I
nte
rna
t
io
na
l J
o
urna
l o
f
Adv
a
nces in Applie
d Science
s
(
I
J
AAS)
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
,
p
p
.
29
~
41
I
SS
N:
2252
-
8
8
1
4
,
DOI
:
1
0
.
1
1
5
9
1
/ijaas
.
v
1
5
.
i
1
.
pp
29
-
41
29
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
a
s
.
ia
esco
r
e.
co
m
Enha
ncing
sleep
diso
rder diag
no
sis
t
hro
ug
h ensem
b
le
ML
mo
dels: a co
mpre
hensiv
e study o
n i
nso
mnia
and slee
p apn
ea
Sa
t
y
a
pra
k
a
s
h Swa
in
1,
4
,
B
ino
d K
um
a
r
P
a
t
t
a
na
y
a
k
1
,
M
ihi
r
Na
ra
y
a
n M
o
ha
nty
2
,
A
m
iy
a
K
um
a
r
Sa
ho
o
3
,
Su
v
endra
K
um
a
r
J
a
y
a
s
ing
h
4
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
I
n
st
i
t
u
t
e
o
f
Te
c
h
n
i
c
a
l
Ed
u
c
a
t
i
o
n
a
n
d
R
e
se
a
r
c
h
,
S
o
A
U
n
i
v
e
r
si
t
y
,
B
h
u
b
a
n
e
sw
a
r
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
I
n
st
i
t
u
t
e
o
f
Te
c
h
n
i
c
a
l
Ed
u
c
a
t
i
o
n
a
n
d
R
e
sea
r
c
h
,
S
o
A
U
n
i
v
e
r
si
t
y
,
B
h
u
b
a
n
e
sw
a
r
,
I
n
d
i
a
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
,
A
r
y
a
n
I
n
st
i
t
u
t
e
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
,
B
h
u
b
a
n
e
sw
a
r
,
I
n
d
i
a
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
,
I
n
st
i
t
u
t
e
o
f
M
a
n
a
g
e
me
n
t
a
n
d
I
n
f
o
r
mat
i
o
n
Te
c
h
n
o
l
o
g
y
,
C
u
t
t
a
c
k
,
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
No
v
1
8
,
2
0
2
4
R
ev
is
ed
J
u
n
1
7
,
2
0
2
5
Acc
ep
ted
J
u
l 1
4
,
2
0
2
5
S
lee
p
d
iso
r
d
e
rs
a
re
c
o
m
m
o
n
a
n
d
c
a
n
sig
n
ifi
c
a
n
tl
y
h
a
rm
h
u
m
a
n
h
e
a
lt
h
,
with
in
so
m
n
ia
a
n
d
sle
e
p
a
p
n
e
a
b
e
in
g
t
h
e
m
o
st
p
re
v
a
len
t
c
o
n
d
it
i
o
n
s.
T
h
e
se
d
iso
rd
e
rs
a
re
o
ften
d
iffi
c
u
lt
t
o
d
e
t
e
c
t
a
n
d
trea
t
a
c
c
u
ra
tely
.
Alth
o
u
g
h
m
a
c
h
in
e
lea
rn
in
g
(M
L)
tec
h
n
i
q
u
e
s
h
a
v
e
sh
o
wn
p
r
o
m
ise
in
imp
ro
v
in
g
d
iag
n
o
stic
p
re
c
isio
n
a
n
d
p
e
rso
n
a
li
z
e
d
trea
tme
n
t,
m
o
st
e
x
isti
n
g
stu
d
ies
re
ly
o
n
si
n
g
le
-
so
u
rc
e
d
a
ta
o
r
c
o
n
v
e
n
ti
o
n
a
l
M
L
m
o
d
e
ls,
wh
ich
li
m
it
t
h
e
ir
ro
b
u
s
tn
e
ss
a
n
d
g
e
n
e
ra
li
z
a
b
il
it
y
a
c
ro
ss
d
iv
e
rse
p
o
p
u
lati
o
n
s.
To
a
d
d
re
ss
th
is
re
se
a
rc
h
g
a
p
,
th
is
stu
d
y
in
te
g
ra
tes
m
u
lt
i
-
m
o
d
a
l
d
a
ta
a
n
d
e
n
se
m
b
le
lea
rn
in
g
tec
h
n
i
q
u
e
s
t
o
e
n
h
a
n
c
e
a
c
c
u
ra
c
y
,
in
ter
p
re
tab
il
it
y
,
a
n
d
re
a
l
-
ti
m
e
a
p
p
l
ica
b
il
it
y
i
n
d
iag
n
o
sin
g
in
so
m
n
ia
a
n
d
sle
e
p
a
p
n
e
a
.
A
d
a
t
a
se
t
o
f
4
0
0
sa
m
p
les
wa
s
c
o
ll
e
c
ted
th
r
o
u
g
h
m
a
n
u
a
l
m
e
th
o
d
s
a
n
d
i
n
tern
e
t
o
f
t
h
in
g
s
(Io
T)
d
e
v
ice
s
fro
m
m
u
l
ti
p
l
e
so
u
rc
e
s.
S
tatisti
c
a
l
tec
h
n
i
q
u
e
s
we
re
a
p
p
l
i
e
d
fo
r
d
a
ta
c
lea
n
in
g
,
f
o
ll
o
we
d
b
y
p
ri
n
c
ip
a
l
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
(
P
CA)
t
o
re
d
u
c
e
d
ime
n
sio
n
a
li
ty
a
n
d
imp
ro
v
e
train
i
n
g
e
fficie
n
c
y
.
F
o
u
r
b
a
se
M
L
m
o
d
e
ls
:
d
e
c
isio
n
tree
(DT),
s
u
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(S
VM),
n
a
i
v
e
Ba
y
e
s
(
NB),
a
n
d
ra
n
d
o
m
fo
re
st
(R
F
)
w
e
re
in
it
iall
y
train
e
d
a
n
d
e
v
a
lu
a
ted
.
S
u
b
se
q
u
e
n
tl
y
,
a
b
o
o
sti
n
g
-
b
a
se
d
e
n
se
m
b
le
m
o
d
e
l
wa
s
imp
lem
e
n
ted
to
f
u
rth
e
r
imp
r
o
v
e
p
e
rfo
rm
a
n
c
e
.
Th
e
p
r
o
p
o
se
d
g
ra
d
ien
t
b
o
o
st
in
g
m
o
d
e
l
with
RF
a
s
t
h
e
b
a
se
lea
rn
e
r
a
c
h
iev
e
d
th
e
h
i
g
h
e
st
d
iag
n
o
stic
a
c
c
u
ra
c
y
o
f
9
6
.
0
1
%
.
Th
e
re
su
lt
s
d
e
m
o
n
stra
te
th
a
t
e
n
se
m
b
le
M
L
m
o
d
e
ls
c
o
m
b
in
e
d
wi
th
m
u
l
ti
-
m
o
d
a
l
d
a
t
a
sig
n
ifi
c
a
n
tl
y
e
n
h
a
n
c
e
th
e
a
c
c
u
ra
c
y
o
f
in
so
m
n
ia an
d
sle
e
p
a
p
n
e
a
d
ia
g
n
o
sis.
K
ey
w
o
r
d
s
:
B
o
o
s
tin
g
m
o
d
el
Dec
is
io
n
tr
ee
Gr
ad
ien
t
b
o
o
s
t
Naiv
e
B
ay
es
Prin
cip
al
co
m
p
o
n
en
t a
n
al
y
s
is
R
an
d
o
m
f
o
r
est
Su
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Su
v
en
d
r
a
K
u
m
ar
J
ay
asin
g
h
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
,
I
n
s
titu
te
o
f
Ma
n
ag
e
m
en
t a
n
d
I
n
f
o
r
m
ati
o
n
T
ec
h
n
o
lo
g
y
Go
u
r
av
Palace
,
Ma
th
a
s
ah
i,
p
o
s
t
-
T
u
ls
ip
u
r
C
u
ttack
-
7
5
3
0
0
8
,
Od
is
h
a,
I
n
d
ia
E
m
ail:
s
jay
asin
g
h
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Millio
n
s
o
f
in
d
iv
id
u
als
wo
r
l
d
wid
e
ar
e
im
p
ac
ted
b
y
th
e
p
r
ev
alen
t
an
d
s
er
io
u
s
p
r
o
b
le
m
o
f
s
leep
p
r
o
b
lem
s
.
Am
o
n
g
th
e
m
an
y
s
leep
-
r
elate
d
is
s
u
es,
in
s
o
m
n
ia
an
d
s
leep
ap
n
ea
ar
e
p
ar
ticu
lar
ly
co
m
m
o
n
a
n
d
p
r
o
v
id
e
s
u
b
s
tan
tial
d
if
f
icu
ltie
s
f
o
r
p
r
o
p
er
d
iag
n
o
s
is
an
d
ef
f
icien
t
tr
ea
tm
en
t.
T
h
ese
s
leep
p
r
o
b
lem
s
ca
n
h
av
e
s
er
io
u
s
ef
f
ec
ts
o
n
o
n
e'
s
p
h
y
s
ical
h
ea
lth
,
m
en
tal
h
ea
lth
,
to
f
in
d
g
e
n
er
al
q
u
ality
o
f
life
if
t
h
ey
ar
e
n
o
t
tr
ea
ted
ef
f
e
c
tiv
ely
.
Fo
r
tu
n
ately
,
th
e
d
ev
elo
p
m
en
t
o
f
m
ac
h
in
e
lear
n
i
n
g
(
ML
)
p
r
esen
ts
a
p
r
o
m
is
in
g
way
to
im
p
r
o
v
e
th
e
ac
cu
r
ac
y
o
f
d
iag
n
o
s
tic
p
r
o
ce
d
u
r
es
an
d
,
m
o
r
e
im
p
o
r
tan
tly
,
t
o
cu
s
to
m
ize
t
r
ea
tm
en
t
p
lan
s
t
o
m
ee
t
t
h
e
s
p
ec
if
ic
n
ee
d
s
o
f
ea
ch
p
er
s
o
n
.
T
h
is
s
t
u
d
y
p
r
o
v
id
e
s
a
n
o
v
el
m
eth
o
d
th
at
u
s
es
ML
m
o
d
els
to
tac
k
le
th
e
ch
allen
g
in
g
p
r
o
b
lem
s
o
f
d
etec
tin
g
in
s
o
m
n
ia
an
d
s
leep
ap
n
ea
.
W
ith
th
e
u
s
e
o
f
co
m
m
o
n
ML
m
o
d
el
s
lik
e
d
ec
is
io
n
tr
ee
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
c
h
20
2
6
:
29
-
41
30
(
DT
)
,
s
u
p
p
o
r
t
v
ec
t
o
r
m
ac
h
in
es
(
SVM)
,
n
aiv
e
B
ay
es
(
N
B
)
,
an
d
r
an
d
o
m
f
o
r
est
(
R
F),
we
h
av
e
p
r
o
p
o
s
ed
a
co
m
p
r
eh
e
n
s
iv
e
h
y
b
r
id
ML
m
o
d
el
af
ter
ad
d
in
g
b
o
o
s
tin
g
m
eth
o
d
s
th
at
in
clu
d
es
in
f
o
r
m
ati
o
n
f
r
o
m
th
ese
f
o
u
r
d
if
f
er
en
t
ML
b
ase
m
eth
o
d
s
.
T
h
ese
d
if
f
er
en
t
m
o
d
els
ea
ch
h
av
e
im
p
r
ess
iv
e
d
iag
n
o
s
tic
ac
cu
r
ac
ies
wh
en
u
s
ed
alo
n
e,
r
an
g
in
g
f
r
o
m
8
7
.
9
6
to
9
1
.
1
7
%.
