I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
,
p
p
.
385
~
396
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
14
.i
1
.
p
p
3
8
5
-
3
9
6
385
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
M
a
chine learning
f
o
r men
tal hea
lth
:
predi
cting
trans
itions
f
ro
m
a
ddiction to ill
nes
s
Ali A
lk
ha
zr
a
j
i
1
,
F
a
t
i
m
a
Als
a
f
i
1
,
M
o
ha
m
ed
Dbo
uk
1
,
Z
ein A
l A
bid
i
n Ib
ra
him
1,
2
,
I
ha
b Sbeity
1,
3
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
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
s,
L
e
b
a
n
e
s
e
U
n
i
v
e
r
si
t
y
,
H
a
d
a
t
C
a
mp
u
s,
B
e
i
r
u
t
,
Le
b
a
n
o
n
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
a
n
d
C
o
mm
u
n
i
c
a
t
i
o
n
s
En
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
L
e
b
a
n
e
se
I
n
t
e
r
n
a
t
i
o
n
a
l
U
n
i
v
e
r
si
t
y
,
B
e
i
r
u
t
,
L
e
b
a
n
o
n
3
D
e
p
a
r
t
me
n
t
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
s
,
L
e
b
a
n
e
se
I
n
t
e
r
n
a
t
i
o
n
a
l
U
n
i
v
e
r
si
t
y
,
B
e
i
r
u
t
,
L
e
b
a
n
o
n
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Mar
30
,
2
0
2
4
R
ev
is
ed
J
u
l
27
,
2
0
2
4
Acc
ep
ted
Au
g
30
,
2
0
2
4
Th
e
i
n
c
re
a
sin
g
p
re
v
a
len
c
e
o
f
in
fe
c
ti
o
n
-
c
a
u
sin
g
d
ise
a
se
s
d
u
e
t
o
e
n
v
i
ro
n
m
e
n
tal
fa
c
to
rs
a
n
d
l
ifes
ty
le
c
h
o
ice
s
h
a
s
stra
in
e
d
t
h
e
h
e
a
lt
h
c
a
re
sy
ste
m
,
n
e
c
e
ss
it
a
ti
n
g
a
d
v
a
n
c
e
d
tec
h
n
iq
u
e
s
to
sa
v
e
li
v
e
s.
Dise
a
se
p
re
d
ictio
n
p
la
y
s
a
c
ru
c
ial
ro
le
in
id
e
n
ti
f
y
in
g
in
d
iv
id
u
a
ls
a
t
r
isk
,
e
n
a
b
li
n
g
e
a
rl
y
trea
tme
n
t,
a
n
d
b
e
n
e
fit
in
g
g
o
v
e
r
n
m
e
n
ts
a
n
d
h
e
a
lt
h
i
n
su
ra
n
c
e
p
ro
v
i
d
e
rs.
Th
e
c
o
l
lab
o
ra
ti
o
n
b
e
twe
e
n
b
io
m
e
d
ici
n
e
a
n
d
d
a
ta
sc
ien
c
e
,
p
a
rti
c
u
larly
a
rti
ficia
l
in
telli
g
e
n
c
e
a
n
d
m
a
c
h
in
e
lea
rn
in
g
,
h
a
s
led
to
si
g
n
ifi
c
a
n
t
a
d
v
a
n
c
e
m
e
n
ts
in
t
h
is
fiel
d
.
Ho
we
v
e
r,
re
se
a
rc
h
e
rs
fa
c
e
c
h
a
ll
e
n
g
e
s
re
late
d
to
d
a
ta
a
v
a
il
a
b
il
it
y
a
n
d
q
u
a
li
t
y
.
Cli
n
ica
l
a
n
d
h
o
sp
it
a
l
d
a
ta,
c
ru
c
ial
fo
r
a
c
c
u
ra
te
p
re
d
ictio
n
s,
a
re
o
ften
c
o
n
f
i
d
e
n
ti
a
l
a
n
d
n
o
t
fre
e
l
y
a
c
c
e
ss
ib
le.
M
o
re
o
v
e
r,
h
e
a
lt
h
c
a
re
d
a
ta
is
p
re
d
o
m
in
a
n
tl
y
u
n
stru
c
t
u
re
d
,
re
q
u
iri
n
g
e
x
ten
siv
e
c
lea
n
in
g
,
p
re
p
ro
c
e
ss
in
g
,
a
n
d
la
b
e
li
n
g
.
