dd
i
ne
F
a
t
ho
un
1
,
M
o
ha
m
ed
Ridd
a
L
a
o
ua
r
1
,
Sa
f
a
Abid
1
,
Sea
n B
.
E
o
m
2
1
LA
M
I
S
La
b
o
r
a
t
o
r
y
,
Ec
h
a
h
i
d
C
h
e
k
h
La
a
r
b
i
Te
b
e
ss
i
U
n
i
v
e
r
si
t
y
,
Te
b
e
ssa
,
A
l
g
e
r
i
a
2
D
e
p
a
r
t
me
n
t
o
f
M
a
n
a
g
e
me
n
t
,
S
o
u
t
h
e
a
st
M
i
ss
o
u
r
i
S
t
a
t
e
U
n
i
v
e
r
s
i
t
y
,
C
a
p
e
G
i
r
a
r
d
e
a
u
,
U
n
i
t
e
d
S
t
a
t
e
s
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 9
,
2
0
2
5
R
ev
is
ed
Ap
r
2
2
,
2
0
2
6
Acc
ep
ted
Ma
y
1
7
,
2
0
2
6
In
-
h
o
sp
it
a
l
c
a
rd
iac
a
rre
st
in
in
t
e
n
siv
e
c
a
re
re
m
a
in
s
fre
q
u
e
n
t
(o
ften
c
it
e
d
in
c
id
e
n
c
e
ro
u
g
h
l
y
0
.
5
%
-
7
.
8
%
o
f
a
d
m
issio
n
s),
w
h
il
e
c
a
u
se
s
d
iffe
r
in
w
h
a
t
sta
ff
a
n
d
e
q
u
i
p
m
e
n
t
m
u
st
b
e
re
a
d
y
.
We
a
sk
wh
e
t
h
e
r
v
it
a
l
-
sig
n
tr
a
jec
to
ries
fro
m
a
sta
n
d
a
rd
EHR
c
a
n
c
las
sify
wh
ich
o
f
th
re
e
c
a
rd
iac
-
re
late
d
m
e
c
h
a
n
ism
s
is
m
o
st
sa
li
e
n
t
a
rrh
y
th
m
ia,
a
c
u
te
m
y
o
c
a
rd
ial
in
fa
rc
ti
o
n
(AMI
),
o
r
r
e
sp
irato
ry
fa
il
u
re
o
r
h
y
p
o
x
ia
s
o
ICU
re
so
u
rc
e
s
c
a
n
b
e
a
li
g
n
e
d
wi
th
ri
sk
.
Us
i
n
g
M
IM
IC
-
IV,
we
e
x
trac
ted
d
iag
n
o
s
e
s
a
n
d
c
h
a
rted
v
it
a
ls
i
n
t
h
e
1
2
h
o
u
rs
b
e
fo
r
e
th
e
in
d
e
x
e
v
e
n
t,
a
p
p
li
e
d
c
lea
n
i
n
g
,
a
g
g
re
g
a
ti
o
n
,
lab
e
l
e
n
c
o
d
in
g
,
se
q
u
e
n
c
e
p
a
d
d
i
n
g
,
a
n
d
c
las
s
b
a
lan
c
in
g
(
3
,
0
0
0
c
a
se
s
p
e
r
c
las
s),
th
e
n
tr
a
in
e
d
a
n
d
c
o
m
p
a
re
d
e
Xtre
m
e
g
ra
d
ien
t
b
o
o
stin
g
(
XG
Bo
o
st
)
,
ra
n
d
o
m
f
o
r
e
st
(RF
)
,
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(
S
VM)
,
a
n
d
l
o
g
isti
c
re
g
re
ss
io
n
(LR)
w
it
h
5
-
fo
ld
c
ro
ss
-
v
a
li
d
a
ti
o
n
o
n
a
n
8
0
/
2
0
sp
li
t.
XG
Bo
o
st
p
e
rfo
rm
e
d
b
e
st
(a
b
o
u
t
9
3
%
a
c
c
u
ra
c
y
;
se
n
siti
v
it
y
8
9
.
1
5
%
;
sp
e
c
ifi
c
it
y
9
0
.
4
3
%
;
AU
C
-
ROC
0
.
9
4
).
F
e
a
tu
re
imp
o
rtan
c
e
h
ig
h
li
g
h
ted
h
e
a
rt
ra
te,
o
x
y
g
e
n
sa
tu
ra
ti
o
n
,
a
n
d
b
l
o
o
d
p
r
e
ss
u
re
p
a
tt
e
rn
s
c
o
n
siste
n
t
wit
h
b
e
d
sid
e
m
o
n
it
o
rin
g
p
ra
c
ti
c
e
.
Th
e
st
u
d
y
su
p
p
o
rts
m
e
c
h
a
n
ism
-
o
rien
ted
tri
a
g
e
lab
e
ls
d
e
riv
e
d
fro
m
wid
e
l
y
re
c
o
rd
e
d
v
it
a
ls,
a
s
a
c
o
m
p
lem
e
n
t
to
g
e
n
e
ric
e
a
rly
wa
rn
in
g
sc
o
re
s,
fo
r
p
rio
ri
ti
z
in
g
tele
m
e
try
,
re
sp
irato
ry
s
u
p
p
o
rt
,
a
n
d
c
a
rd
i
o
lo
g
y
p
a
th
wa
y
s.
Ex
tern
a
l
v
a
li
d
a
ti
o
n
a
n
d
p
ro
sp
e
c
ti
v
e
e
v
a
lu
a
ti
o
n
a
re
n
e
e
d
e
d
b
e
fo
re
d
e
p
lo
y
m
e
n
t.
K
ey
w
o
r
d
s
:
C
ar
d
iac
ar
r
est
Ho
s
p
ital r
eso
u
r
ce
m
an
a
g
em
en
t
Ma
ch
in
e
lear
n
in
g
Vital
s
ig
n
s
XGBo
o
s
t in
ten
s
iv
e
ca
r
e
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
:
C
h
am
s
E
d
d
in
e
Fath
o
u
n
L
AM
I
S L
ab
o
r
ato
r
y
,
E
ch
a
h
id
C
h
ek
h
L
aa
r
b
i T
e
b
ess
i U
n
iv
er
s
ity
T
eb
ess
a,
Alg
er
ia
E
m
ail:
ch
am
s
.
f
ath
o
u
n
@
u
n
iv
-
t
eb
ess
a.
d
z
1.
I
NT
RO
D
UCT
I
O
N
In
in
ten
s
iv
e
ca
r
e,
s
er
io
u
s
ca
r
d
iac
co
n
d
itio
n
s
n
o
tab
l
y
ar
r
h
y
th
m
ia,
ac
u
te
m
y
o
ca
r
d
ial
in
f
a
r
ctio
n
(
AM
I
)
,
an
d
r
esp
ir
ato
r
y
f
ailu
r
e
with
h
y
p
o
x
ia
o
f
ten
d
em
an
d
d
i
f
f
er
en
t
s
taf
f
s
k
ills
an
d
e
q
u
ip
m
e
n
t.
H
o
s
p
ital
in
cid
en
ce
o
f
ca
r
d
iac
ar
r
est
am
o
n
g
I
C
U
ad
m
is
s
io
n
s
is
f
r
eq
u
en
tly
r
ep
o
r
te
d
in
th
e
r
a
n
g
e
o
f
r
o
u
g
h
ly
0
.
5
%
-
7
.
8
%
[
1
]
a
n
d
th
e
ca
u
s
es
ar
e
d
iv
er
s
e,
wh
ich
m
ak
es
g
en
er
ic
aler
ts
an
awk
war
d
f
it
f
o
r
r
eso
u
r
ce
p
lan
n
in
g
[
2
]
T
h
e
p
r
ac
tical
q
u
esti
o
n
b
eh
in
d
th
is
wo
r
k
is
wh
eth
er
r
o
u
tin
ely
r
ec
o
r
d
e
d
v
ital
s
ig
n
s
,
ta
k
en
in
a
f
i
x
ed
win
d
o
w
b
e
f
o
r
e
an
in
d
e
x
ev
e
n
t,
ca
n
s
u
p
p
o
r
t
m
ec
h
an
is
m
-
o
r
ie
n
ted
t
r
iag
e
(
telem
etr
y
v
er
s
u
s
r
esp
ir
a
to
r
y
s
u
p
p
o
r
t
v
e
r
s
u
s
ca
r
d
io
lo
g
y
p
ath
way
s
)
,
r
at
h
er
th
an
o
n
l
y
a
s
in
g
le
u
n
d
if
f
e
r
en
ti
ated
r
is
k
s
co
r
e.
C
las
s
ical
ea
r
ly
war
n
in
g
s
co
r
es
(
e.
g
.
,
ME
W
S
[
3
]
,
NE
W
S
[
4
]
,
[
5
]
)
s
u
m
m
ar
ize
war
d
-
le
v
e
l
r
is
k
b
u
t
o
f
ten
lack
th
e
tem
p
o
r
al
r
eso
lu
tio
n
n
ee
d
ed
f
o
r
t
he
h
o
u
r
s
i
m
m
ed
iately
b
ef
o
r
e
co
llap
s
e
[
6
]
,
[
7
]
.
L
ar
g
e
E
HR
co
r
p
o
r
a
s
u
ch
as
MI
MI
C
-
IV
e
n
ab
le
lea
r
n
in
g
f
r
o
m
m
an
y
I
C
U
s
tay
s
with
r
ic
h
v
ital
an
d
lab
o
r
ato
r
y
d
ata
[
1
]
,
[
8
]
,
[
9
]
;
co
m
p
lem
en
tar
y
s
ig
n
al
ar
ch
iv
es
s
u
ch
as
Ph
y
s
io
Net
h
av
e
s
u
p
p
o
r
ted
r
elate
d
m
eth
o
d
o
lo
g
ical
p
r
o
g
r
ess
[
1
0
]
.
Ma
ch
in
e
lear
n
in
g
s
tu
d
ies
r
e
p
o
r
t
h
i
g
h
d
is
cr
im
in
atio
n
f
o
r
in
h
o
s
p
ital
ca
r
d
iac
e
v
en
ts
,
in
cl
u
d
in
g
f
r
o
m
c
o
m
p
ac
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
I
n
tellig
en
t e
n
g
in
ee
r
in
g
fr
a
mewo
r
k
fo
r
ma
n
a
g
in
g
h
o
s
p
ita
l c
a
r
d
ia
c
a
r
r
est r
eso
u
r
ce
s
(
C
h
a
ms E
d
d
in
e
F
a
th
o
u
n
)
1291
s
tr
u
ctu
r
ed
f
iel
d
s
[
1
1
]
-
[
1
4
]
.
S
u
b
g
r
o
u
p
d
if
f
er
e
n
ce
s
b
etwe
en
h
ea
r
t
-
f
ailu
r
e
an
d
ac
u
te
co
r
o
n
ar
y
p
r
esen
tati
o
n
s
r
em
ain
im
p
o
r
tan
t [
1
5
]
-
[
1
8
]
.
C
h
allen
g
es.
E
HR
tim
e
s
er
ies
ar
e
s
p
ar
s
e
an
d
u
n
ev
en
l
y
o
b
s
er
v
ed
[
1
2
]
,
[
1
3
]
;
ar
r
h
y
th
m
ia,
AM
I
,
an
d
h
y
p
o
x
i
a
f
o
l
l
o
w
d
i
f
f
e
r
e
n
t
p
h
y
s
io
l
o
g
y
[
1
6
]
,
[
1
8
]
;
o
u
t
c
o
m
e
s
a
r
e
i
m
b
a
l
a
n
c
e
d
a
n
d
ar
ea
u
n
d
er
th
e
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
A
UC
-
R
OC
)
a
l
o
n
e
can
m
i
s
l
e
a
d
[
1
9
]
,
[
2
0
]
;
s
tr
o
n
g
p
r
ed
icto
r
s
m
ay
s
till
n
ee
d
p
o
s
t
-
hoc
ex
p
lan
atio
n
s
f
o
r
s
af
e
ad
o
p
tio
n
[
7
]
,
[
2
1
]
;
an
d
d
ee
p
s
eq
u
en
ce
m
o
d
els
ad
d
f
lex
ib
ilit
y
at
a
co
m
p
u
tatio
n
al
p
r
ice
[
2
]
,
[
8
]
.
T
er
m
in
o
lo
g
y
.
C
ar
d
iac
a
r
r
est
is
th
e
ab
r
u
p
t
lo
s
s
of
ef
f
ec
tiv
e
ca
r
d
iac
o
u
tp
u
t
[
2
2
]
.
Ar
r
h
y
th
m
ia
d
en
o
te
s
ab
n
o
r
m
al
h
ea
r
t
r
h
y
th
m
[
2
3
]
.
Hy
p
o
x
ia
is
in
ad
e
q
u
ate
tis
s
u
e
o
x
y
g
e
n
atio
n
[
2
4
]
,
[
2
5
]
.