T
h
is
s
tu
d
y
en
te
r
ed
th
e
f
ield
o
f
b
o
o
s
tin
g
m
o
d
els
lik
e
a
d
ap
tiv
e
b
o
o
s
tin
g
(
Ad
aBo
o
s
t
)
,
g
r
ad
ie
n
t
b
o
o
s
t
in
g
,
e
x
tr
em
e
g
r
ad
ien
t
b
o
o
s
tin
g
(
XGBo
o
s
t
)
,
ca
teg
o
r
ical
b
o
o
s
tin
g
(
C
atB
o
o
s
t
)
,
an
d
lig
h
t
g
r
ad
ien
t
b
o
o
s
tin
g
m
ac
h
in
e
(
L
ig
h
tGB
M
)
,
wh
ich
ar
e
in
clu
d
ed
in
o
u
r
co
m
p
a
r
is
o
n
.
W
ith
an
asto
u
n
d
in
g
ac
cu
r
a
cy
r
ate
o
f
9
6
%,
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
m
o
d
el
s
tan
d
s
o
u
t
as
th
e
clea
r
win
n
er
am
o
n
g
th
em
.
T
h
is
o
u
ts
tan
d
in
g
ac
h
i
ev
em
en
t
h
ig
h
lig
h
ts
th
e
en
o
r
m
o
u
s
p
o
ten
tial
o
f
ML
to
g
r
ea
tly
im
p
r
o
v
e
th
e
ac
cu
r
ac
y
o
f
d
iag
n
o
s
es
f
o
r
s
leep
d
is
o
r
d
er
s
.
A
ca
r
ef
u
lly
s
elec
ted
d
ataset
o
f
4
0
0
d
ata
p
o
in
ts
,
g
ath
er
ed
t
h
r
o
u
g
h
a
c
o
m
b
in
atio
n
o
f
m
an
u
al
g
ath
er
in
g
an
d
d
ata
s
u
p
p
lied
b
y
in
ter
n
et
o
f
th
in
g
s
(
I
o
T
)
d
ev
ic
es,
s
er
v
es
as
th
e
b
asis
f
o
r
th
is
s
tu
d
y
.
T
h
is
d
ataset
co
n
tain
s
a
wid
e
r
an
g
e
o
f
p
eo
p
le
f
r
o
m
d
if
f
e
r
en
t
o
cc
u
p
a
tio
n
s
,
s
u
ch
as
f
a
r
m
er
s
,
s
o
licito
r
s
,
p
r
o
f
ess
o
r
s
,
an
d
d
o
ct
o
r
s
.
I
t
in
clu
d
es
a
wid
e
r
an
g
e
o
f
in
f
o
r
m
atio
n
,
s
u
ch
as
d
em
o
g
r
ap
h
ics,
s
p
ec
if
ic
s
leep
p
atter
n
s
,
th
o
r
o
u
g
h
m
e
d
ical
h
is
to
r
ies
,
an
d
o
th
e
r
im
p
o
r
tan
t
c
h
ar
ac
ter
is
tics
.
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
h
i
g
h
lig
h
t
th
e
p
r
o
f
o
u
n
d
in
f
lu
en
ce
t
h
at
ML
m
o
d
els
ca
n
h
av
e
o
n
th
e
f
ield
o
f
d
iag
n
o
s
in
g
s
le
ep
d
is
o
r
d
er
s
.
I
n
ad
d
itio
n
to
p
r
o
v
id
in
g
a
s
ig
n
if
ican
t
im
p
r
o
v
e
m
en
t
in
d
iag
n
o
s
tic
ac
cu
r
ac
y
,
th
e
b
o
o
s
t
in
g
m
o
d
el,
p
ar
ticu
lar
ly
wh
en
p
ai
r
ed
with
g
r
ad
ien
t
b
o
o
s
tin
g
,
h
as
th
e
p
o
ten
tial
to
lay
th
e
g
r
o
u
n
d
wo
r
k
f
o
r
th
e
cr
ea
tio
n
o
f
an
ad
v
an
ce
d
clin
ical
d
ec
is
io
n
s
u
p
p
o
r
t sy
s
tem
.
B
y
u
tili
zin
g
en
s
em
b
le
ML
m
o
d
els
d
esig
n
ed
to
i
d
en
tify
in
s
o
m
n
ia
an
d
s
leep
ap
n
ea
,
th
is
s
tu
d
y
p
r
esen
ts
a
n
o
v
el
m
eth
o
d
f
o
r
d
iag
n
o
s
in
g
s
leep
d
is
o
r
d
er
s
.
T
h
e
m
a
jo
r
ity
o
f
c
u
r
r
en
t
r
esear
ch
o
n
d
iag
n
o
s
in
g
s
leep
d
is
o
r
d
er
s
u
s
es
s
in
g
le
-
s
o
u
r
ce
d
ata
o
r
co
n
v
en
tio
n
al
ML
m
o
d
els,
wh
ich
f
r
eq
u
en
tly
lack
g
en
er
alis
ab
ilit
y
an
d
r
o
b
u
s
tn
ess
ac
r
o
s
s
a
r
an
g
e
o
f
d
em
o
g
r
ap
h
ics.
T
o
im
p
r
o
v
e
ac
c
u
r
ac
y
,
in
ter
p
r
etab
ilit
y
,
a
n
d
r
ea
l
-
tim
e
ap
p
licab
ilit
y
in
th
e
d
iag
n
o
s
is
o
f
in
s
o
m
n
ia
an
d
s
leep
ap
n
ea
,
a
s
u
b
s
tan
ti
al
r
esear
ch
g
ap
s
till
ex
is
ts
in
th
e
in
teg
r
atio
n
o
f
m
u
lti
-
m
o
d
al
d
ata
an
d
en
s
em
b
le
lear
n
in
g
tec
h
n
iq
u
es.
I
n
c
o
n
tr
ast
to
ea
r
lier
r
esear
ch
th
at
m
o
s
tly
co
n
ce
n
t
r
ated
o
n
in
d
iv
id
u
al
d
is
o
r
d
e
r
s
o
r
u
s
ed
co
n
v
e
n
tio
n
al
d
iag
n
o
s
tic
tech
n
iq
u
es,
t
h
is
wo
r
k
in
co
r
p
o
r
ates
s
o
p
h
is
ticated
en
s
em
b
le
tech
n
iq
u
es
to
in
cr
ea
s
e
th
e
p
r
ec
is
io
n
an
d
d
ep
en
d
ab
ilit
y
o
f
s
leep
d
is
o
r
d
er
d
iag
n
o
s
es.
T
h
e
r
o
b
u
s
tn
ess
o
f
th
e
d
etec
tio
n
p
r
o
ce
s
s
is
im
p
r
o
v
e
d
b
y
t
h
is
wo
r
k
b
y
co
m
b
in
in
g
d
if
f
e
r
en
t
ML
alg
o
r
ith
m
s
,
wh
ich
ad
d
r
ess
es
th
e
s
h
o
r
tco
m
in
g
s
o
f
p
r
ev
i
o
u
s
m
o
d
els
th
at
f
r
eq
u
e
n
tly
ig
n
o
r
e
th
e
co
m
p
lex
ity
a
n
d
o
v
er
la
p
o
f
s
y
m
p
to
m
s
o
f
v
ar
io
u
s
s
leep
d
is
o
r
d
e
r
s
.
B
y
p
r
o
v
id
in
g
a
co
m
p
r
eh
en
s
iv
e,
a
u
t
o
m
ated
s
o
lu
tio
n
t
h
at
n
o
t
o
n
ly
ex
p
ed
ites
d
iag
n
o
s
is
b
u
t
also
lo
wer
s
th
e
p
o
s
s
ib
ili
t
y
o
f
h
u
m
an
er
r
o
r
,
th
is
s
tu
d
y
m
ar
k
s
a
s
ig
n
if
ican
t
ad
v
an
ce
m
en
t
in
th
e
f
ield
an
d
u
ltima
tely
im
p
r
o
v
es
p
atien
t
o
u
tco
m
es
an
d
s
tr
ea
m
lin
es
clin
ical
wo
r
k
f
lo
ws.
A
s
y
s
tem
lik
e
th
is
m
ig
h
t
p
r
o
v
i
d
e
m
ed
ical
p
er
s
o
n
n
el
with
in
d
iv
id
u
alize
d
d
iag
n
o
s
tic
in
s
ig
h
ts
,
p
er
m
itti
n
g
cu
s
to
m
ized
th
er
a
p
y
m
o
d
alities
an
d
ev
en
tu
ally
en
h
an
cin
g
o
u
tc
o
m
e
s
f
o
r
p
eo
p
le
s
u
f
f
er
i
n
g
f
r
o
m
s
le
ep
p
r
o
b
lem
s
.
T
h
is
s
tu
d
y
h
as
r
ev
iewe
d
a
n
u
m
b
er
o
f
wo
r
k
s
r
elate
d
to
h
e
alth
ca
r
e
ap
p
licatio
n
s
u
s
in
g
ML
o
r
d
ee
p
lear
n
in
g
(
DL
)
m
o
d
els.
T
h
ey
h
av
e
tak
en
a
d
if
f
e
r
en
t
p
ar
am
et
er
ized
d
ataset
f
o
r
th
eir
p
r
e
d
ictio
n
wo
r
k
.
B
u
t
h
er
e
we
h
av
e
ta
k
en
a
d
ataset
th
at
c
o
n
tain
s
1
3
p
ar
a
m
eter
s
,
wh
ich
ar
e
m
o
s
t
r
ele
v
an
t
to
s
leep
in
g
d
is
o
r
d
er
s
.
Als
o
,
t
h
is
wo
r
k
h
as
im
p
lem
en
ted
th
e
n
ew
en
s
em
b
le
b
o
o
s
tin
g
m
o
d
els
to
in
cr
ea
s
e
th
e
ac
c
u
r
ac
y
lev
els
f
o
r
g
o
o
d
p
r
ed
ictio
n
s
o
f
d
is
ea
s
e.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
C
ar
m
es
et
a
l
.
[
1
]
in
tr
o
d
u
c
ed
a
s
em
i
-
au
to
m
atic
a
p
p
r
o
ac
h
f
o
r
s
leep
ap
n
ea
d
iag
n
o
s
is
u
tili
zin
g
th
r
ee
-
ch
an
n
el
p
o
ly
s
o
m
n
o
g
r
ap
h
y
(
PS
G)
d
ata
.
T
h
is
m
eth
o
d
em
p
lo
y
s
a
n
E
last
icNe
t
class
if
ier
to
ac
h
iev
e
a
r
em
ar
k
ab
le
9
7
.
9
%
ac
cu
r
ac
y
in
s
leep
ap
n
ea
class
if
icatio
n
.
No
tab
ly
,
it
m
in
im
izes
th
e
tim
e
-
co
n
s
u
m
in
g
m
an
u
al
an
aly
s
is
o
f
PS
G
r
ec
o
r
d
in
g
s
,
ef
f
ec
tiv
ely
m
itig
atin
g
in
ter
-
s
c
o
r
er
v
a
r
iab
ilit
y
.
W
ith
p
o
ten
ti
al
as
a
v
iab
le
s
leep
ap
n
ea
s
cr
ee
n
in
g
to
o
l
in
clin
ica
l
s
ettin
g
s
,
th
is
m
eth
o
d
en
h
an
c
es
ef
f
icien
cy
an
d
r
eliab
ilit
y
,
s
h
o
wca
s
in
g
its
v
alu
e
in
ad
v
an
ci
n
g
d
ia
g
n
o
s
tic
ac
cu
r
ac
y
an
d
s
tr
ea
m
lin
in
g
th
e
ass
ess
m
en
t p
r
o
ce
s
s
f
o
r
s
leep
s
p
ec
ia
lis
ts
.
C
h
o
y
o
n
et
a
l
.
[
2
]
p
r
o
p
o
s
e
a
n
in
v
en
tiv
e
s
o
lu
ti
o
n
b
y
in
te
g
r
atin
g
I
o
T
an
d
ML
to
b
o
ls
ter
h
ea
lth
m
o
n
ito
r
in
g
an
d
co
m
b
at
th
e
p
a
n
d
em
ic
m
o
r
e
ef
f
icien
tly
.