Th
is
stu
d
y
a
ims
to
p
re
d
ict
th
e
li
k
e
li
h
o
o
d
o
f
p
a
ti
e
n
ts
tran
sit
io
n
in
g
to
m
e
n
tal
il
ln
e
ss
b
y
m
o
n
i
to
ri
n
g
a
d
d
ictio
n
c
o
n
d
it
io
n
s
a
n
d
c
o
n
str
u
c
ti
n
g
trea
tme
n
t
p
r
o
t
o
c
o
ls,
wit
h
th
e
g
o
a
l
o
f
m
o
d
ify
i
n
g
th
e
se
p
ro
t
o
c
o
ls
a
c
c
o
rd
i
n
g
l
y
.
We
fo
c
u
s
o
n
p
re
d
ictin
g
su
c
h
tra
n
sfo
rm
a
ti
o
n
s
t
o
i
ll
u
m
i
n
a
te
th
e
u
n
d
e
rly
in
g
fa
c
to
rs
b
e
h
in
d
sh
ift
s
in
m
e
n
tal
h
e
a
lt
h
.
T
o
a
c
h
iev
e
th
is
o
b
j
e
c
ti
v
e
,
d
a
ta
fr
o
m
a
n
Ira
q
i
h
o
sp
it
a
l
h
a
s
b
e
e
n
c
o
ll
e
c
ted
a
n
d
a
n
a
ly
z
e
d
y
ield
i
n
g
p
ro
m
isin
g
re
su
lt
s.
K
ey
w
o
r
d
s
:
Dis
ea
s
e
p
r
ed
ictio
n
Ma
ch
in
e
lear
n
in
g
Me
d
ical
p
r
o
f
ile
Me
n
tal
illn
ess
W
o
r
d
em
b
ed
d
in
g
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Z
ein
Al
Ab
id
in
I
b
r
ah
im
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
Facu
lty
o
f
Scien
ce
s
,
L
eb
a
n
ese
Un
iv
er
s
ity
Had
at
C
am
p
u
s
,
B
eir
u
t,
L
eb
an
o
n
E
m
ail:
ze
in
.
ib
r
ah
im
@
u
l.e
d
u
.
lb
1.
I
NT
RO
D
UCT
I
O
N
I
n
t
o
d
a
y
'
s
d
i
g
i
ta
l
e
r
a
,
t
e
c
h
n
o
l
o
g
i
c
a
l
a
d
v
a
n
c
e
m
e
n
t
s
h
a
v
e
b
r
o
u
g
h
t
a
b
o
u
t
s
i
g
n
i
f
i
c
a
n
t
c
h
a
n
g
e
s
in
l
i
f
e
s
t
y
l
es
,
c
o
n
t
r
i
b
u
t
i
n
g
t
o
t
h
e
s
p
r
e
a
d
o
f
d
i
s
e
as
es
a
l
o
n
g
w
i
t
h
p
o
l
l
u
t
i
o
n
.
T
h
e
h
e
a
l
t
h
c
a
r
e
s
e
ct
o
r
i
s
n
o
w
f
ac
e
d
w
i
t
h
t
h
e
u
r
g
e
n
t
t
a
s
k
o
f
a
d
a
p
t
i
n
g
t
o
t
h
e
s
e
c
h
a
ll
en
g
e
s
t
o
e
n
s
u
r
e
t
h
e
w
el
l
-
b
e
i
n
g
o
f
i
n
d
i
v
i
d
u
a
l
s
a
n
d
t
h
e
o
v
e
r
a
l
l
p
r
o
s
p
e
r
i
t
y
o
f
n
at
i
o
n
s
.
R
is
i
n
g
h
e
a
lt
h
c
a
r
e
c
o
s
ts
a
n
d
t
h
e
i
n
c
r
e
a
s
i
n
g
p
r
e
v
al
e
n
c
e
o
f
p
h
y
s
ica
l
a
n
d
m
e
n
t
a
l
i
ll
n
e
s
s
es
h
a
v
e
m
ad
e
i
t
m
o
r
e
d
i
f
f
ic
u
l
t
f
o
r
g
o
v
e
r
n
m
e
n
t
s
a
n
d
h
e
a
l
t
h
i
n
s
u
r
a
n
c
e
p
r
o
v
i
d
e
r
s
t
o
e
f
f
e
ct
i
v
e
l
y
a
d
d
r
e
s
s
t
h
e
s
e
i
s
s
u
es
.