AM
I
is
m
y
o
ca
r
d
ial
in
ju
r
y
d
u
e
to
s
u
s
tain
ed
is
ch
em
ia
[
2
6
]
.
R
elate
d
wo
r
k
:
C
r
esp
í
et
a
l.
[
2
7
]
p
r
e
d
icted
in
h
o
s
p
ital
ca
r
d
ia
c
ar
r
est
up
to
24
h
o
u
r
s
a
h
ea
d
with
s
tr
o
n
g
d
is
cr
im
in
atio
n
; Rah
ad
ian
et
a
l
.
[
2
8
]
tar
g
ete
d
p
r
e
h
o
s
p
ital
tr
au
m
a
ar
r
est;
So
u
d
an
et
a
l.
[
2
9
]
s
h
o
wed
th
at
r
an
d
o
m
f
o
r
ests
(
R
F)
o
n
s
h
o
r
t
v
ital
-
s
ig
n
h
is
to
r
ies
r
ea
ch
u
s
ef
u
l
ac
cu
r
ac
y
;
B
lo
m
b
e
r
g
et
a
l.
[
3
0
]
ap
p
lied
ML
to
em
er
g
en
cy
ca
lls
f
o
r
o
u
t
of
h
o
s
p
ital
ar
r
est;
Mu
r
u
g
ap
p
a
n
e
t
a
l.
[
3
1
]
u
s
ed
E
C
G
m
o
r
p
h
o
lo
g
y
f
ea
tu
r
es;
an
d
Kim
et
a
l.
[
3
2
]
r
ep
o
r
ted
s
tr
o
n
g
I
C
U
r
esu
lts
with
f
ea
s
ib
le
in
p
u
ts
.
T
o
g
et
h
er
,
t
h
is
lin
e
o
f
wo
r
k
s
u
p
p
o
r
ts
v
itals
led
m
o
d
els
an
d
m
u
lti
h
o
u
r
win
d
o
ws;
we
ex
ten
d
it
b
y
m
ap
p
in
g
p
r
ed
icted
lab
els
to
r
e
s
o
u
r
ce
ty
p
es
alo
n
g
ca
r
d
iac
p
ath
way
s
.
2.
T
H
E
P
RO
P
O
SE
D
AP
P
RO
A
CH
We
ca
s
t
r
is
k
as
s
es
s
m
en
t
as
m
u
lti
-
class
cla
s
s
if
icatio
n
am
o
n
g
a
r
r
h
y
th
m
ia
-
r
elate
d
p
atter
n
s
,
AM
I
-
r
elate
d
p
atter
n
s
,
an
d
r
esp
ir
ato
r
y
f
ailu
r
e
o
r
h
y
p
o
x
ia,
u
s
in
g
g
r
ad
ien
t
-
b
o
o
s
ted
tr
ee
s
[
1
9
]
,
[
2
0
]
o
n
tem
p
o
r
ally
ag
g
r
eg
ated
v
itals
f
r
o
m
MI
MI
C
-
I
V.
E
x
t
r
e
m
e
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
(
XGBo
o
s
t
)
h
an
d
les
m
is
s
in
g
v
alu
es
an
d
n
o
n
lin
ea
r
in
ter
ac
tio
n
s
an
d
s
ca
les
to
lar
g
e
tab
u
lar
E
H
R
ex
tr
ac
ts
[
3
3
]
.
I
n
d
ep
l
o
y
m
en
t,
t
he
in
ten
d
ed
r
o
le
is
d
ec
is
io
n
s
u
p
p
o
r
t:
s
tr
atif
y
in
g
by
m
ec
h
an
is
m
s
h
o
u
ld
h
elp
r
atio
n
alize
co
n
tin
u
o
u
s
telem
etr
y
,
v
en
tilatio
n
r
eso
u
r
ce
s
,
an
d
ca
r
d
io
lo
g
y
p
r
o
ce
d
u
r
es
not
r
ep
lace
clin
ical
ju
d
g
m
en
t.
I
n
ter
p
r
etab
ilit
y
t
o
o
ls
s
u
ch
as
SHAP
[
2
1
]
ca
n
alig
n
m
o
d
el
em
p
h
asis
w
ith
b
ed
s
id
e
m
o
n
ito
r
in
g
p
r
io
r
ities
.
3.
M
E
T
H
O
D
W
e
ex
tr
ac
t
I
C
U
d
ia
g
n
o
s
es
an
d
ch
a
r
ted
v
itals
f
r
o
m
MI
MI
C
-
I
V,
p
r
ep
r
o
ce
s
s
an
d
b
alan
ce
t
h
e
co
h
o
r
t,
th
en
tr
ain
a
n
d
co
m
p
ar
e
s
ev
e
r
a
l
class
if
ier
s
with
cr
o
s
s
-
v
alid
at
io
n
.
Fig
u
r
e
1
s
u
m
m
a
r
izes
th
e
p
ip
elin
e:
d
iag
n
o
s
es
icd
an
d
ch
a
r
tev
en
ts
→
clea
n
in
g
an
d
a
g
g
r
e
g
atio
n
→
en
co
d
in
g
a
n
d
r
esam
p
lin
g
→
tr
ain
in
g
(
XGBo
o
s
t
an
d
b
aselin
es)
→
ev
alu
atio
n
.
3
.
1
.
G
ener
a
l a
rc
hite
ct
ure
Data
ar
e
d
r
awn
f
r
o
m
m
im
ic
h
o
s
p
.
d
iag
n
o
s
es
_
icd
an
d
m
i
m
ic
_
icu
.
ch
ar
tev
en
ts
.
Py
th
o
n
-
b
ased
s
tep
s
h
an
d
le
m
is
s
in
g
v
al
u
es,
t
y
p
e
co
n
v
er
s
io
n
,
ag
g
r
eg
atio
n
b
y
p
atien
t,
lab
el
en
co
d
in
g
,
s
eq
u
en
ce
p
ad
d
in
g
,
class
b
alan
cin
g
(
3
,
0
0
0
ca
s
es
p
er
c
o
n
d
itio
n
)
,
an
8
0
/2
0
s
p
lit,
an
d
5
-
f
o
l
d
cr
o
s
s
-
v
alid
atio
n
o
n
th
e
tr
ain
in
g
p
o
r
tio
n
.
Me
tr
ics
in
clu
d
e
ac
c
u
r
ac
y
,
s
en
s
itiv
ity
,
s
p
ec
if
icity
,
an
d
AUC
-
R
O
C
,
with
co
n
f
u
s
io
n
m
atr
ices
f
o
r
er
r
o
r
s
tr
u
ctu
r
e.
3
.
2
.
Da
t
a
prepro
ce
s
s
ing
3
.
2
.
1
.
Da
t
a
s
o
urce
(
M
I
M
I
C
I
V)
Me
d
ical
in
f
o
r
m
atio
n
m
ar
t
f
o
r
in
ten
s
iv
e
ca
r
e
(
MI
M
I
C
)
-
I
V,
a
lar
g
e
d
eid
e
n
tifie
d
d
ataset
o
f
p
atien
ts
ad
m
itted
to
th
e
em
er
g
en
c
y
d
ep
ar
tm
en
t
or
an
in
ten
s
iv
e
ca
r
e
u
n
it
at
th
e
B
eth
I
s
r
ae
l
Dea
co
n
ess
Me
d
ical
C
en
ter
in
B
o
s
to
n
,
MA
.
MI
MI
C
-
IV
co
n
t
ain
s
d
ata
f
o
r
o
v
er
6
5
,
0
0
0
p
atie
n
ts
ad
m
itted
to
a
n
I
C
U
an
d
o
v
er
2
0
0
,
0
0
0
p
atien
ts
ad
m
itted
to
th
e
em
er
g
e
n
cy
d
ep
ar
tm
en
t.
MI
M
I
C
-
I
V
in
co
r
p
o
r
ates
co
n
tem
p
o
r
ar
y
d
ata
an
d
ad
o
p
ts
a
m
o
d
u
lar
ap
p
r
o
ac
h
to
d
ata
o
r
g
a
n
izatio
n
,
h
ig
h
lig
h
tin
g
d
ata
p
r
o
v
e
n
an
ce
an
d
f
ac
ilit
atin
g
b
o
th
in
d
i
v
id
u
a
l
an
d
co
m
b
in
ed
u
s
e
of
d
is
p
ar
ate
d
ata
s
o
u
r
ce
s
.
MI
MI
C
-
IV
is
in
ten
d
ed
to
ca
r
r
y
on
th
e
s
u
cc
ess
of
MI
MI
C
-
I
I
I
an
d
s
u
p
p
o
r
t
a
b
r
o
ad
s
e
t
o
f
ap
p
licatio
n
s
with
in
h
ea
lth
c
ar
e
[
1
]
.
Fig
u
r
e
2
p
r
esen
ts
a
s
tr
u
ctu
r
al
b
lo
ck
d
ia
g
r
am
o
f
th
e
MI
MI
C
-
I
V
d
atab
ase
ar
ch
itectu
r
e.
T
h
e
im
ag
e
m
a
p
s
out
h
o
w
th
e
d
atab
ase
is
d
iv
id
ed
in
to
d
is
tin
ct,
in
ter
co
n
n
ec
te
d
m
o
d
u
les,
p
r
im
ar
ily
th
e
co
r
e,
h
o
s
p
(
h
o
s
p
ital)
,
an
d
icu
(
in
ten
s
iv
e
ca
r
e)
m
o
d
u
les.
B
y
v
is
u
alizin
g
th
e
r
elatio
n
al
lin
k
s
b
etwe
en
g
en
er
al
p
atien
t
ad
m
is
s
io
n
r
ec
o
r
d
s
a
n
d
h
ig
h
-
f
r
eq
u
en
cy
in
te
n
s
iv
e
ca
r
e
u
n
it
d
ata,
th
e
d
iag
r
a
m
d
em
o
n
s
tr
ates
th
e
p
r
o
v
e
n
an
ce
o
f
th
e
d
ata.
I
t
ex
p
licitly
s
h
o
ws
h
o
w
d
is
p
ar
ate
clin
ical
s
o
u
r
ce
s
ar
e
u
n
if
ied
,
allo
win
g
o
u
r
s
y
s
tem
t
o
p
u
ll c
o
m
p
r
eh
en
s
iv
e
v
ital sig
n
s
an
d
d
ia
g
n
o
s
tic
co
d
es si
m
u
ltan
eo
u
s
ly
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
1
2
9
0
-
1
3
0
2
1292
Fig
u
r
e
1.
Ar
ch
itectu
r
e
d
iag
r
a
m
of
th
e
s
tu
d
y
illu
s
tr
atin
g
th
e
co
m
p
lete
d
ata
p
i
p
e
l
i
n
e
Fig
u
r
e
2.
MI
M
I
C
IV
m
o
d
u
les
[
3
4
]
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
I
n
tellig
en
t e
n
g
in
ee
r
in
g
fr
a
mewo
r
k
fo
r
ma
n
a
g
in
g
h
o
s
p
ita
l c
a
r
d
ia
c
a
r
r
est r
eso
u
r
ce
s
(
C
h
a
ms E
d
d
in
e
F
a
th
o
u
n
)
1293
3
.
2
.
2
.
Da
t
a
ex
t
ra
c
t
io
n
T
h
e
d
ata
ex
tr
ac
tio
n
p
h
ase
o
f
o
u
r
s
tu
d
y
is
p
iv
o
tal
to
s
ettin
g
th
e
f
o
u
n
d
atio
n
f
o
r
o
u
r
an
aly
s
is
.
W
e
u
tili
ze
th
e
MI
MI
C
-
I
V
d
ataset,
a
co
m
p
r
eh
e
n
s
iv
e
clin
ical
d
atab
ase,
to
o
b
tain
ess
en
tial
d
iag
n
o
s
tic
an
d
co
n
tin
u
o
u
s
clin
ical
in
f
o
r
m
atio
n
.
Sp
ec
if
ically
,
we
f
o
cu
s
o
n
t
wo
cr
itical
tab
les:
m
im
ic
h
o
s
p
.
d
iag
n
o
s
es
_
icd
an
d
m
im
ic
_
icu
.
ch
ar
te
v
en
ts
.
First,
we
id
en
tifie
d
th
e
r
o
ws
in
th
e
d
iag
n
o
s
es
_
icd
tab
le
t
h
at
co
r
r
esp
o
n
d
to
o
u
r
tar
g
et
co
n
d
itio
n
s
.
T
o
ac
h
iev
e
th
is
,
we
f
ilter
e
d
th
e
s
p
ec
if
ic
I
C
D
co
d
es:
AM
I
(
co
d
es
s
tar
tin
g
with
4
1
0
)
,
Ar
r
h
y
th
m
ias
(
co
d
es
s
tar
tin
g
with
4
2
7
)
,
a
n
d
R
esp
ir
ato
r
y
Failu
r
e/Hy
p
o
x
ia
(
co
d
es
5
1
8
8
1
,
5
1
8
8
2
,
5
1
8
8
4
,
an
d
7
9
9
0
2
)
.