I
t
h
i
g
h
lig
h
ts
th
e
s
h
o
r
tco
m
in
g
s
o
f
ex
is
tin
g
m
o
n
ito
r
in
g
s
y
s
tem
s
an
d
in
tr
o
d
u
ce
s
an
I
o
T
-
b
ased
m
eth
o
d
f
o
r
r
ea
l
-
t
im
e
h
ea
lth
d
ata
c
o
llectio
n
,
f
ac
ilit
atin
g
p
r
o
m
p
t
d
etec
tio
n
o
f
C
OVI
D
-
1
9
s
ev
er
ity
.
T
h
e
en
v
is
io
n
e
d
s
y
s
tem
s
ee
k
s
to
d
eliv
er
h
o
lis
tic
h
ea
lth
ca
r
e,
r
em
o
te
co
m
m
u
n
icatio
n
,
a
n
d
e
m
er
g
e
n
cy
s
u
p
p
o
r
t,
o
f
f
e
r
in
g
a
p
r
ag
m
atic
ap
p
r
o
ac
h
to
allev
iate
th
e
r
ep
er
c
u
s
s
io
n
s
o
f
th
e
p
an
d
e
m
ic.
Ko
r
k
alain
en
et
a
l
.
[
3
]
p
r
o
p
o
s
ed
a
m
o
d
el
b
y
u
tili
zin
g
DL
an
d
d
iv
er
s
e
ep
o
ch
d
u
r
atio
n
s
t
o
u
n
co
v
er
n
eg
lecte
d
s
leep
-
wak
e
tr
an
s
itio
n
s
.
T
h
e
ex
am
in
atio
n
o
f
4
4
6
o
b
s
tr
u
ctiv
e
s
leep
ap
n
ea
(
OSA)
p
atien
ts
in
d
icate
s
th
at
tr
ad
itio
n
al
ap
p
r
o
ac
h
es
m
ay
u
n
d
e
r
esti
m
ate
s
leep
f
r
ag
m
en
tatio
n
in
s
ev
er
e
OSA.
Sh
o
r
ter
ep
o
ch
-
to
-
ep
o
c
h
d
u
r
atio
n
s
r
e
v
ea
l
h
eig
h
ten
ed
wak
ef
u
ln
ess
an
d
r
ed
u
ce
d
r
ap
id
ey
e
m
o
v
e
m
en
t
/
n
o
n
-
R
E
M
3
(
R
E
M/N3
)
,
em
p
h
asizin
g
th
e
n
ec
ess
ity
f
o
r
a
th
o
r
o
u
g
h
a
n
aly
s
is
o
f
s
leep
ar
ch
itectu
r
e
wh
en
e
v
alu
atin
g
s
le
ep
d
is
o
r
d
er
s
.
Sh
ah
in
et
a
l
.
[
4
]
p
r
esen
t
a
d
u
al
-
p
h
ase
a
u
to
m
ated
m
eth
o
d
to
id
en
tify
in
s
o
m
n
ia
f
r
o
m
o
v
er
n
ig
h
t
elec
tr
o
en
ce
p
h
al
o
g
r
am
(
EEG
)
r
ec
o
r
d
i
n
g
s
,
ad
d
r
ess
in
g
th
e
d
ef
icien
cy
o
f
p
r
o
m
p
t
in
s
o
m
n
ia
d
etec
tio
n
t
o
o
ls
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
E
n
h
a
n
ci
n
g
s
leep
d
is
o
r
d
er d
ia
g
n
o
s
is
th
r
o
u
g
h
e
n
s
emb
le
ML
mo
d
els:
… (
S
a
tya
p
r
a
ka
s
h
S
w
a
in
)
31
T
h
e
ap
p
r
o
ac
h
u
tili
ze
s
d
ee
p
n
eu
r
al
n
etwo
r
k
m
o
d
els,
tr
ain
in
g
b
o
th
s
leep
s
tag
e
an
d
ep
o
ch
-
lev
el
in
s
o
m
n
ia
d
etec
tio
n
m
o
d
els.
Su
b
ject
-
le
v
el
f
ea
tu
r
es
e
x
tr
ac
ted
f
r
o
m
th
ese
m
o
d
els
f
ac
ilit
ate
th
e
u
ltima
te
b
in
a
r
y
class
if
icatio
n
o
f
s
u
b
jects
in
to
co
n
tr
o
l
o
r
in
s
o
m
n
iac
ca
teg
o
r
i
es.
Ass
es
s
m
en
t
o
n
1
1
5
p
a
r
ticip
an
ts
d
em
o
n
s
tr
ates
en
co
u
r
a
g
in
g
o
u
tc
o
m
es,
in
clu
d
in
g
an
F1
-
s
co
r
e
o
f
0
.
8
8
,
8
4
%
s
en
s
itiv
ity
,
an
d
9
1
%
s
p
ec
if
icit
y
,
u
n
d
er
s
co
r
i
n
g
its
p
o
ten
tial c
lin
ical
ap
p
licab
ilit
y
.
Ku
o
an
d
C
h
en
[
5
]
in
tr
o
d
u
c
e
a
n
o
v
el
s
h
o
r
t
-
tim
e
in
s
o
m
n
ia
d
etec
tio
n
s
y
s
tem
u
tili
zin
g
r
ef
in
ed
co
m
p
o
s
ite
m
u
lti
-
s
ca
le
en
tr
o
p
y
(
R
C
MSE
)
an
aly
s
is
o
n
s
in
g
le
-
ch
an
n
el
s
leep
elec
tr
o
o
cu
lo
g
r
ap
h
y
(
E
OG)
.
W
ith
an
SVM
cla
s
s
if
ier
,
th
e
p
r
o
p
o
s
ed
s
y
s
tem
d
em
o
n
s
tr
ated
h
ig
h
ac
cu
r
ac
y
(
8
9
.
3
1
%),
s
en
s
itiv
ity
(
9
6
.
6
3
%),
an
d
F1
-
s
co
r
e
(
9
0
.
0
4
%)
wh
en
t
ested
o
n
3
2
s
u
b
jects,
h
alf
h
ea
l
th
y
an
d
h
alf
with
in
s
o
m
n
ia.
R
C
MSE
em
er
g
es
as
a
v
alu
ab
le
f
ea
tu
r
e
f
o
r
s
h
o
r
t
-
d
u
r
atio
n
in
s
o
m
n
ia
d
etec
tio
n
,
an
d
th
e
s
in
g
le
-
ch
an
n
el
s
leep
E
OG
en
h
an
ce
s
h
o
m
ec
ar
e
a
p
p
licab
ilit
y
f
o
r
p
o
ten
tial in
teg
r
atio
n
in
t
o
p
o
r
tab
le
PS
G
s
y
s
tem
s
.
I
s
lam
et
a
l
.
[
6
]
d
elv
e
in
t
o
th
e
in
f
lu
en
ce
o
f
tec
h
n
o
lo
g
ical
ad
v
an
ce
m
en
ts
o
n
h
u
m
an
h
ea
lth
,
s
p
ec
if
ically
th
e
s
u
r
g
e
in
in
s
o
m
n
ia
lin
k
ed
t
o
v
ir
tu
al
en
g
a
g
em
en
ts
an
d
s
e
d
en
tar
y
h
a
b
its
.
Ack
n
o
wled
g
i
n
g
th
e
d
r
awb
ac
k
s
o
f
co
s
tly
an
d
tim
e
-
in
ten
s
iv
e
m
e
d
ical
test
s
,
th
e
au
th
o
r
s
s
u
g
g
est
an
in
tellig
en
t
ML
m
o
d
el
to
p
r
ed
ict
ch
r
o
n
i
c
in
s
o
m
n
ia.
Utilizin
g
s
ev
en
cla
s
s
if
ier
s
,
th
e
lo
g
is
tic
r
eg
r
ess
io
n
m
o
d
el
d
e
m
o
n
s
tr
ates
ex
ce
p
t
io
n
al
ac
cu
r
ac
y
at
9
8
%,
p
r
esen
tin
g
a
c
o
m
p
ellin
g
an
d
ef
f
ec
tiv
e
m
eth
o
d
f
o
r
d
etec
tin
g
in
s
o
m
n
ia
in
in
d
i
v
id
u
als,
p
ar
ticu
lar
ly
in
s
ettin
g
s
with
lim
ited
r
eso
u
r
ce
s
.
Z
u
lf
ik
er
et
a
l
.
[
7
]
tack
le
th
e
wid
esp
r
ea
d
h
ea
lth
is
s
u
e
o
f
in
s
o
m
n
ia,
a
p
r
ev
ale
n
t
s
leep
d
is
o
r
d
er
ass
o
ciate
d
with
m
en
tal
h
ea
lth
ch
allen
g
es
a
n
d
s
u
b
s
tan
ce
ab
u
s
e.
I
n
tr
o
d
u
cin
g
a
n
ML
a
p
p
r
o
ac
h
,
th
e
m
u
ltil
ay
er
s
tack
in
g
m
o
d
el
p
r
ed
icts
in
s
o
m
n
ia
b
ased
o
n
s
o
cio
-
d
em
o
g
r
ap
h
ic
f
ac
to
r
s
,
ac
h
iev
in
g
a
n
o
tab
le
8
8
.
6
0
%
ac
cu
r
ac
y
.
L
ev
er
ag
in
g
p
r
i
n
cip
al
co
m
p
o
n
en
t
an
aly
s
is
(
PC
A)
f
o
r
f
ea
tu
r
e
r
ed
u
ctio
n
,
th
e
p
r
o
p
o
s
ed
en
s
em
b
le
m
o
d
el
s
u
r
p
ass
es o
th
er
lead
in
g
class
if
ier
s
lik
e
Ad
aBo
o
s
t a
n
d
g
r
ad
ien
t b
o
o
s
t
,
s
h
o
wca
s
in
g
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
ac
r
o
s
s
m
u
ltip
le
ef
f
icac
y
m
et
r
ics.
T
h
is
em
p
h
asizes its
p
o
ten
tial a
s
an
ef
f
ec
tiv
e
to
o
l
f
o
r
p
r
ed
ictin
g
i
n
s
o
m
n
ia.
Alg
h
wir
i
et
a
l
.
[
8
]
in
th
eir
c
r
o
s
s
-
s
ec
tio
n
al
s
tu
d
y
with
1
,
6
0
0
u
n
iv
e
r
s
ity
s
tu
d
en
ts
,
th
e
in
v
esti
g
atio
n
f
o
cu
s
es
o
n
s
leep
d
is
tu
r
b
an
ce
s
,
em
p
lo
y
in
g
lo
g
is
tic
r
eg
r
ess
io
n
,
an
d
ad
v
a
n
ce
d
ML
tech
n
iq
u
es.
T
h
e
s
tu
d
y
d
is
clo
s
es
a
7
0
%
p
r
ev
alen
ce
o
f
p
o
o
r
s
leep
q
u
ality
,
id
e
n
tify
i
n
g
th
e
RF
m
o
d
el
as
th
e
m
o
s
t
ac
cu
r
ate
p
r
ed
icto
r
with
7
4
%
ac
cu
r
ac
y
an
d
9
5
%
s
p
ec
if
icity
.
Facto
r
s
ass
o
ciate
d
with
b
etter
s
leep
q
u
ality
i
n
clu
d
e
ag
e
an
d
tea
co
n
s
u
m
p
tio
n
,
wh
ile
r
is
k
f
ac
to
r
s
co
n
s
is
t
o
f
elec
tr
o
n
ics
u
s
ag
e,
h
ea
d
ac
h
es,
s
y
s
tem
ic
d
is
ea
s
es,
an
d
n
ec
k
p
ain
.
T
h
ese
r
esu
lts
p
r
o
v
id
e
v
alu
ab
le
in
s
ig
h
ts
f
o
r
en
h
an
cin
g
s
tu
d
en
t
well
-
b
ein
g
an
d
cr
af
tin
g
tailo
r
ed
in
ter
v
en
tio
n
s
.
J
ay
asin
g
et
a
l.