W
e
cr
ea
ted
a
n
ew
s
tr
u
ctu
r
ed
tab
le
n
am
ed
ca
s
es
d
iag
n
o
s
to
s
to
r
e
th
ese
f
ilter
ed
r
esu
lts
.
I
n
th
is
tab
le,
we
r
etain
ed
th
e
s
u
b
ject
_
id
,
h
o
s
p
it
al
ad
m
is
s
io
n
I
D
(
h
ad
m
_
id
)
,
s
e
q
u
en
ce
n
u
m
b
e
r
,
I
C
D
co
d
e,
a
n
d
g
en
er
ate
d
a
n
ew
f
ield
ca
lled
c
o
n
d
itio
n
to
ca
teg
o
r
ize
ea
ch
r
o
w.
W
e
also
r
a
n
k
e
d
th
e
d
iag
n
o
s
es
with
in
ea
ch
s
u
b
ject
to
ca
p
tu
r
e
th
e
d
ef
in
itiv
e
f
ir
s
t o
cc
u
r
r
en
ce
o
f
e
ac
h
co
n
d
itio
n
u
s
in
g
t
h
e
R
OW
NUM
B
E
R
(
)
SQL
f
u
n
ctio
n
.
Nex
t,
we
g
ath
er
e
d
th
e
r
ele
v
an
t
v
ital
s
ig
n
m
ea
s
u
r
em
en
ts
f
o
r
ea
ch
id
en
tifie
d
c
o
n
d
itio
n
b
y
j
o
in
in
g
th
e
ca
s
es
_
d
iag
n
o
s
tab
le
with
th
e
ch
ar
tev
en
ts
tab
le,
wh
ich
co
n
tain
s
h
ig
h
-
f
r
eq
u
en
cy
clin
ical
e
v
en
t
d
ata
r
ec
o
r
d
ed
d
u
r
in
g
I
C
U
s
tay
s
.
T
h
is
jo
in
was
ex
ec
u
ted
o
n
th
e
s
u
b
ject
_
id
f
ield
.
C
r
u
cially
,
to
alig
n
w
ith
o
u
r
p
r
ee
m
p
tiv
e
p
r
ed
ictiv
e
f
r
am
ewo
r
k
,
we
s
tr
ictly
ex
tr
ac
ted
c
h
ar
tev
en
ts
r
ec
o
r
d
s
th
at
f
ell
wit
h
in
th
e
d
esig
n
ated
12
-
h
o
u
r
tem
p
o
r
al
win
d
o
w
p
r
io
r
to
th
e
d
o
cu
m
e
n
ted
ev
en
t.
T
o
f
o
cu
s
o
n
th
e
m
o
s
t
p
er
tin
en
t
p
h
y
s
io
lo
g
ica
l
p
ar
am
eter
s
,
we
f
ilter
ed
th
e
c
h
ar
tev
en
ts
tab
le
f
o
r
t
h
e
s
p
ec
i
f
ic
I
T
E
MI
Ds
ass
o
ciate
d
with
ea
ch
r
is
k
f
ac
to
r
,
as
d
etailed
in
T
ab
le
1
.
T
ab
le
1.
Vital
s
ig
n
s
an
d
th
eir
I
T
E
MI
Ds
f
o
r
d
if
f
er
e
n
t
c
o
n
d
i
t
i
o
n
s
C
o
n
d
i
t
i
o
n
V
i
t
a
l
s
i
g
n
I
T
E
M
I
D
A
r
r
h
y
t
h
m
i
a
s
H
e
a
r
t
r
a
t
e
2
2
0
0
4
5
A
r
r
h
y
t
h
m
i
a
s
S
y
st
o
l
i
c
b
l
o
o
d
p
r
e
ss
u
r
e
2
2
0
0
5
0
A
r
r
h
y
t
h
m
i
a
s
D
i
a
st
o
l
i
c
b
l
o
o
d
p
r
e
s
s
u
r
e
2
2
0
0
5
1
A
M
I
T
r
o
p
o
n
i
n
-
T
2
2
7
4
2
9
A
M
I
H
e
a
r
t
r
a
t
e
2
2
0
0
4
5
A
M
I
O
x
y
g
e
n
s
a
t
u
r
a
t
i
o
n
2
2
0
2
2
7
R
e
s
p
i
r
a
t
o
r
y
f
a
i
l
u
r
e
/
H
y
p
o
x
i
a
A
r
t
e
r
i
a
l
b
l
o
o
d
g
a
s
pH
2
2
3
8
3
0
R
e
s
p
i
r
a
t
o
r
y
f
a
i
l
u
r
e
/
H
y
p
o
x
i
a
A
r
t
e
r
i
a
l
b
l
o
o
d
g
a
s
p
C
O
2
2
2
0
2
3
5
R
e
s
p
i
r
a
t
o
r
y
f
a
i
l
u
r
e
/
H
y
p
o
x
i
a
A
r
t
e
r
i
a
l
b
l
o
o
d
g
a
s
p
O
2
2
2
0
2
2
4
R
e
s
p
i
r
a
t
o
r
y
f
a
i
l
u
r
e
/
H
y
p
o
x
i
a
R
e
s
p
i
r
a
t
o
r
y
r
a
t
e
2
2
0
2
1
0
R
e
s
p
i
r
a
t
o
r
y
f
a
i
l
u
r
e
/
H
y
p
o
x
i
a
O
x
y
g
e
n
s
a
t
u
r
a
t
i
o
n
2
2
0
2
2
7
T
h
e
r
esu
ltin
g
d
ataset
is
s
to
r
ed
in
a
n
ewly
cr
ea
ted
tab
le,
r
ef
er
r
ed
to
as c
ases
v
itals
,
wh
ich
co
n
s
o
lid
ates
all
ess
en
tial
in
f
o
r
m
atio
n
r
e
q
u
ir
ed
f
o
r
d
o
w
n
s
tr
ea
m
an
aly
s
is
.
T
h
is
tab
le
co
n
tain
s
s
u
b
ject
i
d
en
tifie
r
s
,
h
o
s
p
ita
l
ad
m
is
s
io
n
I
Ds,
ch
ar
t
tim
es,
item
I
Ds,
r
ec
o
r
d
ed
v
ital
s
ig
n
v
alu
es,
an
d
th
eir
co
r
r
e
s
p
o
n
d
in
g
u
n
its
o
f
m
ea
s
u
r
em
en
t.
I
n
ad
d
itio
n
to
th
ese
co
r
e
f
ield
s
,
we
in
tr
o
d
u
ce
an
ad
d
itio
n
al
d
escr
ip
tiv
e
attr
ib
u
te,
v
ital
s
ig
n
,
wh
ich
ex
p
licitly
s
p
ec
i
f
ies
th
e
ty
p
e
o
f
p
h
y
s
io
lo
g
ical
m
ea
s
u
r
em
en
t
r
e
p
r
esen
ted
i
n
ea
ch
r
ec
o
r
d
.
T
h
is
ad
d
ed
d
escr
ip
to
r
en
h
an
ce
s
th
e
in
ter
p
r
etab
ilit
y
o
f
t
h
e
d
ataset
an
d
f
a
cilitates
ea
s
ier
g
r
o
u
p
in
g
an
d
f
i
lter
in
g
o
f
v
ital
s
ig
n
ca
teg
o
r
ies d
u
r
in
g
s
u
b
s
eq
u
e
n
t
an
aly
tical
task
s
.
3
.
3
.
Da
t
a
prepa
ra
t
i
o
n
3
.
3
.
1
.
M
is
s
i
ng
da
t
a
Ou
t
o
f
th
e
to
tal
d
ataset,
at
t
h
e
p
r
ep
a
r
ato
r
y
s
tag
e,
we
ar
e
f
a
ce
d
with
1
0
4
v
alu
es
f
o
r
wh
ic
h
d
ata
was
m
is
s
in
g
in
o
u
r
s
et,
p
ar
ticu
la
r
ly
in
th
e
v
alu
e
n
u
m
c
o
lu
m
n
.
Ho
wev
er
,
s
in
ce
th
e
n
u
m
b
er
r
ep
r
esen
ts
q
u
ite
a
s
m
aller
p
er
ce
n
tag
e
o
f
th
e
en
tire
d
ataset,
th
e
d
ec
is
io
n
t
o
r
em
o
v
e
th
ese
r
o
ws
h
ad
to
b
e
co
n
s
id
er
ed
.
T
h
is
wo
u
ld
m
ak
e
s
u
r
e
th
e
d
a
ta
is
in
teg
r
al
an
d
co
h
er
en
t
with
o
u
t
s
ev
er
ely
h
am
p
er
in
g
th
e
r
o
b
u
s
tn
ess
o
f
th
e
s
tu
d
y
,
an
d
we
will,
th
er
ef
o
r
e,
co
n
tin
u
e
th
e
s
tu
d
y
with
a
clea
n
s
u
b
s
tan
tial
d
ataset.
T
o
d
em
o
n
s
tr
ate
th
e
n
ec
e
s
s
ity
o
f
o
u
r
d
ata
clea
n
in
g
p
r
o
to
co
ls
,
Fig
u
r
e
3
p
r
o
v
id
e
s
a
d
ir
ec
t
v
is
u
al
s
n
ap
s
h
o
t
o
f
th
e
r
aw
d
ataset
in
tab
u
lar
f
o
r
m
.
T
h
e
im
a
g
e
h
i
g
h
lig
h
ts
s
p
ec
if
ic
in
s
tan
ce
s
o
f
d
ata
in
co
m
p
leten
ess
en
co
u
n
ter
ed
d
u
r
in
g
t
h
e
p
r
ep
ar
ato
r
y
s
tag
e.
I
t
d
is
p
lay
s
s
ev
er
al
co
lu
m
n
s
o
f
cli
n
ical
r
e
co
r
d
s
s
u
ch
as
s
u
b
ject
I
D,
tim
e
s
tam
p
,
an
d
item
I
D
wh
er
e
th
e
v
alu
en
u
m
(
n
u
m
er
ic
al
v
alu
e)
co
lu
m
n
c
o
n
tain
s
n
u
ll
o
r
em
p
ty
ce
lls
.
T
h
is
v
is
u
al
ev
id
en
ce
u
n
d
e
r
s
co
r
es
ex
ac
tly
h
o
w
m
is
s
in
g
p
h
y
s
io
lo
g
ical
m
ea
s
u
r
em
e
n
ts
p
r
esen
t
t
h
em
s
elv
es
in
r
aw
E
HR
ex
tr
ac
ts
b
ef
o
r
e
im
p
u
tatio
n
an
d
clea
n
in
g
ar
e
ap
p
lied
.
3
.
3
.
2
.
Da
t
a
co
nv
er
s
io
n
Du
r
in
g
th
e
d
ata
co
n
v
er
s
io
n
p
h
ase,
two
im
p
o
r
tan
t
s
tep
s
ar
e
tak
en
to
r
ep
r
esen
t
th
e
d
ata
in
an
an
aly
tically
u
s
ab
le
f
o
r
m
at.
F
ir
s
t,
C
h
ar
ttime
,
wh
ich
is
ab
o
u
t
th
e
tim
e
at
w
h
ich
e
v
er
y
clin
ical
ev
en
t
wa
s
r
ec
o
r
d
e
d
,
is
co
n
v
er
te
d
f
r
o
m
its
g
en
er
al
f
o
r
m
at
to
a
s
tan
d
ar
d
ized
o
n
e
in
d
atetim
e
f
o
r
m
at
.
T
em
p
o
r
al
an
al
y
s
is
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
1
2
9
0
-
1
3
0
2
1294
ca
n
b
e
d
o
n
e
o
n
th
is
with
h
ig
h
p
r
ec
is
io
n
,
s
in
ce
th
is
f
o
r
m
at
allo
ws
f
o
r
ea
s
y
,
tim
e
-
b
ased
o
p
er
atio
n
s
o
n
th
e
d
ataset.
Seco
n
d
is
th
e
co
n
v
e
r
s
io
n
o
f
th
e
v
alu
en
u
m
en
tr
ie
s
,
n
u
m
er
ic
in
n
at
u
r
e,
r
e
p
r
esen
tin
g
th
e
d
if
f
er
en
t
m
ea
s
u
r
em
en
ts
o
f
v
ital
s
ig
n
s
to
f
lo
at
d
ata
ty
p
e.
T
h
is
en
s
u
r
e
s
th
at
th
ese
d
if
f
er
en
t
m
ea
s
u
r
e
m
en
ts
ar
e
in
s
o
m
e
u
n
if
o
r
m
n
u
m
e
r
ic
f
o
r
m
at,
h
e
n
ce
ac
cu
r
ate
co
m
p
u
tatio
n
s
an
d
s
tatis
tica
l
an
aly
s
es
ca
n
b
e
p
e
r
f
o
r
m
e
d
o
n
th
em
.
T
h
ese
co
n
v
er
s
io
n
s
ar
e
q
u
ite
i
m
p
o
r
tan
t
f
o
r
m
ai
n
tain
in
g
d
ata
in
teg
r
ity
an
d
m
ak
in
g
clin
ic
al
d
ata
am
en
ab
le
to
p
r
o
ce
s
s
in
g
an
d
an
aly
s
is
in
f
u
r
t
h
er
s
tag
es.
Fig
u
r
e
3.
Sam
p
les
of
r
o
ws
with
m
is
s
in
g
v
a
l
u
e
3
.
3
.
3
.
P
a
t
ient
s
m
e
a
s
urem
ent
s
a
g
g
re
g
a
t
io
n
Ag
g
r
eg
atio
n
o
f
p
atien
t
m
ea
s
u
r
em
en
ts
is
a
v
er
y
s
ig
n
if
ican
t
s
tep
to
co
llate
an
d
s
u
m
m
ar
ize
th
e
ex
ten
s
iv
e
clin
ical
d
ata.
W
e
h
a
v
e
g
r
o
u
p
ed
d
ata
b
y
s
u
b
ject
id
s
o
th
at
all
m
ea
s
u
r
em
en
ts
co
n
ce
r
n
in
g
o
n
e
p
atien
t
will b
e
tak
en
as a
s
in
g
le
g
r
o
u
p
,
wh
ich
ca
p
tu
r
es a
co
m
p
r
eh
en
s
iv
e
p
r
o
f
ile
o
f
ea
ch
p
atien
t’
s
v
ital sig
n
s
o
v
er
tim
e
an
d
allo
ws
th
e
d
ataset
to
b
e
m
o
r
e
s
tr
u
ctu
r
ed
a
n
d
co
h
er
en
t.
P
r
ec
is
ely
,
we
r
etr
iev
ed
all
th
e
v
alu
en
u
m
en
t
r
ies
o
f
ea
ch
p
atien
t,
w
h
ich
g
r
an
ted
u
s
a
f
in
e
-
g
r
ain
ed
s
eq
u
en
ce
d
m
ea
s
u
r
em
en
t.
T
h
is
ag
g
r
e
g
atio
n
clea
n
s
in
g
o
f
th
e
d
ataset
r
ed
u
ce
s
th
e
n
o
is
e
am
o
n
g
th
e
d
ata
an
d
u
n
d
e
r
lin
es
r
ath
er
h
id
d
e
n
p
atter
n
s
,
h
e
n
ce
m
a
k
in
g
f
u
t
u
r
e
an
aly
s
is
an
d
m
o
d
el
tr
ain
in
g
ea
s
ier
f
o
r
u
s
.
Fig
u
r
e
4
s
h
o
wca
s
es
a
tab
u
lar
ex
ce
r
p
t
o
f
th
e
d
ataset
f
o
llo
win
g
th
e
tem
p
o
r
al
ag
g
r
eg
atio
n
p
r
o
ce
s
s
.
T
h
e
v
is
u
al
illu
s
tr
ates
h
o
w
h
ig
h
-
f
r
eq
u
en
cy
,
d
is
jo
in
ted
v
ital
s
ig
n
r
ea
d
in
g
s
h
av
e
b
ee
n
s
y
s
tem
atica
lly
g
r
o
u
p
ed
b
y
s
u
b
ject
id
.
B
y
lo
o
k
in
g
at
th
e
r
o
ws,
o
n
e
ca
n
s
ee
h
o
w
is
o
lated
v
al
u
en
u
m
ch
a
r
t e
v
en
ts
ar
e
co
n
s
o
lid
ated
in
to
u
n
if
ied
tim
e
-
s
er
ies
b
lo
ck
s
f
o
r
in
d
iv
id
u
al
p
atien
ts
.
T
h
is
co
n
f
ir
m
s
v
i
s
u
ally
th
at
th
e
r
a
w
d
ata
h
as
b
ee
n
s
u
cc
ess
f
u
lly
tr
a
n
s
f
o
r
m
ed
in
to
co
h
e
r
en
t,
tim
e
-
s
eq
u
en
ce
d
a
r
r
ay
s
,
r
ed
u
cin
g
n
o
is
e
an
d
p
r
ep
a
r
in
g
th
e
d
ataset
f
o
r
ac
cu
r
ate
m
ac
h
i
n
e
lear
n
in
g
in
g
esti
o
n
.
Fig
u
r
e
4.
Sam
p
les
of
ag
g
r
eg
at
ed
m
e
a
s
u
r
e
m
e
n
ts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
I
n
tellig
en
t e
n
g
in
ee
r
in
g
fr
a
mewo
r
k
fo
r
ma
n
a
g
in
g
h
o
s
p
ita
l c
a
r
d
ia
c
a
r
r
est r
eso
u
r
ce
s
(
C
h
a
ms E
d
d
in
e
F
a
th
o
u
n
)
1295
3
.
3
.
4
.
E
nco
din
g
a
nd
f
ix
ing
s
e
qu
ence
s
W
h
en
p
r
ep
ar
i
n
g
th
e
d
ata
f
o
r
m
ac
h
in
e
lear
n
in
g
,
it
is
cr
u
c
ial
to
en
co
d
e
co
n
d
itio
n
lab
el
s
an
d
p
ad
s
eq
u
en
ce
s
to
a
f
ix
ed
le
n
g
th
.
W
e
f
ir
s
t
co
n
v
er
t
th
e
ca
teg
o
r
ic
al
lab
els,
wh
ich
r
ep
r
esen
t
d
if
f
er
en
t
co
n
d
itio
n
s
s
u
c
h
as
AM
I
,
Ar
r
h
y
th
m
ias,
an
d
R
esp
ir
ato
r
y
Failu
r
e/Hy
p
o
x
ia,
in
to
n
u
m
er
ical
f
o
r
m
at
u
s
in
g
a
L
a
b
elE
n
co
d
er
.
T
h
is
s
tep
is
es
s
en
tial
b
ec
au
s
e
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
r
eq
u
ir
e
n
u
m
er
ical
in
p
u
t
to
p
r
o
c
ess
an
d
an
aly
ze
th
e
d
ata
ef
f
ec
tiv
ely
.
On
ce
t
h
e
c
o
n
d
itio
n
lab
els
ar
e
en
co
d
ed
,
we
p
r
o
ce
ed
to
p
ad
th
e
s
eq
u
en
ce
s
o
f
v
alu
en
u
m
e
n
tr
ies
to
a
f
ix
ed
len
g
th
u
s
in
g
th
e
p
ad
s
eq
u
en
ce
s
f
u
n
ctio
n
.
Pad
d
in
g
en
s
u
r
es
th
at
all
s
eq
u
en
ce
s
h
av
e
a
u
n
if
o
r
m
len
g
th
,
wh
ich
f
ac
ilit
ates
co
n
s
is
ten
t
in
p
u
t
d
im
e
n
s
io
n
s
f
o
r
th
e
m
ac
h
i
n
e
lear
n
in
g
m
o
d
el.
B
y
p
ad
d
in
g
th
e
s
eq
u
en
ce
s
,
we
ca
n
h
an
d
le
v
ar
y
in
g
le
n
g
th
s
o
f
p
atie
n
t
m
ea
s
u
r
em
e
n
t
d
a
ta
wh
ile
p
r
eser
v
in
g
th
e
tem
p
o
r
al
o
r
d
er
o
f
t
h
e
m
ea
s
u
r
em
en
ts
.
T
h
ese
s
tep
s
en
h
an
ce
th
e
s
tr
u
ct
u
r
e
an
d
u
s
a
b
ilit
y
o
f
th
e
d
ata,
en
ab
lin
g
ac
cu
r
ate
an
d
e
f
f
icien
t
p
r
ed
i
ctiv
e
m
o
d
elin
g
i
n
s
u
b
s
eq
u
en
t p
h
ases
.
3
.
3
.
5
.
Da
t
a
re
s
a
m
pli
ng
T
h
e
i
n
i
t
i
al
d
a
t
as
e
t
e
x
h
i
b
it
e
d
s
i
g
n
i
f
i
c
a
n
t
c
l
ass
i
m
b
a
l
a
n
c
e
a
c
r
o
s
s
t
h
e
t
h
r
e
e
t
a
r
g
e
t
c
o
n
d
i
ti
o
n
s
:
A
r
r
h
y
t
h
m
i
a
s
(
9
,
9
8
5
ca
s
es),
R
esp
ir
ato
r
y
Failu
r
e/Hy
p
o
x
ia
(
6
,
8
5
9
ca
s
e
s
)
,
an
d
AM
I
(
3
,
1
6
5
ca
s
es).
T
o
a
d
d
r
ess
th
is
,
we
ap
p
lied
r
esam
p
lin
g
tech
n
iq
u
es
s
p
ec
if
ically
,
r
an
d
o
m
u
n
d
er
s
am
p
lin
g
f
o
r
t
h
e
m
ajo
r
ity
class
es
an
d
r
an
d
o
m
o
v
er
s
am
p
lin
g
f
o
r
th
e
m
in
o
r
ity
class
to
ac
h
iev
e
a
b
alan
ce
d
r
ep
r
esen
tatio
n
.
T
h
e
d
ataset
was
s
tan
d
ar
d
ized
t
o
ex
ac
tly
3
,
0
0
0
ca
s
es p
er
c
o
n
d
it
io
n
.
T
h
is
r
esam
p
lin
g
s
tep
is
cr
itic
al
f
o
r
p
r
ev
e
n
tin
g
m
o
d
el
b
ias
to
war
d
th
e
m
ajo
r
ity
class
.
I
n
h
ig
h
-
s
tak
es
clin
ical
s
ce
n
ar
io
s
,
s
u
ch
as
p
r
e
d
ictin
g
im
p
en
d
i
n
g
ca
r
d
iac
ar
r
est,
a
m
o
d
el
b
iased
to
war
d
a
h
ea
v
ily
r
ep
r
esen
ted
class
y
ield
s
u
n
ac
ce
p
tab
ly
h
ig
h
f
alse
-
n
eg
ativ
e
r
ates
f
o
r
m
in
o
r
ity
co
n
d
itio
n
s
.
B
alan
cin
g
th
e
d
ata
e
n
s
u
r
es
th
a
t
th
e
m
ac
h
in
e
lear
n
in
g
m
o
d
el
m
ain
tain
s
eq
u
al
p
r
ed
ictiv
e
s
en
s
itiv
ity
ac
r
o
s
s
all
th
r
ee
c
r
itical
r
is
k
f
ac
to
r
s
,
y
ield
in
g
h
ig
h
ly
r
eliab
le
an
d
clin
ically
ac
tio
n
ab
le
p
r
ed
icti
o
n
s
[
3
5
]
.
T
a
b
le
2
ex
p
licitly
illu
s
tr
ate
th
e
clas
s
d
is
tr
ib
u
tio
n
b
ef
o
r
e
a
n
d
af
ter
th
is
v
ital r
esam
p
lin
g
p
r
o
ce
s
s
.
T
ab
le
2.
N
u
m
b
er
of
ca
s
es
b
ef
o
r
e
an
d
af
ter
r
e
s
a
m
p
l
i
n
g
C
o
n
d
i
t
i
o
n
B
e
f
o
r
e
r
e
s
a
m
p
l
i
n
g
A
f
t
e
r
r
e
s
a
m
p
l
i
n
g
A
r
r
h
y
t
h
m
i
a
s
9
,
9
8
5
3
,
0
0
0
R
e
s
p
i
r
a
t
o
r
y
f
a
i
l
u
r
e
/
H
y
p
o
x
i
a
6
,
8
5
9
3
,
0
0
0
A
M
I
3
,
1
6
5
3
,
0
0
0
3
.
4
.
M
o
del desig
n
T
h
is
s
u
b
s
ec
tio
n
d
escr
ib
es
th
e
m
o
d
el
d
ev
elo
p
m
en
t
p
r
o
ce
s
s
,
i
n
clu
d
in
g
d
ata
s
p
litt
in
g
,
m
o
d
el
s
elec
tio
n
,
h
y
p
er
p
ar
am
eter
tu
n
in
g
,
a
n
d
e
v
alu
atio
n
p
r
o
ce
d
u
r
es.
W
e
ev
alu
ated
m
u
ltip
le
m
ac
h
in
e
lear
n
in
g
al
g
o
r
ith
m
s
t
o
id
en
tify
th
e
m
o
s
t su
itab
le
ap
p
r
o
ac
h
f
o
r
ca
r
d
iac
a
r
r
est p
r
ed
icti
o
n
.
3
.
4
.
1
.
Da
t
a
s
pli
t
t
ing
a
nd
cr
o
s
s
-
v
a
lid
a
t
io
n
T
h
e
r
esam
p
led
d
ataset
was
s
p
lit
in
to
tr
ain
in
g
an
d
test
in
g
s
e
ts
u
s
in
g
an
8
0
/2
0
r
atio
,
wh
er
e
8
0
%
was
u
s
ed
f
o
r
tr
ain
in
g
an
d
2
0
%
was
r
eser
v
ed
f
o
r
test
in
g
.