[
9
]
p
r
o
p
o
s
ed
a
m
o
d
el
f
o
r
wea
th
e
r
f
o
r
ec
asti
n
g
,
h
ig
h
lig
h
tin
g
t
h
at
its
n
o
n
lin
e
a
r
n
atu
r
e
is
im
p
ac
ted
b
y
clim
ate
c
h
an
g
e
a
n
d
th
e
b
u
tter
f
ly
e
f
f
ec
t.
T
h
e
r
e
s
ea
r
ch
in
tr
o
d
u
ce
s
h
y
b
r
id
s
o
f
t
co
m
p
u
tin
g
m
o
d
els,
m
er
g
in
g
SVM
,
m
u
lti
-
lay
er
p
er
ce
p
tio
n
,
an
d
f
u
zz
y
l
o
g
ic
,
aim
ed
at
im
p
r
o
v
in
g
th
e
ac
cu
r
ac
y
o
f
wea
th
e
r
p
r
ed
ictio
n
s
f
o
r
Delh
i.
Ass
ess
m
en
t
m
etr
ics
s
u
ch
as
m
ea
n
ab
s
o
lu
te
er
r
o
r
(
MA
E
)
,
r
o
o
t
m
ea
n
s
q
u
a
r
ed
er
r
o
r
(
R
MSE
)
,
r
elativ
e
ab
s
o
lu
te
e
r
r
o
r
(
R
AE
)
,
a
n
d
r
o
o
t
r
elativ
e
s
q
u
ar
ed
e
r
r
o
r
(
R
R
SE
)
s
u
b
s
tan
tiate
th
e
ef
f
ec
tiv
e
n
ess
o
f
th
ese
m
o
d
els in
en
h
an
ci
n
g
f
o
r
ec
asti
n
g
p
r
ec
is
io
n
.
Swain
et
a
l
.
[
1
0
]
in
th
eir
s
tu
d
y
co
m
p
ar
e
d
if
f
e
r
en
t M
L
m
o
d
el
s
f
o
r
tack
l
in
g
th
e
co
m
p
lex
tas
k
o
f
illn
ess
p
r
o
g
n
o
s
is
,
s
p
ec
if
ically
f
o
cu
s
in
g
o
n
in
d
iv
id
u
als
ag
ed
2
0
a
n
d
ab
o
v
e,
p
ar
ticu
la
r
ly
th
o
s
e
with
Hb
A1
c
lev
els
s
u
r
p
ass
in
g
6
.
5
%,
in
d
icativ
e
o
f
d
iab
etic
d
is
ea
s
es.
E
m
p
lo
y
in
g
I
o
T
,
th
e
r
esear
ch
ass
ess
es
r
is
k
f
ac
to
r
s
ak
in
to
h
ea
r
t
d
is
ea
s
es,
h
ig
h
lig
h
tin
g
th
e
s
ig
n
if
ican
ce
o
f
ML
in
d
iag
n
o
s
is
an
d
p
r
e
v
en
tio
n
.
T
h
e
in
co
r
p
o
r
atio
n
o
f
ad
v
an
ce
d
tec
h
n
o
lo
g
ies
s
u
ch
as
I
o
T
,
clo
u
d
a
p
p
licatio
n
s
,
a
n
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
h
o
ld
s
p
o
ten
tial
f
o
r
im
p
r
o
v
in
g
h
ea
lth
ca
r
e
s
tr
ateg
ies in
ad
d
r
ess
in
g
h
ea
r
t
d
is
ea
s
es.
Mo
n
d
al
et
a
l
.
[
1
1
]
f
o
cu
s
o
n
th
e
in
cr
ea
s
in
g
r
is
k
o
f
co
r
o
n
ar
y
h
ea
r
t
d
is
ea
s
e
(
C
HD)
in
ad
u
lt
s
,
u
tili
zin
g
ML
f
o
r
tim
ely
d
etec
tio
n
.
W
ith
a
d
ataset
o
f
4
,
2
4
0
in
s
tan
ce
s
an
d
1
5
f
ea
tu
r
es,
d
iv
er
s
e
ML
m
o
d
els
wer
e
ex
am
in
ed
,
h
ig
h
lig
h
tin
g
RF
an
d
g
r
ad
ie
n
t
b
o
o
s
t
class
if
ier
s
f
o
r
th
eir
ac
cu
r
ac
y
.
T
h
e
u
ltima
te
en
s
em
b
le
m
o
d
el,
co
m
b
in
in
g
b
o
o
s
tin
g
m
o
d
els,
d
em
o
n
s
tr
ated
a
n
o
u
ts
tan
d
in
g
9
2
.
2
6
%
ac
c
u
r
ac
y
,
h
i
g
h
lig
h
tin
g
its
ef
f
icac
y
in
ea
r
ly
C
HD
r
is
k
p
r
ed
ictio
n
with
m
in
i
m
al
f
alse n
eg
ativ
es.
Ku
m
ar
an
d
R
ed
d
y
[
1
2
]
d
elv
e
in
to
d
if
f
er
e
n
t
ca
r
d
iac
d
is
ea
s
es,
h
ig
h
lig
h
tin
g
h
ea
r
t
f
ailu
r
e
(
HF)
,
co
r
o
n
a
r
y
a
r
ter
y
d
is
ea
s
e
(
C
AD)
,
an
d
ca
r
d
io
v
ascu
lar
d
is
ea
s
e
(
C
V
D
)
.
T
h
e
p
r
o
p
o
s
ed
d
iag
n
o
s
tic
s
y
s
tem
u
tili
ze
s
s
u
p
er
v
is
ed
lear
n
in
g
,
s
p
ec
if
ic
ally
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
te
ch
n
iq
u
e,
to
ac
cu
r
ately
d
etec
t
HF
,
u
tili
zin
g
th
e
C
lev
elan
d
d
ataset
.
Ach
iev
in
g
an
o
u
ts
tan
d
in
g
9
7
.
1
0
%
ac
cu
r
a
cy
,
th
e
m
o
d
el
p
r
o
v
es
ef
f
ec
tiv
e
in
au
to
m
ated
HF
d
iag
n
o
s
is
,
o
u
t
p
er
f
o
r
m
in
g
alter
n
ativ
e
m
eth
o
d
s
.
T
h
e
em
p
h
a
s
is
o
n
g
r
ad
ie
n
t
b
o
o
s
tin
g
s
ig
n
if
ies
a
s
ig
n
if
ican
t
ad
v
an
ce
m
e
n
t in
h
ea
r
t
d
is
ea
s
e
d
etec
tio
n
tech
n
iq
u
es.
Das
et
a
l
.
[
1
3
]
ad
d
r
ess
in
g
th
e
u
r
g
e
n
t
n
ee
d
f
o
r
ea
r
ly
h
ea
r
t
d
is
ea
s
e
d
etec
tio
n
,
th
is
s
tu
d
y
u
t
ilizes
ML
with
f
iv
e
b
o
o
s
tin
g
m
eth
o
d
s
:
Ad
aBo
o
s
t
,
g
r
ad
ien
t
b
o
o
s
tin
g
,
XGBo
o
s
t
,
C
atB
o
o
s
t,
an
d
L
ig
h
tGB
M
o
n
th
e
Un
iv
er
s
ity
o
f
C
alif
o
r
n
ia,
I
r
v
in
e
(
UC
I
)
C
lev
elan
d
HD
d
a
taset.
C
o
m
p
ar
ativ
e
a
n
aly
s
is
d
em
o
n
s
tr
ates
th
at
Ad
aBo
o
s
t
an
d
XGBo
o
s
t
ex
c
el
with
th
e
h
ig
h
est
ac
cu
r
ac
y
(
≈
9
3
%),
p
r
ec
is
io
n
(
0
.
9
4
)
,
r
ec
all
(
0
.
9
3
)
,
an
d
F1
-
s
co
r
es
(
0
.
9
3
)
.
T
h
ese
r
esu
l
ts
s
u
r
p
ass
r
ec
en
t
en
s
em
b
le
te
ch
n
iq
u
es
a
n
d
s
tack
in
g
class
if
ier
s
ap
p
lied
to
th
e
s
am
e
d
ataset.
Sen
an
d
Ver
m
a
[
1
4
]
ad
v
o
ca
te
s
f
o
r
ad
v
an
ce
d
d
iag
n
o
s
tic
m
eth
o
d
s
u
s
in
g
ML
.
T
h
e
p
r
o
p
o
s
al
in
tr
o
d
u
ce
s
a
s
o
f
t
v
o
tin
g
m
eta
-
class
if
ier
co
m
p
o
s
ed
o
f
C
atB
o
o
s
t
,
L
ig
h
tGB
M
,
G
au
s
s
ian
n
aiv
e
B
ay
es
(
GNB)
,
RF
,
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
c
h
20
2
6
:
29
-
41
32
XGBo
o
s
t
,
s
u
r
p
ass
in
g
tr
ad
itio
n
al
ap
p
r
o
ac
h
es.
C
o
n
d
u
cted
o
n
a
co
m
b
in
e
d
UC
I
h
ea
r
t
d
is
ea
s
e
an
d
Stat
lo
g
d
ataset,
th
e
ex
p
er
im
en
t
y
ield
s
r
em
ar
k
ab
le
r
esu
lts
,
ac
h
iev
in
g
a
9
1
.
8
5
%
ac
cu
r
ac
y
an
d
a
0
.
9
3
4
4
ar
ea
u
n
d
e
r
th
e
cu
r
v
e
s
co
r
e
(
AUC)
.
T
h
ese
o
u
tco
m
es
s
ig
n
if
y
en
h
a
n
ce
d
d
iag
n
o
s
tic
ef
f
icac
y
in
p
r
e
d
ictin
g
h
ea
r
t
d
is
ea
s
e,
s
h
o
wca
s
in
g
th
e
p
o
ten
tial
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el.
Af
r
ee
n
et
a
l
.
[
1
5
]
u
tili
ze
ML
,
em
p
lo
y
in
g
C
atB
o
o
s
t
an
d
L
ig
h
tGB
M
m
o
d
els
to
p
r
ed
ict
an
d
ca
teg
o
r
ize
liv
er
d
is
ea
s
e,
ad
d
r
ess
in
g
th
e
i
n
cr
ea
s
in
g
p
r
ev
alen
ce
o
f
p
atien
ts
with
d
iv
er
s
e
co
m
p
licatio
n
s
.
I
m
p
lem
en
tin
g
a
g
r
ad
ien
t
b
o
o
s
tin
g
-
b
ased
class
if
ier
with
f
ea
tu
r
e
s
elec
tio
n
t
h
r
o
u
g
h
p
r
ep
r
o
ce
s
s
in
g
en
h
a
n
ce
s
m
o
d
el
ac
c
u
r
ac
y
.
C
atB
o
o
s
t
ac
h
iev
es
th
e
h
ig
h
est
ac
cu
r
ac
y
at
8
6
.
8
%,
wh
ile
L
ig
h
tGB
M
attain
s
8
2
.
6
%,
d
em
o
n
s
tr
atin
g
th
ei
r
ef
f
icac
y
in
ea
r
l
y
id
en
tific
atio
n
o
f
liv
er
d
is
ea
s
es f
o
r
im
p
r
o
v
ed
p
atien
t o
u
tco
m
es.
3.
RE
S
E
ARCH
D
E
SI
G
N
3
.
1
.
M
a
chine le
a
rning
A
cr
u
cial
ar
ea
o
f
ar
tific
ial
in
te
llig
en
ce
(
AI
)
is
ML
[
1
6
]
,
wh
ich
en
ab
les co
m
p
u
ter
s
to
m
ak
e
j
u
d
g
m
e
n
ts
an
d
lear
n
f
r
o
m
d
ata
o
n
th
eir
o
wn
.
I
t
is
a
f
ast
-
d
ev
el
o
p
in
g
d
is
cip
lin
e
with
a
wid
e
r
an
g
e
o
f
ap
p
licatio
n
s
,
all
o
f
wh
ich
ar
e
b
asically
b
ased
o
n
d
ata
an
aly
s
is
to
id
en
tif
y
t
r
en
d
s
an
d
lin
k
ag
es.