T
o
e
n
s
u
r
e
r
o
b
u
s
t
ev
alu
atio
n
an
d
p
r
ev
en
t
o
v
er
f
itti
n
g
,
we
im
p
lem
en
ted
5
-
f
o
ld
c
r
o
s
s
-
v
alid
atio
n
o
n
th
e
tr
ain
in
g
s
et.
T
h
is
ap
p
r
o
ac
h
d
iv
id
es
th
e
tr
ai
n
in
g
d
ata
in
to
f
iv
e
s
u
b
s
ets,
tr
ain
in
g
th
e
m
o
d
el
o
n
f
o
u
r
f
o
ld
s
an
d
v
alid
atin
g
o
n
th
e
r
em
ain
in
g
f
o
l
d
,
r
o
tatin
g
th
r
o
u
g
h
all
f
o
ld
s
.
T
h
e
cr
o
s
s
-
v
alid
atio
n
p
r
o
ce
s
s
p
r
o
v
id
es
a
m
o
r
e
r
eliab
le
esti
m
ate
o
f
m
o
d
el
p
e
r
f
o
r
m
an
ce
an
d
h
elp
s
id
en
tif
y
p
o
ten
tial o
v
er
f
itti
n
g
is
s
u
es.
3
.
4
.
2
.
M
o
del selec
t
io
n a
nd
co
m
pa
riso
n
We
ev
alu
ated
m
u
ltip
le
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
to
i
d
en
tif
y
th
e
m
o
s
t
ef
f
ec
tiv
e
ap
p
r
o
ac
h
f
o
r
ca
r
d
iac
ar
r
est
p
r
ed
ictio
n
.
T
h
e
alg
o
r
it
h
m
s
test
ed
in
clu
d
ed
:
(
i)
XGBo
o
s
t
[
3
3
]
,
a
g
r
a
d
ien
t
b
o
o
s
tin
g
f
r
am
ewo
r
k
k
n
o
wn
f
o
r
its
h
ig
h
p
er
f
o
r
m
an
ce
o
n
s
tr
u
ctu
r
ed
d
ata;
(
ii)
R
F
[
3
6
]
,
an
en
s
em
b
le
m
eth
o
d
u
s
in
g
m
u
ltip
le
d
ec
is
io
n
tr
ee
s
;
(
iii)
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
[
3
7
]
with
r
a
d
ial
b
asis
f
u
n
ctio
n
k
er
n
el;
an
d
(
iv
)
lo
g
is
tic
r
eg
r
ess
io
n
(
L
R
)
[
3
8
]
,
a
b
aselin
e
lin
ea
r
class
if
ier
.
E
ac
h
m
o
d
el
was
tr
ain
ed
u
s
in
g
t
h
e
s
am
e
p
r
ep
r
o
ce
s
s
in
g
p
ip
elin
e
an
d
ev
alu
ated
u
s
in
g
i
d
en
tical
m
etr
ics to
en
s
u
r
e
f
air
co
m
p
ar
is
o
n
.
3
.
4
.
3
.
H
y
perpa
ra
m
et
er
t
un
ing
Fo
r
th
e
XGBo
o
s
t
clas
s
if
ier
,
we
p
er
f
o
r
m
ed
h
y
p
er
p
ar
am
ete
r
tu
n
in
g
u
s
in
g
g
r
id
s
ea
r
ch
with
5
-
f
o
l
d
cr
o
s
s
-
v
alid
atio
n
.
T
h
e
h
y
p
er
p
ar
am
eter
s
o
p
tim
ized
in
clu
d
ed
:
lear
n
in
g
r
ate
(
0
.
0
1
,
0
.
1
,
0
.
3
)
,
m
ax
im
u
m
d
ep
th
(
3
,
5,
7
,
1
0
)
,
m
in
im
u
m
ch
ild
weig
h
t
(
1
,
3,
5
)
,
s
u
b
s
am
p
le
(
0
.
8
,
0
.
9
,
1
.
0
)
,
a
n
d
n
u
m
b
er
of
esti
m
ato
r
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
1
2
9
0
-
1
3
0
2
1296
(
1
0
0
,
2
0
0
,
3
0
0
)
.
T
h
e
o
p
tim
al
h
y
p
er
p
ar
am
eter
s
wer
e
s
elec
ted
b
ased
on
th
e
h
ig
h
est
cr
o
s
s
-
v
alid
atio
n
ac
c
u
r
ac
y
wh
ile
m
ain
tain
in
g
g
o
o
d
g
en
er
aliza
tio
n
p
er
f
o
r
m
an
ce
.
3
.
4
.
4
.
1
2
-
ho
ur
predict
io
n m
ec
ha
nis
m
To
p
r
ed
ict
ca
r
d
iac
ar
r
est
12
h
o
u
r
s
in
ad
v
a
n
ce
,
we
im
p
le
m
en
ted
a
tem
p
o
r
al
win
d
o
win
g
ap
p
r
o
ac
h
.
Fo
r
ea
ch
p
atien
t,
we
ex
tr
ac
te
d
v
ital
s
ig
n
m
ea
s
u
r
em
en
ts
r
e
co
r
d
ed
with
in
a
1
2
-
h
o
u
r
win
d
o
w
p
r
ec
ed
i
n
g
th
e
ca
r
d
iac
ar
r
est
ev
en
t
(
f
o
r
p
o
s
itiv
e
ca
s
es)
o
r
a
r
an
d
o
m
l
y
s
elec
ted
1
2
-
h
o
u
r
win
d
o
w
d
u
r
in
g
I
C
U
s
tay
(
f
o
r
n
eg
ativ
e
ca
s
es).
T
h
e
m
ea
s
u
r
em
en
ts
wer
e
ag
g
r
eg
ated
u
s
in
g
s
tati
s
tical
f
ea
tu
r
es
in
clu
d
in
g
m
ea
n
,
s
tan
d
ar
d
d
ev
iatio
n
,
m
in
im
u
m
,
m
ax
im
u
m
,
an
d
tr
en
d
in
d
icato
r
s
.
T
h
is
ap
p
r
o
ac
h
ca
p
tu
r
es
b
o
th
th
e
a
b
s
o
lu
te
v
alu
es
an
d
tem
p
o
r
al
p
atter
n
s
of
v
ital
s
ig
n
s
th
at
m
ay
p
r
ec
ed
e
ca
r
d
iac
ar
r
est,
en
ab
lin
g
ea
r
ly
p
r
ed
ictio
n
wh
ile
m
ain
tain
in
g
clin
ical
r
elev
an
ce
.
3
.
4
.
5
.
M
o
del t
ra
ini
ng
a
nd
ev
a
lua
t
io
n
T
h
e
s
elec
ted
XGBo
o
s
t
m
o
d
el
was
tr
ain
ed
on
th
e
tr
ain
in
g
s
et
u
s
in
g
th
e
o
p
tim
al
h
y
p
er
p
ar
am
eter
s
id
en
tifie
d
th
r
o
u
g
h
g
r
id
s
ea
r
c
h
.
T
h
e
m
o
d
el’
s
p
er
f
o
r
m
an
ce
was
ev
alu
ated
on
th
e
in
d
ep
e
n
d
en
t
test
s
et
u
s
in
g
m
u
ltip
le
m
etr
ics
in
clu
d
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
(
s
en
s
itiv
ity
)
,
s
p
ec
if
icity
,
F1
-
s
co
r
e,
an
d
ar
ea
u
n
d
er
th
e
R
OC
cu
r
v
e
(
AUC
-
R
OC
)
.
We
also
g
en
er
ated
a
co
n
f
u
s
io
n
m
atr
ix
to
v
is
u
alize
th
e
class
if
icatio
n
p
e
r
f
o
r
m
an
ce
ac
r
o
s
s
d
if
f
er
en
t
co
n
d
itio
n
cl
ass
es
an
d
p
lo
tted
R
OC
an
d
p
r
ec
is
io
n
-
r
ec
all
c
u
r
v
es
to
ass
es
s
th
e
m
o
d
el’
s
d
is
cr
im
in
ativ
e
ab
ilit
y
.
3
.
4
.
6
.
XG
B
o
o
s
t
o
bje
ct
iv
e
(
s
um
m
a
r
y
)
XGBo
o
s
t
f
its
an
ad
d
itiv
e
en
s
em
b
le
of
r
eg
r
ess
io
n
tr
ee
s
by
m
i
n
im
izin
g
ℒ
=
∑
(
̂
,
)
+
∑
Ω
(
)
w
i
t
h
̂
=
∑
(
)
a
n
d
l
e
a
f
-
w
e
i
g
h
t
r
e
g
u
lar
izat
io
n
Ω
(
)
=
+
2
∑
2
[
3
3
]
;
m
u
l
t
i
c
l
as
s
o
u
t
p
u
ts
u
s
e
s
o
f
t
m
a
x
c
r
o
s
s
-
e
n
t
r
o
p
y
.
F
u
ll
d
e
r
i
v
a
ti
o
n
o
f
g
r
a
d
i
e
n
t
s
a
n
d
s
p
li
t
g
a
i
n
s
f
o
l
lo
w
s
C
h
e
n
a
n
d
G
u
e
s
t
r
i
n
[
3
3
]
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
4
.
1
.
M
o
del c
o
m
pa
riso
n
re
s
u
lt
s
W
e
ev
alu
ated
f
o
u
r
m
ac
h
in
e
l
ea
r
n
in
g
alg
o
r
ith
m
s
to
id
en
tify
th
e
m
o
s
t
ef
f
ec
tiv
e
ap
p
r
o
ac
h
f
o
r
ca
r
d
iac
ar
r
est
p
r
ed
ictio
n
.
T
ab
le
3
p
r
es
en
ts
th
e
p
er
f
o
r
m
a
n
ce
co
m
p
ar
i
s
o
n
of
XGBo
o
s
t,
R
F,
SVM,
a
n
d
L
R
u
s
in
g
5
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
o
n
th
e
tr
ai
n
in
g
s
et.
XGBo
o
s
t
d
em
o
n
s
tr
ated
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
a
cr
o
s
s
all
m
etr
ics,
ac
h
iev
in
g
th
e
h
ig
h
est
ac
c
u
r
ac
y
(
9
3
.
0
%),
s
en
s
itiv
ity
(
8
9
.
1
5
%),
s
p
ec
if
icity
(
9
0
.
4
3
%),
a
n
d
AUC
-
R
OC
(
0
.
9
4
)
.
R
F
ac
h
iev
ed
th
e
s
ec
o
n
d
-
b
est
p
er
f
o
r
m
an
ce
with
8
8
.
5
%
a
cc
u
r
ac
y
,
f
o
llo
wed
b
y
SVM
(
8
5
.
2
%)
an
d
LR
(
8
2
.
1
%).
T
h
e
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
of
XGBo
o
s
t
can
be
attr
ib
u
ted
to
its
ab
ilit
y
to
h
an
d
le
n
o
n
-
l
in
ea
r
r
elatio
n
s
h
ip
s
,
m
is
s
in
g
v
alu
es,
an
d
f
ea
t
u
r
e
in
t
er
ac
tio
n
s
ef
f
ec
tiv
ely
.
T
ab
le
3.
Per
f
o
r
m
a
n
ce
co
m
p
a
r
i
s
o
n
of
m
ac
h
in
e
lear
n
in
g
m
o
d
e
l
s
M
o
d
e
l
A
c
c
u
r
a
c
y
(
%
)
S
e
n
s
i
t
i
v
i
t
y
(
%
)
S
p
e
c
i
f
i
c
i
t
y
(
%
)
AUC
-
R
O
C
X
G
B
o
o
s
t
9
3
.
0
8
9
.
1
5
9
0
.
4
3
0
.
9
4
RF
8
8
.
5
8
5
.
2
8
7
.
8
0
.
9
1
S
V
M
8
5
.
2
8
1
.
5
8
4
.
3
0
.
8
8
LR
8
2
.
1
7
8
.
9
8
1
.
2
0
.
8
5
4
.
2
.
Cla
s
s
if
ica
t
io
n r
epo
rt
T
T
h
e
class
if
icatio
n
r
ep
o
r
t
p
r
esen
ts
d
etailed
m
etr
ics
s
u
ch
a
s
p
r
ec
is
io
n
,
r
ec
all,
a
n
d
F1
-
s
co
r
e
f
o
r
ea
c
h
co
n
d
itio
n
class
,
p
r
o
v
id
i
n
g
in
s
ig
h
ts
in
to
th
e
m
o
d
el’
s
p
er
f
o
r
m
an
ce
.
T
h
e
XGBo
o
s
t
m
o
d
el
ac
h
iev
ed
h
ig
h
p
er
f
o
r
m
an
ce
ac
r
o
s
s
all
th
r
ee
class
es,
wi
th
C
las
s
0
(
ar
r
h
y
th
m
ias)
s
h
o
win
g
0
.
8
9
p
r
ec
i
s
io
n
an
d
0
.
9
0
r
ec
all
,
C
las
s
1
(
r
esp
ir
at
o
r
y
f
ailu
r
e/
h
y
p
o
x
ia)
ac
h
iev
in
g
0
.
9
1
p
r
ec
is
io
n
an
d
0
.
8
9
r
ec
all,
a
n
d
C
lass
2
(
AM
I
)
d
em
o
n
s
tr
atin
g
1
.
0
0
p
r
ec
is
io
n
an
d
1
.
0
0
r
ec
all.
T
h
e
p
er
f
ec
t
p
er
f
o
r
m
a
n
ce
f
o
r
C
lass
2
m
ay
b
e
attr
ib
u
ted
to
th
e
d
is
tin
ct
p
h
y
s
io
lo
g
ical
p
atte
r
n
s
ass
o
ciate
d
w
ith
AM
I
,
wh
ic
h
ar
e
m
o
r
e
ea
s
ily
d
is
tin
g
u
is
h
ab
le
in
th
e
f
ea
tu
r
e
s
p
ac
e,
b
u
t
th
is
r
esu
lt
r
eq
u
ir
es
ca
r
ef
u
l
in
ter
p
r
etatio
n
a
n
d
m
ay
in
d
icate
th
e
n
ee
d
f
o
r
a
d
d
it
io
n
al
v
alid
atio
n
o
n
in
d
ep
en
d
en
t d
atasets
to
en
s
u
r
e
g
en
er
aliza
b
ilit
y
.
4
.
3
.
Sens
it
iv
it
y
a
nd
s
pecif
icit
y
R
ep
o
r
ted
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
f
o
llo
w
th
e
s
tan
d
ar
d
d
e
f
in
itio
n
s
T
P/(T
P+FN)
T
P/(T
P
+FN)
an
d
T
N/(
T
N+
FP
)
T
N/(
T
N+
F
P),
r
esp
ec
tiv
ely
.
On
th
e
h
eld
-
o
u
t
t
est
s
p
lit,
th
e
f
itted
m
o
d
el
y
ield
ed
ap
p
r
o
x
im
ately
0
.
8
9
1
5
s
en
s
itiv
ity
an
d
0
.
9
0
4
3
s
p
ec
if
icity
.
T
h
ese
v
alu
es
in
d
icate
th
at
th
e
class
if
ier
co
r
r
ec
tly
id
en
tifie
s
m
o
s
t
tr
u
e
p
o
s
itiv
e
ca
s
es wh
ile
also
m
ain
tain
in
g
a
lo
w
r
ate
o
f
f
alse p
o
s
itiv
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
I
n
tellig
en
t e
n
g
in
ee
r
in
g
fr
a
mewo
r
k
fo
r
ma
n
a
g
in
g
h
o
s
p
ita
l c
a
r
d
ia
c
a
r
r
est r
eso
u
r
ce
s
(
C
h
a
ms E
d
d
in
e
F
a
th
o
u
n
)
1297
4
.
4
.
Co
nfusi
o
n m
a
t
rix
T
h
e
co
n
f
u
s
io
n
m
atr
ix
p
r
o
v
id
e
s
a
d
etailed
b
r
ea
k
d
o
wn
o
f
th
e
m
o
d
el’
s
p
er
f
o
r
m
an
c
e
ac
r
o
s
s
t
h
e
d
if
f
er
en
t
co
n
d
itio
n
class
es,
s
h
o
win
g
tr
u
e
p
o
s
itiv
es,
f
alse
p
o
s
itiv
es,
tr
u
e
n
eg
ativ
es,
an
d
f
alse
n
eg
ati
v
es.
T
h
e
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
o
f
th
e
XGBo
o
s
t
class
if
ier
is
v
is
u
ally
d
etailed
i
n
th
e
3
×
3
g
r
id
o
f
th
e
c
o
n
f
u
s
io
n
m
atr
ix
s
h
o
wn
i
n
Fig
u
r
e
5.
T
h
e
r
o
ws
o
f
th
e
g
r
id
co
r
r
esp
o
n
d
to
th
e
tr
u
e
clin
ical
lab
els
(
0
:
Ar
r
h
y
th
m
ias,
1
:
R
esp
ir
ato
r
y
Failu
r
e/Hy
p
o
x
ia,
an
d
2
:
AM
I
)
,
wh
ile
th
e
co
lu
m
n
s
r
ep
r
esen
t
th
e
lab
els
p
r
ed
icted
b
y
th
e
m
o
d
el.
T
h
e
d
iag
o
n
al
ce
lls
ar
e
p
o
p
u
lated
with
th
e
h
i
g
h
est
n
u
m
b
er
s
,
v
is
u
ally
c
o
n
f
ir
m
in
g
th
e
m
o
d
el’
s
h
ig
h
v
o
lu
m
e
o
f
co
r
r
ec
t
p
r
ed
ictio
n
s
(
t
r
u
e
p
o
s
itiv
es)
a
cr
o
s
s
all
ca
teg
o
r
ies.
C
o
n
v
er
s
ely
,
th
e
o
f
f
-
d
iag
o
n
al
ce
lls
c
o
n
tain
n
o
tab
ly
lo
w
n
u
m
b
er
s
,
illu
s
tr
atin
g
t
h
e
m
o
d
e
l’
s
m
in
im
al
cr
o
s
s
-
class
m
is
cla
s
s
if
icatio
n
an
d
lo
w
f
alse
-
p
o
s
itiv
e
r
ates.
Fig
u
r
e
5
.
C
o
n
f
u
s
io
n
m
atr
ix
ev
alu
atin
g
th
e
XGBo
o
s
t
class
if
ier
’
s
p
e
r
f
o
r
m
a
n
c
e
4
.
5
.
RO
C
curv
e
a
nd
prec
is
i
o
n
-
re
ca
ll c
urv
e
T
h
e
R
OC
cu
r
v
e
an
d
p
r
ec
is
io
n
-
r
ec
all
cu
r
v
e
p
r
o
v
id
e
in
s
ig
h
ts
in
to
th
e
tr
ad
e
-
o
f
f
s
b
etwe
en
tr
u
e
p
o
s
itiv
e
r
ates
an
d
f
alse
p
o
s
itiv
e
r
ate
s
,
an
d
p
r
ec
is
io
n
v
er
s
u
s
r
ec
a
ll,
r
esp
ec
tiv
ely
.
Sp
ec
if
ically
,
th
e
d
is
cr
im
in
ato
r
y
ca
p
ab
ilit
y
o
f
th
e
XGBo
o
s
t
class
if
ier
is
v
is
u
ally
v
alid
ated
b
y
th
e
R
OC
cu
r
v
e
illu
s
tr
ated
i
n
F
ig
u
r
e
6.
T
h
e
g
r
ap
h
p
lo
ts
th
e
tr
u
e
p
o
s
itiv
e
r
ate
(
s
en
s
itiv
ity
)
ag
ain
s
t
th
e
f
a
ls
e
p
o
s
itiv
e
r
ate
(
1
-
s
p
ec
if
ici
ty
)
ac
r
o
s
s
v
ar
io
u
s
class
if
icatio
n
th
r
esh
o
ld
s
f
o
r
th
e
th
r
ee
d
is
tin
ct
r
is
k
f
ac
to
r
s
.
T
h
e
cu
r
v
es
p
l
o
ttin
g
clo
s
ely
to
t
h
e
to
p
-
lef
t
co
r
n
er
a
n
d
th
eir
co
r
r
esp
o
n
d
in
g
h
ig
h
ar
ea
u
n
d
er
t
h
e
c
u
r
v
e
(
AUC)
v
alu
es
v
is
u
ally
co
n
f
ir
m
th
e
m
o
d
e
l’
s
r
o
b
u
s
t
ab
ilit
y
t
o
co
r
r
ec
tly
d
is
tin
g
u
is
h
b
etwe
en
Ar
r
h
y
th
m
ias,
AM
I
,
an
d
R
esp
ir
ato
r
y
Failu
r
e.
T
h
is
h
ig
h
d
e
g
r
ee
of
s
ep
a
r
ab
ilit
y
en
s
u
r
es
th
at
clin
ical
r
eso
u
r
ce
s
a
re
not
m
is
allo
ca
ted
d
u
e
to
f
al
s
e
alar
m
s
.
Fig
u
r
e
6
.
R
OC
cu
r
v
e
ev
alu
atin
g
th
e
d
iag
n
o
s
tic
ab
ilit
y
of
th
e
XGBo
o
s
t
clas
s
if
ier
ac
r
o
s
s
th
e
d
is
tin
ct
r
is
k
f
ac
to
r
c
a
t
e
g
o
r
i
e
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
1
2
9
0
-
1
3
0
2
1298
4
.
6
.
F
e
a
t
ure
im
po
rt
a
nce
a
na
ly
s
is
XGBo
o
s
t
g
ain
-
b
ased
im
p
o
r
ta
n
ce
r
an
k
e
d
h
ea
r
t
r
ate
(
lev
el
a
n
d
v
ar
iab
ilit
y
)
,
o
x
y
g
en
s
atu
r
a
tio
n
,
b
lo
o
d
p
r
ess
u
r
e,
an
d
r
esp
ir
ato
r
y
r
ate
h
ig
h
ly
.
T
h
at
o
r
d
er
in
g
m
atch
es
h
o
w
I
C
U
team
s
p
r
io
r
itize
h
em
o
d
y
n
am
ic
an
d
r
esp
ir
ato
r
y
m
o
n
ito
r
in
g
;
it
d
o
e
s
n
o
t
b
y
its
elf
p
r
o
v
e
ca
u
s
ality
,
b
u
t
it
in
cr
ea
s
es
co
n
f
id
en
ce
t
h
at
th
e
m
o
d
el
is
n
o
t
d
r
iv
en
b
y
o
b
s
cu
r
e
ar
tifa
ct
f
ea
tu
r
es.
4
.
7
.
Su
mm
a
ry
o
f
k
ey
f
ind
in
g
s
X
G
B
o
o
s
t
r
e
ac
h
e
d
a
b
o
u
t
9
3
%
ac
c
u
r
a
c
y
w
i
t
h
8
9
.
1
5
% s
e
n
s
i
ti
v
i
ty
a
n
d
9
0
.
4
3
%
s
p
e
c
i
f
i
c
i
t
y
a
n
d
A
U
C
-
R
O
C
0
.
9
4
;
5
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
o
n
th
e
tr
ain
in
g
p
o
r
tio
n
was
s
tab
le
(
m
ea
n
ac
cu
r
ac
y
a
b
o
u
t
9
2
.
8
%,
SD
1
.
2
%).
Per
class
m
etr
ics we
r
e
s
tr
o
n
g
o
v
er
all; p
er
f
ec
t sco
r
es f
o
r
o
n
e
c
lass
war
r
an
t e
x
ter
n
al
ch
ec
k
s
f
o
r
o
v
e
r
f
itti
n
g
.
4
.
8
.
Dis
cus
s
io
n
Ou
r
s
tu
d
y
s
u
g
g
ests
th
at
th
e
X
GB
o
o
s
t
-
b
ased
ap
p
r
o
ac
h
ac
h
ie
v
es
h
ig
h
er
ac
cu
r
ac
y
(
9
3
.
0
%)
co
m
p
ar
e
d
to
ex
is
tin
g
m
o
d
els
in
th
e
liter
atu
r
e,
s
u
ch
as
C
r
e
s
p
í
et
a
l.
(
8
2
.
4
%),
B
lo
m
b
er
g
et
a
l.
(
8
4
.
5
%),
an
d
Kim
et
a
l.
(
8
8
.
6
%).
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
m
ay
b
e
n
ef
it
f
r
o
m
th
e
co
m
p
r
eh
en
s
iv
e
f
ea
tu
r
e
en
g
in
ee
r
in
g
an
d
tem
p
o
r
a
l
win
d
o
win
g
ap
p
r
o
ac
h
with
o
u
t
ad
v
er
s
ely
im
p
ac
tin
g
c
o
m
p
u
tat
io
n
al
ef
f
icien
cy
.
T
h
e
m
o
d
el’
s
ab
ilit
y
to
p
r
ed
ict
ca
r
d
iac
ar
r
est
1
2
h
o
u
r
s
in
ad
v
an
ce
p
r
o
v
id
es
a
s
ig
n
if
ica
n
t
ad
v
an
tag
e
o
v
er
e
x
is
tin
g
ap
p
r
o
ac
h
es
th
at
ty
p
ically
f
o
cu
s
on
s
h
o
r
ter
p
r
e
d
ictio
n
win
d
o
ws
or
lack
th
e
tem
p
o
r
al
r
eso
lu
tio
n
n
ee
d
ed
f
o
r
ea
r
ly
i
n
t
e
r
v
e
n
t
i
o
n
.