ML
e
n
c
o
m
p
ass
es
v
ar
io
u
s
ap
p
r
o
ac
h
es,
in
clu
d
in
g
:
–
Su
p
er
v
is
ed
lear
n
in
g
,
wh
er
e
al
g
o
r
ith
m
s
u
s
e
lab
eled
d
ata
f
o
r
l
ea
r
n
in
g
.
–
Un
s
u
p
er
v
is
ed
lear
n
in
g
f
o
r
u
n
c
o
v
er
in
g
h
id
d
e
n
p
atter
n
s
in
u
n
l
ab
eled
d
ata
.
–
R
ein
f
o
r
ce
m
en
t le
ar
n
i
n
g
,
a
p
p
li
ed
to
in
ter
ac
tiv
e
s
ettin
g
s
.
ML
is
u
s
ed
i
n
s
p
ee
ch
an
d
i
m
ag
e
r
ec
o
g
n
itio
n
,
r
ec
o
m
m
e
n
d
atio
n
s
y
s
tem
s
,
au
t
o
n
o
m
o
u
s
v
eh
icles,
h
ea
lth
ca
r
e,
an
d
r
o
b
o
tics
,
ad
v
an
cin
g
co
m
p
u
ter
v
is
io
n
,
lan
g
u
ag
e
u
n
d
er
s
tan
d
in
g
,
an
d
r
o
b
o
tics
.
Data
q
u
ality
,
m
o
d
el
in
ter
p
r
eta
b
ilit
y
,
an
d
et
h
ical
is
s
u
es
ar
e
ch
allen
g
es.
N
ev
er
th
eless
,
ML
co
n
tin
u
es
to
r
esh
ap
e
in
d
u
s
tr
ies,
r
ea
f
f
ir
m
in
g
its
p
iv
o
tal
p
o
s
itio
n
in
co
n
tem
p
o
r
ar
y
tech
n
o
lo
g
y
an
d
AI
,
a
n
d
h
av
in
g
a
s
ig
n
if
ic
an
t
im
p
ac
t
o
n
h
o
w
we
liv
e
o
u
r
liv
es.
3
.
2
.
M
a
chine le
a
rning
a
rc
hite
ct
ure
Data
co
llectio
n
an
d
p
r
e
p
r
o
ce
s
s
in
g
ar
e
th
e
f
ir
s
t
s
tep
s
in
th
e
ML
ar
ch
itectu
r
e
p
r
o
c
ess
,
as
s
h
o
wn
in
Fig
u
r
e
1
.
Nex
t,
d
ep
e
n
d
in
g
o
n
th
e
task
at
h
an
d
,
a
s
u
itab
le
ML
m
o
d
el
is
ch
o
s
en
.
Fo
r
th
e
p
u
r
p
o
s
es
o
f
m
o
d
el
tr
ain
in
g
an
d
ev
alu
atio
n
,
th
e
d
ata
is
d
iv
id
ed
in
to
tr
ain
in
g
an
d
v
alid
atio
n
s
ets.
W
h
en
h
y
p
e
r
p
ar
am
eter
tu
n
i
n
g
is
s
u
cc
ess
f
u
l
in
o
p
tim
izin
g
m
o
d
el
p
er
f
o
r
m
an
ce
,
th
e
m
o
d
el
is
u
s
ed
to
m
ak
e
p
r
e
d
ictio
n
s
a
b
o
u
t
th
e
r
ea
l
w
o
r
ld
.
T
o
au
to
m
ate
d
ec
is
io
n
-
m
ak
i
n
g
an
d
p
atter
n
id
en
tific
atio
n
ac
r
o
s
s
a
v
ar
iety
o
f
ap
p
licatio
n
s
,
ML
is
a
p
o
wer
f
u
l
to
o
l.
C
o
n
s
tan
t m
o
n
ito
r
in
g
an
d
i
n
ter
p
r
etab
ilit
y
en
s
u
r
e
th
e
m
o
d
el
r
e
m
ain
s
ef
f
ec
tiv
e
an
d
u
n
d
er
s
tan
d
ab
le.
Fig
u
r
e
1
.
Ar
c
h
itectu
r
e
o
f
ML
3
.
3
.
E
ns
em
ble
lea
rning
A
ML
tech
n
iq
u
e
ter
m
ed
en
s
e
m
b
le
lear
n
in
g
[
1
7
]
co
m
b
in
es th
e
p
r
ed
ictio
n
s
o
f
v
ar
io
u
s
s
ep
a
r
ate
m
o
d
els
(
co
m
m
o
n
ly
r
ef
er
r
ed
to
as
"b
ase
m
o
d
els"
o
r
"we
ak
lear
n
er
s
")
in
o
r
d
er
to
en
h
an
ce
o
v
er
all
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
.
T
h
e
th
e
o
r
y
b
e
h
i
n
d
en
s
em
b
le
lear
n
in
g
is
th
at
b
y
co
m
b
in
in
g
th
e
ad
v
an
ta
g
es
o
f
v
ar
io
u
s
m
o
d
els,
it
ca
n
less
en
th
e
d
r
aw
b
ac
k
s
o
f
ea
ch
m
o
d
el
s
ep
ar
ately
an
d
g
en
er
ate
p
r
ed
ictio
n
s
th
at
a
r
e
m
o
r
e
r
eliab
le
an
d
ac
cu
r
ate.
T
h
er
e
ar
e
s
ev
e
r
al
en
s
em
b
le
lear
n
in
g
m
o
d
els
av
ailab
le,
s
u
ch
as
b
ag
g
in
g
,
b
o
o
s
tin
g
,
v
o
tin
g
,
an
d
s
tack
in
g
.
T
h
is
wo
r
k
h
as im
p
le
m
en
ted
a
v
o
tin
g
an
d
b
o
o
s
tin
g
m
o
d
el
with
th
e
b
est esti
m
ato
r
,
wh
ich
ar
e
d
escr
ib
e
as f
o
llo
ws
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
E
n
h
a
n
ci
n
g
s
leep
d
is
o
r
d
er d
ia
g
n
o
s
is
th
r
o
u
g
h
e
n
s
emb
le
ML
mo
d
els:
… (
S
a
tya
p
r
a
ka
s
h
S
w
a
in
)
33
4.
M
E
T
H
O
DO
L
O
G
Y
4
.
1
.
F
l
o
w
o
f
wo
r
k
Fig
u
r
e
2
s
h
o
w
s
th
e
p
r
o
p
o
s
ed
wo
r
k
m
o
d
el
to
ac
h
iev
e
th
e
tar
g
et.
T
h
e
r
aw
d
ataset
is
p
r
ep
r
o
ce
s
s
ed
b
y
u
s
in
g
d
if
f
e
r
en
t
s
tatis
tical
m
eth
o
d
s
lik
e
m
ea
n
,
m
o
d
e
,
an
d
m
ed
ian
to
r
ed
u
ce
em
p
ty
,
n
u
ll
,
o
r
ir
r
elev
an
t
d
ata
item
s
in
th
e
d
ataset.
T
h
en
PC
A
tech
n
iq
u
es
is
u
s
ed
to
r
ed
u
c
e
th
e
d
im
en
s
io
n
o
f
t
h
e
d
atase
t.
7
0
%
d
ata
in
t
h
e
d
ataset
is
u
s
ed
f
o
r
m
ac
h
in
e
tr
a
in
in
g
,
2
0
%
is
f
o
r
test
in
g
,
an
d
1
0
%
is
u
s
ed
f
o
r
v
alid
atio
n
wo
r
k
.
Fo
u
r
b
ase
-
lev
el
ML
v
o
tin
g
m
o
d
els
wer
e
u
s
ed
f
o
r
tr
ain
i
n
g
p
u
r
p
o
s
e
an
d
f
iv
e
ty
p
e
b
o
o
s
tin
g
m
o
d
els
we
r
e
u
s
ed
to
en
h
an
ce
th
e
ac
cu
r
ac
y
r
esu
lts
o
f
t
h
e
b
ase
v
o
tin
g
m
o
d
els.
Fig
u
r
e
2
.
Pro
p
o
s
ed
m
o
d
el
o
f
wo
r
k
4
.
2
.
Da
t
a
s
et
4
.
2
.
1
.
Da
t
a
co
llect
io
n
A
r
ea
l
tim
e
d
ata
is
g
ath
er
ed
f
r
o
m
4
0
0
p
atien
ts
to
cr
ea
te
a
co
m
p
lete
d
ataset.
T
h
is
wo
r
k
d
ataset
co
n
tain
s
a
wid
e
r
an
g
e
o
f
p
ati
en
t
-
r
elate
d
d
ata
th
at
was
g
ath
er
ed
u
s
in
g
b
o
t
h
m
o
d
e
r
n
I
o
T
d
ev
ices
an
d
m
an
u
a
l
r
ec
o
r
d
in
g
tech
n
iq
u
es.
T
h
e
d
at
aset
in
clu
d
es
cr
u
cial
d
em
o
g
r
a
p
h
ic
in
f
o
r
m
atio
n
lik
e
g
en
d
er
an
d
ag
e
,
as
well
a
s
im
p
o
r
tan
t
h
ea
lth
-
r
elate
d
f
ac
to
r
s
lik
e
s
leep
d
u
r
atio
n
an
d
q
u
ality
,
p
h
y
s
ical
ac
tiv
ity
lev
els,
s
tr
ess
lev
els,
b
o
d
y
m
ass
in
d
ex
(
B
MI
)
ca
te
g
o
r
y
,
b
lo
o
d
p
r
ess
u
r
e,
a
n
d
h
ea
r
t
r
ate.
T
h
e
i
n
co
r
p
o
r
atio
n
o
f
I
o
T
d
e
v
ices
im
p
r
o
v
e
d
th
e
ac
cu
r
ac
y
a
n
d
b
r
ea
d
th
o
f
o
u
r
r
esear
ch
f
in
d
in
g
s
b
y
en
ab
lin
g
u
s
to
ac
q
u
ir
e
p
r
ec
is
e
an
d
r
ea
l
-
tim
e
m
ea
s
u
r
em
en
ts
.
W
ith
th
e
u
s
e
o
f
th
is
co
m
p
r
eh
en
s
iv
e
d
ataset,
wh
ich
s
er
v
es
as
th
e
b
asi
s
f
o
r
o
u
r
r
esear
ch
,
we
wer
e
ab
le
to
ex
am
in
e
an
d
ev
alu
ate
n
u
m
er
o
u
s
f
ac
ets
o
f
p
atien
t
h
ea
lth
an
d
well
-
b
ein
g
an
d
o
f
f
er
in
s
ig
h
tf
u
l
co
n
tr
ib
u
tio
n
s
to
th
e
f
ield
s
o
f
m
e
d
icin
e
an
d
h
ea
lth
ca
r
e.
4
.
2
.
2
.
Da
t
a
s
et
cle
a
nin
g
T
h
er
e
is
a
ch
an
ce
o
f
ir
r
ele
v
a
n
t
d
ata
in
th
e
r
aw
d
ataset
,
wh
ich
m
ay
ca
u
s
e
wr
o
n
g
m
ac
h
i
n
e
tr
ain
in
g
.
Dif
f
er
en
t
s
tatis
tical
m
eth
o
d
s
l
ik
e
m
ea
n
,
m
o
d
e
,
an
d
m
ed
ian
ar
e
u
s
ed
to
clea
n
th
e
d
ataset
.
T
h
ese
tech
n
iq
u
es
h
elp
to
r
e
d
u
ce
v
al
u
es lik
e
em
p
ty
ce
lls
,
g
ar
b
ag
e
v
alu
e
s
,
an
d
n
o
n
-
r
elate
d
v
alu
es.
4
.
2
.
3
.
P
rincipa
l
co
m
po
nent
a
na
ly
s
is
A
d
im
en
s
io
n
ality
r
e
d
u
ctio
n
te
ch
n
iq
u
e
u
s
ed
in
b
o
th
d
ata
a
n
a
ly
s
is
an
d
ML
is
PC
A
[
1
8
]
.
I
ts
m
ain
g
o
al
is
to
p
r
eser
v
e
th
e
m
o
s
t
im
p
o
r
tan
t
in
f
o
r
m
atio
n
wh
ile
r
e
d
u
c
in
g
th
e
co
m
p
lex
ity
p
r
esen
t
in
h
ig
h
-
d
im
en
s
io
n
al
d
atasets
.
So
m
e
co
m
m
o
n
s
ig
n
i
f
ican
ce
s
o
f
PC
A
ar
e
ap
p
lied
o
v
er
o
u
r
c
o
llected
d
ataset
,
s
u
ch
as:
d
im
en
s
io
n
ality
r
ed
u
ctio
n
,
v
is
u
aliza
tio
n
,
n
o
is
e
r
ed
u
ctio
n
,
f
ea
tu
r
e
en
g
i
n
ee
r
in
g
,
m
u
lti
-
co
l
lin
ea
r
ity
m
itig
atio
n
,
d
ata
co
m
p
r
ess
io
n
,
an
d
an
o
m
aly
d
etec
tio
n
.
A
u
s
ef
u
l
m
eth
o
d
with
m
a
n
y
a
p
p
licatio
n
s
in
n
u
m
er
o
u
s
s
tu
d
y
d
o
m
ain
s
is
PC
A.
I
t
is
a
cr
u
ci
al
to
o
l
f
o
r
d
er
iv
in
g
v
alu
ab
le
in
s
ig
h
ts
f
r
o
m
co
m
p
licated
d
atasets
b
ec
au
s
e
o
f
its
ca
p
ac
ity
to
r
ed
u
ce
d
im
en
s
io
n
ality
,
in
cr
ea
s
e
v
is
u
aliza
tio
n
,
an
d
i
m
p
r
o
v
e
d
ata
q
u
ality
.
I
t
en
a
b
les
th
em
to
d
etec
t
h
id
d
en
p
atter
n
s
,
f
ilter
o
u
t
b
ac
k
g
r
o
u
n
d
n
o
is
e
,
an
d
o
p
tim
ize
s
u
b
s
eq
u
en
t
s
tep
s
.
Fig
u
r
e
3
d
is
p
lay
s
a
g
r
ap
h
ic
d
e
p
ictio
n
o
f
o
u
r
o
b
tain
e
d
d
ataset
af
ter
PC
A
was a
p
p
lied
to
it.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
c
h
20
2
6
:
29
-
41
34
Fig
u
r
e
3
.
Gr
a
p
h
ical
r
ep
r
esen
tatio
n
o
f
PC
A
o
v
er
c
o
llected
d
at
aset
5.
SI
M
UL
A
T
I
O
N
A
ND
RE
SU
L
T
AN
AL
Y
SI
S
5
.
1
.
B
a
s
e
mo
del
B
ase
m
o
d
el
(
v
o
tin
g
)
class
if
ier
s
[
1
9
]
–
[
2
3
]
ar
e
ML
m
o
d
els
th
at
an
ticip
ate
an
o
u
tp
u
t b
ased
o
n
th
e
class
th
at
h
as
th
e
h
ig
h
est
p
o
s
s
ib
ilit
y
o
f
b
ein
g
th
e
o
u
tp
u
t
,
as
s
h
o
wn
in
Fig
u
r
e
4
.
T
h
ese
ar
e
g
a
th
er
ed
ex
p
e
r
tis
e
b
y
tr
ain
in
g
o
n
a
c
o
llectio
n
o
f
v
ar
io
u
s
m
o
d
els.
I
t
s
im
p
ly
av
e
r
ag
es
th
e
o
u
tco
m
es
o
f
ea
c
h
class
if
ier
th
at
wa
s
s
u
b
m
itted
in
to
th
e
v
o
tin
g
class
if
ier
in
o
r
d
e
r
to
f
o
r
ec
ast th
e
o
u
tp
u
t c
lass
b
ased
o
n
th
e
h
ig
h
est
m
ajo
r
ity
o
f
v
o
tes.
I
n
s
tead
o
f
cr
ea
tin
g
in
d
iv
id
u
al
s
p
ec
ialized
m
o
d
els
an
d
ev
al
u
atin
g
th
ei
r
co
r
r
ec
tn
ess
,
th
e
id
ea
is
to
d
ev
elo
p
a
s
in
g
le
m
o
d
el
th
at
lear
n
s
f
r
o
m
m
u
ltip
le
m
o
d
els
an
d
p
r
ed
icts
o
u
tp
u
t
b
ased
o
n
th
e
ag
g
r
eg
ate
m
ajo
r
ity
o
f
v
o
tin
g
f
o
r
ea
ch
o
u
tp
u
t
class
[
3
]
,
[
2
4
]
,
[
2
5
]
.
T
h
e
f
i
n
al
p
r
e
d
ictio
n
“y
”
is
d
eter
m
in
ed
b
y
co
u
n
tin
g
t
h
e
v
o
tes
f
r
o
m
ea
ch
b
ase
m
o
d
el
as in
(
1
)
.
=
(
∑
1
[
ℎ
(
)
−
]
=
1
)
(
1
)
L
et
N
i
s
n
u
m
b
er
o
f
b
ase
m
o
d
els,
h
i(
x)
be
th
e
p
r
ed
ictio
n
m
a
d
e
b
y
th
e
ith
b
ase
m
o
d
el
f
o
r
in
p
u
t
x
,
an
d
y
b
e
th
e
f
in
al
en
s
em
b
le
p
r
e
d
ictio
n
.
Fig
u
r
e
4
.
Vo
tin
g
m
o
d
el
m
ad
e
u
s
in
g
4
d
if
f
er
en
t
m
l m
o
d
els
H
e
r
e
,
f
o
u
r
ML
m
o
d
e
l
s
l
ik
e
D
T
,
S
V
M
,
N
B
,
a
n
d
R
F
a
r
e
s
e
l
e
c
t
e
d
f
o
r
e
n
s
em
b
l
e
l
e
ar
n
in
g
v
o
t
i
n
g
m
o
d
e
l
c
r
e
a
t
i
o
n
:
i)
DT
:
a
tr
ee
-
lik
e
s
tr
u
ctu
r
e
ca
lle
d
a
d
ec
is
io
n
tr
ee
is
em
p
lo
y
ed
f
o
r
class
if
icatio
n
an
d
r
eg
r
ess
io
n
ap
p
licatio
n
s
.
I
n
o
r
d
er
to
d
ec
id
e
o
r
p
r
e
d
ict
s
o
m
eth
in
g
,
it
d
iv
id
es
th
e
d
ata
in
to
b
r
an
c
h
es
d
e
p
en
d
in
g
o
n
f
ea
tu
r
e
v
al
u
es.
E
v
er
y
n
o
d
e
an
d
b
r
a
n
ch
i
n
t
h
e
tr
ee
in
d
icate
s
d
if
f
er
e
n
t
f
e
atu
r
es
an
d
o
p
tio
n
s
,
r
esp
ec
tiv
ely
.
DT
s
ar
e
r
en
o
wn
ed
f
o
r
b
ein
g
ea
s
y
to
u
n
d
er
s
tan
d
an
d
co
m
p
r
eh
en
d
.
ii)
SVM:
a
p
o
ten
t c
lass
if
ier
th
at
l
o
ca
tes a
h
y
p
er
p
lan
e
i
n
h
ig
h
-
d
i
m
en
s
io
n
al
s
p
ac
e
to
d
iv
id
e
v
a
r
io
u
s
class
es.
I
t
is
ef
f
icien
t
f
o
r
b
o
th
lin
ea
r
a
n
d
n
o
n
-
lin
ea
r
class
if
icatio
n
jo
b
s
s
in
ce
it
s
ee
k
s
to
m
a
x
im
is
e
th
e
m
ar
g
i
n
b
etwe
en
class
es.
SVM
i
s
r
en
o
wn
ed
f
o
r
its
d
ep
en
d
a
b
ilit
y
an
d
ad
ap
tab
ilit
y
.
iii)
NB
:
b
ased
o
n
B
ay
es'
th
eo
r
em
,
NB
is
a
p
r
o
b
a
b
ilis
tic
clas
s
if
icatio
n
alg
o
r
ith
m
.
C
alcu
latio
n
s
ar
e
m
ad
e
s
im
p
ler
b
y
th
e
ass
u
m
p
tio
n
th
at
ch
ar
ac
ter
is
tics
ar
e
in
d
ep
e
n
d
en
t.
NB
is
well
-
r
en
o
wn
ed
f
o
r
b
ein
g
q
u
ick
an
d
ef
f
ec
tiv
e
an
d
is
f
r
e
q
u
en
tly
u
s
ed
in
tex
t c
lass
if
icatio
n
an
d
s
p
am
d
etec
tio
n
.
iv
)
R
F:
a
g
r
o
u
p
o
f
DT
s
is
ca
lled
a
r
an
d
o
m
.
T
o
less
en
o
v
e
r
f
itti
n
g
an
d
in
c
r
ea
s
e
p
r
ed
ictio
n
ac
cu
r
ac
y
,
it
m
ix
es
th
e
p
r
ed
ictio
n
s
o
f
v
ar
io
u
s
tr
e
es.
Sin
ce
R
F
is
s
o
r
esil
ien
t
a
n
d
ad
a
p
tab
le,
it
ca
n
b
e
u
s
ed
f
o
r
a
v
ar
iety
o
f
class
if
icatio
n
an
d
r
eg
r
ess
io
n
a
p
p
licatio
n
s
.
I
t is r
en
o
w
n
ed
f
o
r
h
an
d
lin
g
in
tr
icate
d
ata
r
elatio
n
s
h
ip
s
well.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
E
n
h
a
n
ci
n
g
s
leep
d
is
o
r
d
er d
ia
g
n
o
s
is
th
r
o
u
g
h
e
n
s
emb
le
ML
mo
d
els:
… (
S
a
tya
p
r
a
ka
s
h
S
w
a
in
)
35
Her
e
,
th
e
m
ain
d
ataset
,
wh
ich
co
n
tain
s
4
0
0
s
am
p
les
,
is
d
iv
id
ed
in
to
4
eq
u
al
n
u
m
b
e
r
s
o
f
in
s
tan
ce
s
,
an
d
ap
p
l
y
v
o
tin
g
m
o
d
el
is
ap
p
lied
o
n
ea
ch
i
n
s
tan
ce
.
B
ased
o
n
th
e
tak
en
d
ataset
an
d
its
Py
th
o
n
s
im
u
latio
n
,
it
is
clea
r
ly
s
ee
n
th
at
R
F
g
iv
es
th
e
b
est
ac
cu
r
ac
y
am
o
n
g
th
e
o
t
h
er
th
r
ee
m
o
d
els
,
wh
ic
h
is
s
h
o
wn
in
T
a
b
le
1
a
n
d
its
g
r
ap
h
ical
r
ep
r
esen
tatio
n
in
Fig
u
r
e
5
.
T
h
en
,
th
e
co
m
p
a
r
is
o
n
s
b
etwe
en
f
o
u
r
m
o
d
els
ar
e
p
r
o
v
id
e
d
b
y
tak
in
g
th
e
p
ar
am
eter
s
lik
e
MA
E
,
R
M
SE,
R
AE
,
an
d
R
R
SE
f
o
r
b
etter
u
n
d
e
r
s
tan
d
in
g
.
T
ab
le
1
.
Acc
u
r
ac
y
tab
le
o
f
v
o
tin
g
m
o
d
el
o
f
DT
,
SVM,
NB
,
an
d
R
F o
v
er
4
eq
u
al
ly
d
iv
id
e
d
d
atasets
D
a
t
a
s
e
t
i
n
st
a
n
c
e
S
V
M
DT
NB
RF
1
-
1
0
0
9
0
.
2
9
2
.
3
9
8
8
.
0
4
8
9
.
1
3
1
0
0
-
200
8
8
.