T
ab
le
4
p
r
o
v
id
es
a
co
m
p
ar
is
o
n
of
o
u
r
m
o
d
el
with
r
elate
d
w
o
r
k
s
in
ter
m
s
of
p
r
ed
ictio
n
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
s
p
ec
if
icity
,
an
d
d
ataset
ch
ar
ac
t
er
is
tics
.
T
ab
le
4.
C
o
m
p
ar
is
o
n
of
p
r
ed
ic
tio
n
m
o
d
els
f
o
r
ca
r
d
iac
e
v
e
n
t
s
A
u
t
h
o
r
s
P
r
e
d
i
c
t
i
o
n
t
i
m
i
n
g
A
c
c
u
r
a
c
y
(
%
)
S
e
n
s
i
t
i
v
i
t
y
(
%
)
S
p
e
c
i
f
i
c
i
t
y
(
%
)
D
a
t
a
s
e
t
C
r
e
s
p
í
e
t
a
l
.
[
2
7
]
24
h
o
u
r
s
i
n
a
d
v
a
n
c
e
8
2
.
4
8
0
.
8
-
D
a
t
a
f
r
o
m
750
r
e
c
o
r
d
s
o
f
6
2
0
p
a
t
i
e
n
t
s
f
r
o
m
M
a
y
2
0
0
9
t
o
D
e
c
e
m
b
e
r
2
0
2
1
R
a
h
a
d
i
a
n
e
t
a
l
.
[
2
8
]
P
r
e
h
o
s
p
i
t
a
l
,
on
sce
n
e
or
e
n
r
o
u
t
e
-
7
6
.
7
-
P
a
n
-
A
si
a
Tr
a
u
m
a
O
u
t
c
o
mes
S
t
u
d
y
(
P
A
TO
S
)
d
a
t
a
b
a
s
e
S
o
u
d
a
n
et
a
l
.
[
2
9
]
1
-
1
2
h
o
u
r
s
,
60
m
i
n
u
t
e
s
p
r
e
c
e
d
i
n
g
CA
80
-
-
V
i
t
a
l
s
i
g
n
s
f
r
o
m
H
E
R
s
y
s
t
e
m
B
l
o
m
b
e
r
g
e
t
a
l
.
[
3
0
]
E
m
e
r
g
e
n
c
y
c
a
l
l
s
8
4
.
5
8
4
.
5
9
7
.
1
C
o
p
e
n
h
a
g
e
n
E
M
S
d
a
t
a
b
a
s
e
M
u
r
u
g
a
p
p
a
n
et
a
l
.
[
3
1
]
5
mi
n
u
t
e
s
be
f
o
r
e
VF
o
n
s
e
t
1
0
0
9
7
.
1
2
9
7
.
1
2
M
I
T
-
B
I
H
d
a
t
a
b
a
s
e
s
K
i
m
et
a
l
.
[
3
2
]
1
-
6
h
o
u
r
s
b
e
f
o
r
e
o
c
c
u
r
-
r
e
n
c
e
8
8
.
6
5
0
.
7
4
1
.
2
D
a
t
a
f
r
o
m
2
9
,
1
8
1
pa
t
i
e
n
t
s
in
I
C
U
of
t
w
o
h
o
s
-
p
i
t
a
l
s
O
u
r
S
t
u
d
y
12
h
o
u
r
s
i
n
a
d
v
a
n
c
e
9
3
.
0
8
9
.
1
5
9
0
.
4
3
V
i
t
a
l
s
i
g
n
s
f
r
o
m
c
o
m
p
r
e
h
e
n
s
i
v
e
I
C
U
d
a
t
a
se
t
M
I
M
I
C
IV
9
0
0
0
p
a
t
i
e
n
t
s
4
.
8
.
1
.
P
re
dict
io
n a
cc
ura
cy
a
nd
s
en
s
it
iv
it
y
O
u
r
m
o
d
e
l
a
c
h
i
e
v
e
d
an
a
c
c
u
r
a
c
y
of
9
3
%
,
h
i
g
h
e
r
t
h
a
n
t
h
o
s
e
r
e
p
o
r
t
e
d
by
C
r
e
s
p
í
e
t
a
l
.
[
2
7
]
(
8
2
.
4
%
)
,
B
lo
m
b
er
g
et
a
l.
[
3
0
]
(
8
4
.
5
%),
an
d
Kim
et
a
l.
[
3
2
]
(
8
8
.
6
%).
Als
o
,
o
u
r
m
o
d
el
’
s
s
en
s
itiv
ity
o
f
8
9
.
1
5
%
is
h
ig
h
e
r
th
an
th
at
o
f
R
ah
ad
ian
et
a
l.
[
2
8
]
(
7
6
.
7
%)
an
d
Kim
et
a
l.
(
5
0
.
7
%).
4
.
8
.
2
.
Sp
ec
if
icit
y
T
h
e
s
p
ec
if
icity
o
f
o
u
r
m
o
d
el
was
also
v
er
y
h
ig
h
at
9
0
.
4
3
%
.
T
h
is
v
al
u
e
is
clo
s
e
to
th
e
s
p
ec
if
icities
r
ep
o
r
ted
b
y
B
lo
m
b
er
g
et
a
l.
(
9
7
.
1
%)
an
d
Mu
r
u
g
a
p
p
an
e
t
a
l.
(
9
7
.
1
2
%).
Su
ch
co
m
p
a
r
ab
le
p
er
f
o
r
m
an
ce
s
u
g
g
ests
th
at
o
u
r
class
if
ier
m
ain
tain
s
a
lo
w
r
ate
o
f
f
alse
p
o
s
i
tiv
es
r
elativ
e
to
p
r
ev
io
u
s
ly
p
u
b
lis
h
ed
ap
p
r
o
ac
h
es
o
n
r
elate
d
ca
r
d
iac
p
r
ed
ictio
n
t
ask
s
.
4
.
8
.
3
.
Da
t
a
s
et
cha
r
a
ct
er
is
t
ics
I
n
o
u
r
p
r
o
jec
t,
t
h
e
u
s
e
d
d
atas
et
was c
o
m
p
r
e
h
e
n
s
i
v
e
in
n
at
u
r
e
,
r
e
g
ar
d
i
n
g
I
C
U
v
it
al
s
i
g
n
m
e
as
u
r
em
e
n
ts
,
en
h
a
n
ci
n
g
t
h
e
p
r
e
d
i
cti
v
e
p
o
w
er
o
f
t
h
e
m
o
d
el.
T
h
is
c
o
n
tr
a
s
ts
wit
h
s
tu
d
i
es,
s
u
c
h
as
C
r
e
s
p
í
et
a
l
.
[
2
7
]
a
n
d
R
ah
a
d
i
a
n
et
a
l
.
[
2
8
]
,
w
h
e
r
e
d
i
f
f
e
r
e
n
t
d
at
ase
ts
wit
h
d
if
f
e
r
en
t p
ar
am
ete
r
s
we
r
e
u
t
ili
ze
d
.
O
u
r
m
o
d
el
o
f
f
o
r
e
ca
s
t
in
g
m
e
d
ic
al
c
o
n
d
iti
o
n
s
u
p
to
1
2
h
o
u
r
s
i
n
a
d
v
a
n
c
e
b
r
i
n
g
s
s
e
v
e
r
al
a
d
v
a
n
ta
g
es i
n
r
es
o
u
r
c
e
m
a
n
ag
e
m
e
n
t
t
o
t
h
e
I
C
U
.
4
.
8
.
4
.
E
a
rly
det
ec
t
io
n a
nd
inte
rv
ent
io
n
By
p
r
e
d
i
ct
i
n
g
c
o
n
d
i
ti
o
n
s
s
u
c
h
as
a
r
r
h
y
t
h
m
i
as
,
h
y
p
o
x
i
a
,
a
n
d
AM
I
m
u
c
h
in
a
d
v
a
n
c
e
,
our
m
o
d
el
h
e
l
p
s
t
h
e
h
e
a
l
t
h
c
a
r
e
p
r
o
v
i
d
e
r
a
l
l
o
c
a
t
e
r
es
o
u
r
c
e
s
e
f
f
i
c
ie
n
t
l
y
,
p
r
e
p
a
r
e
f
o
r
p
o
s
s
i
b
l
e
c
o
m
p
l
i
c
at
i
o
n
s
,
a
n
d
h
e
n
c
e
t
a
k
e
p
r
e
v
e
n
t
i
v
e
m
ea
s
u
r
es.
T
h
is
p
r
o
ac
tiv
e
a
p
p
r
o
ac
h
g
r
ea
tly
en
h
an
ce
s
p
atie
n
t
o
u
tco
m
es
wh
ile
m
a
k
in
g
u
ltim
ate
u
s
e
of
r
eso
u
r
ce
s
s
p
en
t in
th
e
I
C
U.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
I
n
tellig
en
t e
n
g
in
ee
r
in
g
fr
a
mewo
r
k
fo
r
ma
n
a
g
in
g
h
o
s
p
ita
l c
a
r
d
ia
c
a
r
r
est r
eso
u
r
ce
s
(
C
h
a
ms E
d
d
in
e
F
a
th
o
u
n
)
1299
4
.
8
.
5
.
Co
m
pr
ehens
iv
e
m
o
nit
o
ring
C
o
n
tin
u
al
m
o
n
ito
r
in
g
o
f
t
h
e
p
atien
t’
s
v
ital
s
ig
n
s
,
co
m
b
in
e
d
with
m
u
ltip
le
p
r
e
d
ictiv
e
v
ar
ia
b
les,
h
elp
s
th
e
m
o
d
el
ca
p
tu
r
e
a
m
o
r
e
c
o
m
p
lete
clin
ical
p
ictu
r
e.
T
h
i
s
co
m
p
r
eh
en
s
iv
e
r
ep
r
esen
tatio
n
s
u
p
p
o
r
ts
m
o
r
e
ac
cu
r
ate
id
en
tific
atio
n
o
f
s
u
b
tl
e
p
h
y
s
io
lo
g
ical
ch
a
n
g
es
p
r
ec
e
d
in
g
ca
r
d
iac
d
eter
io
r
atio
n
.
As
a
r
esu
lt,
th
e
m
o
d
el
ca
n
g
en
er
ate
p
r
ed
ictio
n
s
th
at
a
r
e
b
o
th
m
o
r
e
r
eliab
le
an
d
clin
ically
m
ea
n
in
g
f
u
l f
o
r
d
ec
is
io
n
s
u
p
p
o
r
t.
4
.
8
.
6
.
Sca
la
bil
it
y
a
nd
a
da
pta
bil
it
y
B
ec
au
s
e
our
m
o
d
el
is
b
ased
on
ea
s
ily
av
ai
lab
le
I
C
U
d
ata
an
d
a
s
tr
o
n
g
m
ac
h
in
e
lear
n
i
n
g
te
ch
n
iq
u
e,
it
is
s
ca
lab
le
an
d
ad
ap
ta
b
le
to
d
if
f
er
en
t
clin
ical
s
ettin
g
s
.
He
n
ce
,
b
r
o
ad
a
p
p
licab
ilit
y
an
d
i
m
p
ac
t
ar
e
e
n
s
u
r
ed
.
T
h
is
s
tu
d
y
ex
p
lo
r
ed
a
co
m
p
r
eh
en
s
iv
e
p
r
e
d
ictiv
e
m
o
d
el
f
o
r
ca
r
d
iac
a
r
r
est
in
I
C
U
h
ea
r
t
p
atien
ts
u
s
in
g
th
e
MI
MI
C
-
IV
d
atab
ase
with
9
,
0
0
0
p
atien
ts
.
Ho
wev
er
,
f
u
r
t
h
er
a
n
d
in
-
d
e
p
th
s
tu
d
ies
m
ay
be
n
e
ed
ed
to
c
o
n
f
i
r
m
its
g
en
er
aliza
b
ilit
y
,
esp
ec
ially
r
e
g
ar
d
in
g
ex
te
r
n
al
v
alid
atio
n
o
n
in
d
ep
en
d
en
t
d
atasets
f
r
o
m
d
if
f
er
en
t
h
ea
lth
ca
r
e
in
s
titu
tio
n
s
.
Th
e
m
o
d
el’
s
p
er
f
o
r
m
an
ce
o
n
C
lass
2
(
AM
I
)
with
1
0
0
%
p
r
ec
is
io
n
an
d
r
ec
all,
wh
ile
im
p
r
ess
iv
e
,
r
e
q
u
i
r
e
s
c
a
r
e
f
u
l
i
n
t
e
r
p
r
e
ta
t
i
o
n
an
d
m
a
y
b
e
n
e
f
i
t
f
r
o
m
a
d
d
i
ti
o
n
a
l
v
a
l
i
d
a
ti
o
n
to
e
n
s
u
r
e
it
is
not
o
v
e
r
f
i
t
t
i
n
g
to
s
p
e
ci
f
i
c
p
a
t
t
e
r
n
s
in
t
h
e
t
r
ai
n
i
n
g
d
a
t
a
.