0
9
8
6
.
9
8
5
.
7
9
0
.
4
7
2
0
0
-
300
88
85
85
92
3
0
0
-
400
9
1
.
8
3
9
1
.
8
3
9
0
.
8
1
9
2
.
8
5
Fig
u
r
e
5
.
Gr
a
p
h
ical
r
ep
r
esen
tatio
n
o
f
ac
c
u
r
ac
y
o
f
DT
,
SVM,
NB
,
an
d
R
F
5
.
1
.
1
.
M
ea
n
a
bs
o
lute
er
ro
r
T
h
e
av
er
a
g
e
s
ize
o
f
t
h
e
er
r
o
r
s
b
etwe
en
th
e
ex
p
e
cted
an
d
a
ctu
al
v
alu
es
is
m
ea
s
u
r
ed
s
tatis
tically
as
MA
E
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
ML
m
o
d
els
is
ev
alu
ated
u
s
in
g
th
e
MA
E
,
a
p
o
p
u
lar
m
ea
s
u
r
e
o
f
f
o
r
ec
ast
er
r
o
r
in
tim
e
s
er
ies
r
esear
ch
.
I
t
is
a
p
ar
ticu
lar
ly
h
el
p
f
u
l
in
d
icato
r
f
o
r
ass
ess
in
g
co
n
tin
u
o
u
s
v
alu
e
p
r
ed
ictio
n
m
o
d
els.
T
h
e
MA
E
v
alu
e
p
r
esen
ted
i
n
T
ab
le
2
a
n
d
th
e
g
r
ap
h
ic
r
ep
r
esen
tatio
n
in
Fig
u
r
e
6
w
h
en
ev
alu
atin
g
th
e
p
er
f
o
r
m
an
ce
o
f
o
u
r
m
o
d
el.
T
h
e
f
o
r
m
u
la
f
o
r
MA
E
is
(
2
)
.
=
(
1
/
)
(
=
1
)
|
–
ŷ
|
(
2
)
W
h
er
e
n
is
s
am
p
le
s
ize,
is
ac
tu
al
v
alu
e
at
o
b
s
er
v
atio
n
i
,
an
d
ŷ
is
th
e
p
r
ed
icted
v
alu
e
at
o
b
s
er
v
atio
n
i
.
T
ab
le
2
.
MA
E
v
al
u
e
o
f
v
o
tin
g
m
o
d
el
o
f
DT
,
SVM,
NB
,
an
d
R
F
D
a
t
a
s
e
t
i
n
st
a
n
c
e
S
V
M
DT
NB
RF
1
-
1
0
0
0
.
0
6
0
.
1
0
.
0
8
0
.
0
9
1
0
0
-
200
0
.
0
7
0
.
1
4
0
.
0
9
0
.
0
9
2
0
0
-
300
0
.
0
8
0
.
1
9
0
.
0
9
0
.
0
7
3
0
0
-
400
0
.
0
5
0
.
1
5
0
.
0
6
0
.
0
6
Fig
u
r
e
6
.
Gr
a
p
h
ical
r
ep
r
esen
tatio
n
o
f
MA
E
d
ata
o
v
e
r
DT
,
S
VM
,
NB
,
an
d
R
F
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
c
h
20
2
6
:
29
-
41
36
5
.
1
.
2
.
Ro
o
t
m
e
a
n sq
ua
re
d e
r
ro
r
T
h
e
s
tatis
tical
m
ea
s
u
r
em
en
t
o
f
th
e
d
is
cr
ep
a
n
cy
b
etwe
en
ex
p
ec
ted
an
d
ac
t
u
al
v
alu
es
is
ca
lled
R
MSE
.
I
t
is
ca
lcu
lated
b
y
u
s
in
g
th
e
MSE
s
q
u
ar
e
r
o
o
t
as
in
(
3
)
.
I
n
m
o
d
el
p
er
f
o
r
m
an
ce
ev
alu
atio
n
,
th
e
R
MSE
v
alu
e
is
s
h
o
wn
in
T
ab
le
3
,
an
d
a
g
r
ap
h
ical
r
ep
r
esen
tatio
n
is
s
h
o
wn
in
Fig
u
r
e
7
.
=
(
)
(
3
)
T
ab
le
3
.
R
MSE
v
alu
e
o
f
v
o
tin
g
m
o
d
el
o
f
DT
,
SVM,
NB
,
an
d
R
F
D
a
t
a
s
e
t
i
n
st
a
n
c
e
S
V
M
DT
NB
RF
1
-
1
0
0
0
.
2
5
0
.
2
0
.
2
7
0
.
2
3
1
0
0
-
200
0
.
2
8
0
.
2
0
.
2
9
0
.
2
5
2
0
0
-
300
0
.
2
8
0
.
2
9
0
.
3
0
.
2
1
3
0
0
-
400
0
.
2
3
0
.
2
4
0
.
2
4
0
.
2
Fig
u
r
e
7
.
Gr
a
p
h
ical
r
ep
r
esen
tatio
n
o
f
R
MSE
d
ata
o
v
er
DT
,
S
VM
,
NB
,
an
d
R
F
m
o
d
el
s
5
.
1
.
3
.
Rela
t
iv
e
a
bs
o
lute
er
ro
r
A
s
tati
s
tical
in
d
icato
r
o
f
a
p
r
ed
ictio
n
'
s
ac
cu
r
ac
y
in
r
elatio
n
to
th
e
ac
tu
al
v
alu
e
is
th
e
R
AE
.
B
y
d
iv
id
in
g
th
e
ab
s
o
lu
te
er
r
o
r
b
y
th
e
tr
u
e
v
alu
e,
it
is
ca
lcu
lated
.
T
h
e
f
o
r
m
u
la
f
o
r
R
AE
is
(
4
)
.
Fo
r
o
u
r
m
o
d
el
p
er
f
o
r
m
an
ce
ev
alu
atio
n
,
th
e
R
AE
v
alu
e
is
s
h
o
wn
in
T
ab
le
4
,
an
d
a
g
r
ap
h
ical
r
e
p
r
esen
tatio
n
is
s
h
o
w
n
in
Fig
u
r
e
8
.
=
|
(
−
ŷ
)
|
/
(
4
)
W
h
er
e
is
th
e
ac
tu
al
v
alu
e
at
o
b
s
er
v
atio
n
i
,
ŷ
is
th
e
p
r
ed
icted
v
alu
e
at
o
b
s
er
v
atio
n
i
.
T
ab
le
4
.
R
AE
v
alu
e
o
f
v
o
tin
g
m
o
d
el
DT
,
SVM,
NB
,
an
d
R
F
D
a
t
a
s
e
t
i
n
st
a
n
c
e
S
V
M
DT
NB
RF
1
-
1
0
0
3
6
.
6
1
5
8
.
2
48
5
1
.
5
1
1
0
0
-
200
5
0
.
6
9
5
.
5
6
0
.
4
6
6
1
.
8
2
0
0
-
300
2
1
.
4
5
2
.
5
3
2
6
.
1
9
2
1
.
1
4
3
0
0
-
400
1
4
.
4
4
2
.
0
5
1
6
.
2
5
1
7
.
4
7
Fig
u
r
e
8
.
Gr
a
p
h
ical
r
ep
r
esen
tatio
n
o
f
R
AE
v
alu
e
o
v
er
DT
,
SVM,
NB
,
an
d
R
F
m
o
d
el
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ad
v
Ap
p
l Sci
I
SS
N:
2252
-
8
8
1
4
E
n
h
a
n
ci
n
g
s
leep
d
is
o
r
d
er d
ia
g
n
o
s
is
th
r
o
u
g
h
e
n
s
emb
le
ML
mo
d
els:
… (
S
a
tya
p
r
a
ka
s
h
S
w
a
in
)
37
5
.
1
.
4
.
Ro
o
t
re
la
t
iv
e
s
qu
a
re
d e
rr
o
r
A
s
tatis
tica
l
in
d
icato
r
o
f
a
p
r
e
d
ictio
n
'
s
ac
cu
r
ac
y
in
r
elatio
n
t
o
th
e
m
ea
n
v
al
u
e
o
f
th
e
ac
t
u
al
d
ata
is
th
e
R
R
SE.
I
t
is
ca
lcu
lated
b
y
d
iv
i
d
in
g
th
e
ac
tu
al
d
ata'
s
m
ea
n
v
alu
e
b
y
th
e
R
MSE
.
T
h
e
f
o
r
m
u
l
a
f
o
r
R
R
SE
is
(
5
)
.
I
n
o
u
r
m
o
d
el
p
er
f
o
r
m
a
n
ce
ev
alu
atio
n
,
th
e
R
R
SE
v
alu
e
s
h
o
wn
in
T
ab
le
5
an
d
its
g
r
ap
h
ic
al
r
ep
r
esen
tatio
n
ar
e
s
h
o
wn
in
Fig
u
r
e
9
.
=
/
(
)
(
5
)
T
ab
le
5
.
R
R
SE
v
alu
e
o
f
v
o
tin
g
m
o
d
el
o
v
er
DT
,
SVM,
NB
,
an
d
R
F
D
a
t
a
s
e
t
i
n
st
a
n
c
e
S
V
M
DT
NB
RF
1
-
1
0
0
8
7
.
7
7
3
.
7
9
5
.
4
4
8
2
.
4
2
1
0
0
-
200
1
0
3
.
7
1
0
2
.
6
6
1
0
9
.
3
1
9
2
.
6
2
2
0
0
-
300
6
5
.
5
9
6
9
.
3
8
7
1
.
8
1
4
9
.
4
8
3
0
0
-
400
5
3
.
8
7
5
7
.
5
1
5
6
.
7
4
7
.
3
6
Fig
u
r
e
9
.
Gr
a
p
h
ical
r
ep
r
esen
tatio
n
o
f
R
R
SE
v
alu
e
o
v
e
r
DT
,
SVM,
NB
,
an
d
R
F
m
o
d
el
s
5
.
2
.
B
o
o
s
t
ing
m
o
del
B
o
o
s
t
i
n
g
i
s
a
w
e
l
l
-
k
n
o
w
n
en
s
e
m
b
l
e
ML
te
c
h
n
i
q
u
e
t
h
a
t
c
r
e
a
t
e
s
a
r
o
b
u
s
t
l
e
ar
n
e
r
b
y
c
o
m
b
i
n
in
g
p
r
ed
i
c
t
i
o
n
s
f
r
o
m
s
e
v
er
a
l
w
e
ak
le
a
r
n
e
r
s
,
f
r
eq
u
en
t
l
y
DT
s
o
r
s
h
a
l
l
o
w
m
o
d
e
l
s
.
I
t
s
m
a
i
n
g
o
a
l
i
s
t
o
i
n
c
r
e
a
s
e
m
o
d
e
l
ac
c
u
r
a
c
y
b
y
g
iv
i
n
g
c
as
e
s
t
h
a
t
w
e
r
e
p
r
ev
i
o
u
s
l
y
e
r
r
o
n
e
o
u
s
l
y
c
l
a
s
s
i
f
i
e
d
m
o
r
e
w
e
ig
h
t
.
T
h
e
ap
p
r
o
a
ch
i
s
d
e
v
e
lo
p
e
d
in
i
t
e
r
a
t
iv
e
s
t
ep
s
,
w
i
t
h
a
s
t
r
o
n
g
e
m
p
h
a
s
i
s
o
n
co
r
r
ec
t
i
n
g
t
h
e
m
i
s
t
a
k
e
s
m
a
d
e
b
y
e
ar
l
i
e
r
m
o
d
e
l
s
,
a
s
s
h
o
w
n
i
n
F
i
g
u
r
e
1
0
.
I
ts
ch
ar
ac
ter
is
tics
in
clu
d
e
ad
ap
tab
ilit
y
,
s
eq
u
en
tial
b
u
ild
in
g
,
th
e
elim
in
atio
n
o
f
b
ias,
an
d
v
u
ln
er
ab
ilit
y
to
p
o
ten
tial
o
v
er
-
f
itti
n
g
in
v
a
r
ian
ce
f
in
d
in
g
.
T
o
r
ed
u
ce
th
e
r
is
k
o
f
o
v
er
-
f
itti
n
g
,
m
eth
o
d
s
lik
e
ea
r
ly
s
to
p
p
in
g
an
d
r
estrictin
g
th
e
d
ep
t
h
o
f
p
o
o
r
lear
n
er
s
ar
e
f
r
eq
u
e
n
tly
u
s
ed
.
Her
e,
we
o
b
s
er
v
e
d
th
at
R
F
d
eliv
er
s
th
e
b
est
p
r
ed
ictio
n
ac
c
u
r
ac
y
w
h
en
co
m
p
ar
ed
to
o
th
er
m
o
d
els
b
y
u
s
in
g
DT
,
SVM
,
an
d
NB
.
I
n
o
r
d
er
to
co
m
p
ar
e,
we
u
s
e
v
ar
io
u
s
b
o
o
s
tin
g
m
o
d
els,
s
u
ch
as
Ad
aBo
o
s
t
,
g
r
ad
ien
t
b
o
o
s
tin
g
,
XGBo
o
s
t
,
L
ig
h
tGB
M
,
an
d
C
atB
o
o
s
t
f
o
r
p
r
ed
ictio
n
o
v
er
th
e
d
ataset.
Af
ter
ap
p
ly
in
g
s
ev
e
r
al
b
o
o
s
t
in
g
m
o
d
els
o
v
e
r
R
F,
we
f
o
u
n
d
th
e
p
r
ed
ictio
n
ac
cu
r
ac
y
,
w
h
ich
is
s
to
r
e
d
in
T
ab
le
6
,
an
d
its
g
r
ap
h
ical
r
ep
r
esen
tatio
n
is
s
h
o
wn
i
n
F
ig
u
r
e
1
1
.
T
a
b
le
7
h
ig
h
lig
h
ts
th
e
in
n
o
v
atio
n
s
an
d
en
h
an
ce
m
e
n
ts
m
ad
e
in
th
is
r
esear
ch
an
d
o
f
f
e
r
s
a
clea
r
co
m
p
ar
is
o
n
b
etwe
en
th
is
wo
r
k
an
d
t
h
e
ex
is
tin
g
r
ela
ted
wo
r
k
s
.
Fig
u
r
e
1
0
.
B
o
o
s
tin
g
m
o
d
el
ar
c
h
itectu
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
8
1
4
I
n
t J Ad
v
Ap
p
l Sci
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
c
h
20
2
6
:
29
-
41
38
T
ab
le
6
.
Acc
u
r
ac
y
tab
le
a
f
ter
ap
p
ly
in
g
b
o
o
s
tin
g
m
o
d
els o
v
e
r
RF
M
o
d
e
l
s
A
c
c
u
r
a
c
y
(
%)
A
d
a
B
o
o
st
9
0
.
5
6
G
r
a
d
i
e
n
t
b
o
o
s
t
9
6
.
0
1
X
G
B
o
o
st
9
2
.
4
7
C
a
t
B
o
o
s
t
9
0
.
5
6
Li
g
h
t
G
B
M
8
5
.
2
0
Fig
u
r
e
1
1
.
Gr
ap
h
ical
r
e
p
r
esen
t
atio
n
ac
cu
r
ac
y
o
f
b
o
o
s
tin
g
m
o
d
els
T
ab
le
7
.
C
o
m
p
a
r
is
o
n
s
b
etwe
en
th
is
wo
r
k
with
ex
is
tin
g
w
o
r
k
s
S
t
u
d
y
H
e
a
l
t
h
c
o
n
d
i
t
i
o
n
M
e
t
h
o
d
o
l
o
g
y
D
a
t
a
s
e
t
A
c
c
u
r
a
c
y
(
%)
N
o
t
a
b
l
e
f
i
n
d
i
n
g
s
K
o
r
k
a
l
a
i
n
e
n
e
t
a
l
.
[
3
]
O
S
A
DL
a
n
d
d
i
v
e
r
se
e
p
o
c
h
d
u
r
a
t
i
o
n
s
4
4
6
O
S
A
p
a
t
i
e
n
t
s
N
/
A
R
e
v
e
a
l
e
d
u
n
d
e
r
e
s
t
i
mat
i
o
n
o
f
sl
e
e
p
f
r
a
g
me
n
t
a
t
i
o
n
i
n
s
e
v
e
r
e
O
S
A
a
n
d
e
mp
h
a
s
i
z
i
n
g
t
h
e
n
e
e
d
f
o
r
d
e
t
a
i
l
e
d
sl
e
e
p
a
r
c
h
i
t
e
c
t
u
r
e
a
n
a
l
y
si
s
.
K
u
o
a
n
d
C
h
e
n
[
5
]
I
n
so
mn
i
a
S
V
M
a
n
d
R
C
M
S
E
a
n
a
l
y
si
s
3
2
su
b
j
e
c
t
s
8
9
.
3
1
U
t
i
l
i
z
e
s
si
n
g
l
e
-
c
h
a
n
n
e
l
sl
e
e
p
EO
G
f
o
r
sh
o
r
t
-
d
u
r
a
t
i
o
n
i
n
s
o
m
n
i
a
d
e
t
e
c
t
i
o
n
a
n
d
su
i
t
a
b
l
e
f
o
r
p
o
r
t
a
b
l
e
P
S
G
sy
st
e
ms.
Zu
l
f
i
k
e
r
e
t
a
l
.
[
7
]
I
n
so
mn
i
a
M
u
l
t
i
l
a
y
e
r
st
a
c
k
i
n
g
m
o
d
e
l
a
n
d
P
C
A
S
o
c
i
o
-
d
e
m
o
g
r
a
p
h
i
c
d
a
t
a
se
t
8
8
.
6
0
En
se
mb
l
e
m
o
d
e
l
s
u
r
p
a
sses
l
e
a
d
i
n
g
c
l
a
ssi
f
i
e
r
s
a
n
d
p
r
o
v
i
d
i
n
g
su
p
e
r
i
o
r
p
e
r
f
o
r
m
a
n
c
e
i
n
p
r
e
d
i
c
t
i
n
g
i
n
s
o
m
n
i
a
b
a
s
e
d
o
n
s
o
c
i
o
-
d
e
m
o
g
r
a
p
h
i
c
f
a
c
t
o
r
s.
A
l
g
h
w
i
r
i
e
t
a
l
.
[
8
]
S
l
e
e
p
q
u
a
l
i
t
y
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
a
n
d
RF
1
,
6
0
0
u
n
i
v
e
r
s
i
t
y
st
u
d
e
n
t
s
74
I
d
e
n
t
i
f
i
e
d
k
e
y
f
a
c
t
o
r
s
a
f
f
e
c
t
i
n
g
sl
e
e
p
q
u
a
l
i
t
y
,
w
i
t
h
RF
s
h
o
w
i
n
g
t
h
e
h
i
g
h
e
st
p
r
e
d
i
c
t
i
o
n
a
c
c
u
r
a
c
y
.
Th
i
s
st
u
d
y
S
l
e
e
p
d
i
s
o
r
d
e
r
S
t
a
t
i
st
i
c
a
l
met
h
o
d
s
f
o
r
d
a
t
a
c
l
e
a
n
i
n
g
,
P
C
A
,
B
a
se
ML
m
o
d
e
l
s
,
a
n
d
M
L
b
o
o
s
t
i
n
g
mo
d
e
l
s
R
e
a
l
-
t
i
me
d
a
t
a
se
t
o
f
4
0
0
p
a
t
i
e
n
t
s
9
6
.
0
1
N
o
v
e
l
i
n
t
e
g
r
a
t
i
o
n
o
f
v
a
r
i
o
u
s
M
L
t
e
c
h
n
i
q
u
e
s
f
o
r
e
n
h
a
n
c
e
d
d
i
a
g
n
o
st
i
c
a
c
c
u
r
a
c
y
o
n
s
l
e
e
p
d
i
s
o
r
d
e
r
p
r
e
d
i
c
t
i
o
n
,
w
i
t
h
a
f
o
c
u
s
o
n
e
a
r
l
y
a
n
d
r
e
l
i
a
b
l
e
d
e
t
e
c
t
i
o
n
.
6.
CO
NCLU
SI
O
N
I
n
o
r
d
er
to
tack
le
th
e
d
if
f
icu
lt
task
o
f
d
etec
tin
g
s
leep
in
g
d
is
o
r
d
er
s
,
th
is
s
tu
d
y
h
as
ef
f
ec
tiv
el
y
ap
p
lie
d
en
s
em
b
le
lear
n
in
g
alg
o
r
ith
m
s
,
s
p
ec
if
ically
v
o
tin
g
an
d
b
o
o
s
tin
g
,
o
n
th
e
d
ata
d
ataset
g
ath
e
r
ed
b
y
I
o
T
s
en
s
o
r
s
an
d
m
an
u
ally
.
T
h
e
s
tr
ateg
y
o
f
th
is
s
tu
d
y
in
v
o
lv
ed
co
m
b
in
in
g
f
o
u
r
d
if
f
er
en
t
b
ase
m
o
d
els
in
to
a
v
o
tin
g
en
s
em
b
le:
DT
,
SVM,
N
B
,
a
n
d
R
F.
Usi
n
g
a
d
ataset
m
ad
e
u
p
o
f
4
0
0
p
atien
t
r
ec
o
r
d
s
,
th
ese
b
ase
m
o
d
els
d
em
o
n
s
tr
ated
ef
f
icien
cy
in
im
p
r
o
v
in
g
th
e
p
r
e
d
ictab
ilit
y
o
f
s
leep
in
g
d
is
o
r
d
er
s
.
T
h
e
boo
s
tin
g
ap
p
r
o
ac
h
,
wh
ich
b
u
ilt
o
n
th
e
ad
v
a
n
tag
es
o
f
th
e
b
ase
m
o
d
els
an
d
ad
d
ed
f
i
v
e
b
o
o
s
tin
g
-
lear
n
e
r
s
:
Ad
aBo
o
s
t
,
g
r
ad
ien
t
b
o
o
s
tin
g
,
XGBo
o
s
t
,
L
ig
h
tGB
M
,
an
d
C
a
tB
o
o
s
t
,
was
r
esp
o
n
s
ib
le
f
o
r
t
h
e
r
ea
l
b
r
ea
k
th
r
o
u
g
h
.
Gr
ad
ien
t
b
o
o
s
tin
g
ex
ce
lled
am
o
n
g
th
e
m
,
attain
in
g
a
r
e
m
ar
k
ab
le
ac
cu
r
ac
y
r
ate
o
f
9
6
.
0
1
%,
d
em
o
n
s
tr
atin
g
t
h
e
ca
p
ab
ilit
y
o
f
b
o
o
s
tin
g
to
en
h
an
ce
m
o
d
el
p
er
f
o
r
m
a
n
ce
.
Fu
tu
r
e
r
esear
ch
h
as
a
n
u
m
b
er
o
f
p
o
ten
tial
d
ir
ec
tio
n
s
,
t
o
b
e
s
u
r
e.
M
o
d
el
r
o
b
u
s
tn
ess
m
ay
b
e
im
p
r
o
v
ed
b
y
en
lar
g
in
g
th
e
d
ataset
an
d
a
d
d
in
g
n
ew
ch
ar
ac
ter
is
tics
,
lik
e
s
leep
p
atter
n
s
an
d
life
s
ty
le
v
ar
iab
les.
T
h
e
ac
cu
r
ac
y
o
f
p
r
e
d
ictio
n
s
m
a
y
b
e
f
u
r
th
er
im
p
r
o
v
e
d
b
y
i
n
v
esti
g
atin
g
cu
ttin
g
-
ed
g
e
DL
ar
ch
itectu
r
es to
r
ep
lace
t
h
e
n
e
u
r
al
n
etwo
r
k
co
m
p
o
n
e
n
t.
Evaluation Warning : The document was created with Spire.PDF for Python.