Ad
d
i
t
i
o
n
a
l
l
y
,
t
h
e
s
t
u
d
y
f
o
c
u
s
e
d
on
t
h
r
e
e
s
p
ec
i
f
i
c
c
a
r
d
i
a
c
c
o
n
d
i
ti
o
n
s
;
f
u
t
u
r
e
r
e
s
ea
r
c
h
s
h
o
u
ld
ex
p
lo
r
e
th
e
m
o
d
el’
s
a
p
p
licab
ilit
y
to
a
b
r
o
ad
er
r
an
g
e
o
f
ca
r
d
iac
e
v
en
ts
an
d
p
atien
t p
o
p
u
latio
n
s
.
Ou
r
s
tu
d
y
s
u
g
g
ests
th
at
g
r
ad
i
en
t
-
b
o
o
s
ted
m
o
d
els
o
n
s
tr
u
ct
u
r
ed
v
ital
s
ca
n
o
u
tp
e
r
f
o
r
m
g
e
n
er
ic
ea
r
l
y
w
a
r
n
i
n
g
s
c
o
r
e
s
w
h
e
n
t
h
e
t
a
s
k
is
f
r
a
m
e
d
as
m
u
l
t
i
-
c
l
a
s
s
m
e
c
h
a
n
i
s
m
t
r
i
a
g
e
.
F
o
l
l
o
w
o
n
w
o
r
k
c
o
u
l
d
a
d
d
l
a
b
o
r
a
t
o
r
i
e
s
an
d
m
ed
icatio
n
s
,
u
s
e
SHAP
-
s
ty
le
attr
ib
u
tio
n
s
[
2
1
]
an
d
L
I
ME
[
3
9
]
at
t
h
e
b
e
d
s
id
e
[
4
0
]
,
an
d
test
r
ea
l
-
tim
e
o
r
p
r
o
s
p
ec
tiv
e
d
e
p
lo
y
m
e
n
ts
.
5.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
s
u
cc
ess
f
u
lly
d
ev
e
lo
p
ed
an
d
v
alid
ated
a
m
ac
h
i
n
e
lear
n
in
g
-
b
ased
p
r
ed
ictio
n
m
o
d
el
f
o
r
ca
r
d
iac
ar
r
est
r
is
k
in
I
C
U
h
ea
r
t
p
atien
ts
u
s
in
g
v
ital
s
ig
n
m
e
asu
r
em
en
ts
r
ec
o
r
d
ed
1
2
h
o
u
r
s
b
ef
o
r
e
th
e
ev
en
t
.
T
h
e
XGBo
o
s
t
class
if
ier
d
em
o
n
s
tr
ated
s
u
p
e
r
io
r
p
er
f
o
r
m
a
n
ce
co
m
p
ar
ed
to
th
e
o
th
er
ev
alu
ated
alg
o
r
ith
m
s
,
ac
h
iev
in
g
9
3
%
ac
cu
r
ac
y
,
8
9
.
1
5
%
s
en
s
itiv
ity
,
an
d
9
0
.
4
3
%
s
p
ec
if
icity
u
n
d
e
r
5
-
f
o
l
d
cr
o
s
s
-
v
alid
atio
n
.
T
h
e
ab
ilit
y
to
p
r
ed
ict
ca
r
d
iac
ar
r
est
1
2
h
o
u
r
s
in
ad
v
an
ce
p
r
o
v
id
es
clin
ician
s
with
a
m
ea
n
in
g
f
u
l
win
d
o
w
f
o
r
ear
ly
in
ter
v
e
n
tio
n
an
d
m
o
r
e
ef
f
ec
tiv
e
allo
ca
tio
n
o
f
cr
itical
ca
r
e
r
eso
u
r
ce
s
.
T
h
e
k
e
y
co
n
tr
ib
u
tio
n
s
o
f
th
is
wo
r
k
ar
e
f
o
u
r
f
o
ld
an
d
clo
s
e
ly
in
ter
co
n
n
ec
ted
.
First,
we
p
r
o
p
o
s
ed
a
p
r
ed
ictio
n
m
o
d
el
s
p
ec
if
ically
tailo
r
ed
to
I
C
U
h
ea
r
t
p
atien
ts
,
f
o
cu
s
in
g
o
n
clin
ically
r
elev
an
t
m
ec
h
an
is
m
s
o
f
d
eter
io
r
atio
n
.
Seco
n
d
,
we
in
tr
o
d
u
ce
d
a
tem
p
o
r
al
win
d
o
win
g
s
tr
ateg
y
t
h
at
ca
p
t
u
r
es
v
ital
s
ig
n
p
atter
n
s
in
th
e
1
2
h
o
u
r
s
p
r
ec
e
d
in
g
ca
r
d
iac
ar
r
est,
allo
win
g
th
e
m
o
d
el
to
ex
p
lo
it
d
y
n
am
ic
c
h
an
g
es
r
ath
er
th
an
is
o
lated
m
ea
s
u
r
em
en
ts
.
T
h
ir
d
,
we
s
y
s
tem
atica
lly
co
m
p
ar
ed
s
ev
er
al
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
an
d
f
o
u
n
d
th
a
t
XGBo
o
s
t
d
eliv
er
ed
th
e
b
est
o
v
er
all
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
s
tan
d
ar
d
class
if
icatio
n
m
et
r
ics.
Fo
u
r
th
,
we
co
n
d
u
cte
d
f
ea
tu
r
e
im
p
o
r
tan
ce
a
n
aly
s
is
,
wh
ich
h
ig
h
lig
h
ted
h
ea
r
t
r
ate,
o
x
y
g
e
n
s
atu
r
a
tio
n
,
an
d
b
lo
o
d
p
r
ess
u
r
e
as
lead
in
g
p
r
ed
icto
r
s
,
th
e
r
eb
y
r
ein
f
o
r
cin
g
th
e
clin
ical
p
lau
s
ib
ilit
y
an
d
i
n
ter
p
r
etab
ilit
y
o
f
t
h
e
m
o
d
el’
s
o
u
tp
u
ts
.
Desp
ite
th
ese
s
tr
en
g
th
s
,
s
ev
er
al
lim
itatio
n
s
s
h
o
u
ld
b
e
ac
k
n
o
wled
g
e
d
to
s
u
p
p
o
r
t
a
b
alan
ce
d
in
ter
p
r
etatio
n
o
f
th
e
f
in
d
i
n
g
s
.
T
h
e
m
o
d
el
was
tr
ain
e
d
an
d
v
a
lid
ated
o
n
a
s
in
g
le
d
ataset,
s
o
ex
ter
n
al
v
alid
ati
o
n
u
s
in
g
d
ata
f
r
o
m
o
t
h
er
h
o
s
p
itals
an
d
h
ea
lth
ca
r
e
s
y
s
tem
s
is
n
ee
d
ed
to
co
n
f
ir
m
its
g
e
n
er
aliza
b
ilit
y
.
T
h
e
p
e
r
f
ec
t
p
er
f
o
r
m
an
ce
o
b
s
er
v
e
d
f
o
r
th
e
AM
I
class
s
u
g
g
ests
a
r
is
k
o
f
o
v
e
r
f
itti
n
g
f
o
r
th
is
s
u
b
g
r
o
u
p
a
n
d
war
r
an
ts
f
u
r
th
er
an
aly
s
is
with
ad
d
itio
n
al
an
d
m
o
r
e
d
iv
e
r
s
e
s
am
p
les.
I
n
ad
d
itio
n
,
th
e
cu
r
r
e
n
t
wo
r
k
f
o
cu
s
e
d
o
n
th
r
ee
s
p
ec
if
ic
ca
r
d
iac
co
n
d
itio
n
s
,
wh
ich
m
a
y
lim
it
th
e
ap
p
licab
ilit
y
o
f
th
e
ap
p
r
o
ac
h
to
o
th
er
f
o
r
m
s
o
f
c
ar
d
iac
d
eter
io
r
atio
n
o
r
to
b
r
o
a
d
er
p
atien
t
p
o
p
u
la
tio
n
s
.
Fin
ally
,
th
e
m
o
d
el
h
as
n
o
t
y
et
b
ee
n
ev
alu
ated
in
r
ea
l
-
tim
e
clin
ical
wo
r
k
f
lo
ws,
an
d
p
r
o
s
p
ec
tiv
e
s
tu
d
ies
ar
e
r
eq
u
ir
ed
to
ass
ess
it
s
p
r
ac
tical
im
p
ac
t
o
n
d
ec
is
io
n
m
ak
in
g
,
r
eso
u
r
ce
u
tili
za
tio
n
,
an
d
p
atien
t o
u
tco
m
es.
B
u
ild
in
g
o
n
th
ese
r
esu
lts
,
s
ev
er
al
d
ir
ec
ti
o
n
s
f
o
r
f
u
tu
r
e
r
e
s
ea
r
ch
ca
n
en
h
a
n
ce
b
o
th
ac
c
u
r
ac
y
an
d
clin
ical
u
s
ef
u
ln
ess
.
On
e
im
p
o
r
tan
t
a
v
en
u
e
is
to
in
teg
r
ate
ad
d
itio
n
al
clin
ical
v
ar
iab
les,
s
u
ch
as
lab
o
r
at
o
r
y
v
alu
es,
m
ed
icatio
n
ex
p
o
s
u
r
e,
an
d
d
em
o
g
r
ap
h
ic
ch
ar
ac
te
r
is
tics
,
to
en
r
ich
t
h
e
f
ea
tu
r
e
s
p
ac
e
an
d
p
o
ten
tially
im
p
r
o
v
e
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
.
An
o
th
er
p
r
o
m
is
in
g
d
ir
ec
ti
o
n
is
to
in
co
r
p
o
r
ate
ex
p
lain
a
b
le
AI
tech
n
iq
u
es
to
m
ak
e
in
d
iv
id
u
al
p
r
e
d
ictio
n
s
m
o
r
e
tr
an
s
p
ar
e
n
t a
n
d
to
h
elp
cli
n
ician
s
u
n
d
er
s
tan
d
w
h
ich
v
ital
s
ig
n
p
atter
n
s
d
r
iv
e
r
is
k
esti
m
ates.
Fu
r
th
er
wo
r
k
is
also
n
ee
d
ed
to
d
ev
elo
p
an
d
test
r
ea
l
-
tim
e
p
r
ed
ictio
n
s
y
s
tem
s
th
at
ca
n
b
e
em
b
ed
d
e
d
with
in
ex
is
tin
g
I
C
U
m
o
n
ito
r
in
g
in
f
r
astru
ctu
r
e,
e
n
ab
lin
g
c
o
n
tin
u
o
u
s
r
is
k
esti
m
atio
n
at
th
e
b
e
d
s
id
e
.
Pro
s
p
ec
tiv
e
v
alid
atio
n
s
tu
d
i
es sh
o
u
ld
b
e
co
n
d
u
cted
to
q
u
a
n
tify
h
o
w
s
u
ch
s
y
s
tem
s
in
f
lu
en
ce
clin
ical
d
ec
is
io
n
s
,
r
eso
u
r
ce
m
an
ag
em
e
n
t,
an
d
p
atien
t
o
u
tco
m
es
in
r
o
u
tin
e
p
r
ac
tice.
I
n
p
ar
allel,
ex
p
lo
r
in
g
d
ee
p
lear
n
in
g
ar
ch
itectu
r
es
th
at
b
etter
ca
p
tu
r
e
co
m
p
lex
tem
p
o
r
al
d
ep
e
n
d
en
cies
in
v
ital
s
ig
n
d
ata
m
ay
y
ield
ad
d
itio
n
al
p
er
f
o
r
m
an
ce
g
ai
n
s
,
p
r
o
v
id
ed
t
h
at
in
ter
p
r
etab
ilit
y
a
n
d
co
m
p
u
tatio
n
al
ef
f
icien
cy
a
r
e
p
r
eser
v
ed
.
I
n
c
o
n
clu
s
io
n
,
th
is
s
tu
d
y
d
e
m
o
n
s
tr
ates
th
e
p
o
ten
tial
o
f
m
ac
h
in
e
lear
n
in
g
,
an
d
in
p
ar
tic
u
lar
XGBo
o
s
t,
f
o
r
ea
r
ly
p
r
e
d
ictio
n
o
f
ca
r
d
iac
ar
r
est
in
I
C
U
h
ea
r
t
p
ati
en
ts
b
ased
o
n
r
o
u
tin
ely
co
l
lecte
d
v
ital
s
ig
n
s
.
T
h
e
m
o
d
el’
s
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
an
d
1
2
-
h
o
u
r
p
r
ed
ictio
n
h
o
r
izo
n
s
u
g
g
est
th
at
it
ca
n
s
er
v
e
as
a
v
alu
ab
le
to
o
l
Evaluation Warning : The document was created with Spire.PDF for Python.