T
h
e
d
i
g
it
i
za
t
i
o
n
a
n
d
p
r
o
c
e
s
s
i
n
g
o
f
v
a
s
t
a
m
o
u
n
t
s
o
f
d
a
t
a
h
a
v
e
o
p
e
n
e
d
n
e
w
p
o
s
s
i
b
il
i
ti
e
s
f
o
r
d
i
s
e
as
e
p
r
e
d
i
c
t
i
o
n
,
e
a
r
l
y
i
n
te
r
v
e
n
t
i
o
n
,
a
n
d
t
h
e
d
e
v
e
l
o
p
m
e
n
t
o
f
i
n
n
o
v
a
t
i
v
e
t
r
e
a
t
m
e
n
t
o
p
t
i
o
n
s
.
U
t
il
i
zi
n
g
c
o
m
p
l
e
x
c
o
m
p
u
t
a
t
i
o
n
t
e
c
h
n
o
l
o
g
i
e
s
,
p
a
r
t
i
c
u
l
a
r
l
y
a
r
ti
f
i
c
i
al
i
n
t
e
l
li
g
e
n
c
e
a
n
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m
s
,
h
e
a
l
t
h
c
a
r
e
p
r
o
f
e
s
s
i
o
n
a
l
s
c
a
n
l
e
v
e
r
a
g
e
v
a
r
i
o
u
s
d
a
t
a
s
o
u
r
c
e
s
s
u
c
h
as
i
m
a
g
e
s
,
te
x
t
,
a
n
d
s
o
u
n
d
s
t
o
p
r
ed
i
c
t
d
i
s
e
a
s
e
s
y
m
p
t
o
m
s
a
n
d
t
a
k
e
p
r
e
v
e
n
t
i
v
e
m
e
a
s
u
r
es
w
el
l
i
n
ad
v
a
n
c
e
.
T
h
e
co
llab
o
r
atio
n
b
etwe
en
t
h
e
b
io
m
ed
ical
f
ield
a
n
d
d
ata
s
cien
ce
h
as
led
to
s
ig
n
if
ica
n
t
ad
v
a
n
ce
m
en
ts
in
d
is
ea
s
e
p
r
ed
ictio
n
m
o
d
els
an
d
d
atasets
o
v
er
th
e
p
ast
f
e
w
y
ea
r
s
.
Du
e
to
th
e
g
lo
b
al
r
is
e
in
ad
d
ictio
n
r
ates,
r
esear
ch
er
s
m
u
s
t seek
s
o
lu
tio
n
s
to
ad
d
r
ess
th
i
s
p
r
e
v
alen
t iss
u
e.
Ad
d
itio
n
ally
,
ad
d
icts
ar
e
u
n
d
er
g
o
in
g
tr
ea
tm
e
n
t
p
r
o
to
co
ls
an
d
u
tili
zin
g
s
p
ec
ia
lized
s
y
s
tem
s
f
o
r
d
r
u
g
d
is
p
o
s
al.
Ho
wev
er
,
s
o
m
e
in
d
iv
id
u
al
s
with
in
th
is
g
r
o
u
p
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
385
-
3
9
6
386
m
ig
h
t
e
x
p
er
ien
ce
a
s
h
if
t
in
d
i
ag
n
o
s
is
f
r
o
m
a
d
d
ictio
n
to
an
o
th
er
p
s
y
c
h
o
lo
g
ical
d
is
o
r
d
er
.
T
h
is
ch
an
g
e
p
r
esen
ts
n
ew
ch
allen
g
es
in
im
p
lem
en
tin
g
ef
f
ec
tiv
e
th
er
ap
e
u
tic
p
r
o
to
co
ls
,
as
th
e
u
n
d
er
ly
i
n
g
r
ea
s
o
n
s
f
o
r
th
is
tr
an
s
f
o
r
m
atio
n
r
em
ain
u
n
k
n
o
wn
.
Hen
ce
,
ac
cu
r
ately
p
r
ed
ic
tin
g
th
e
p
r
o
g
r
ess
io
n
f
r
o
m
ad
d
ictio
n
to
an
o
t
h
er
p
s
y
ch
iatr
ic
co
n
d
itio
n
wo
u
l
d
en
ab
le
s
k
illed
d
o
cto
r
s
to
p
r
o
ac
t
iv
ely
im
p
lem
en
t p
r
ev
e
n
tiv
e
m
ea
s
u
r
es a
n
d
p
r
o
v
id
e
r
esear
ch
er
s
with
th
e
o
p
p
o
r
tu
n
ity
to
in
v
esti
g
ate
th
is
p
h
en
o
m
en
o
n
.
T
h
e
aim
o
f
th
is
wo
r
k
is
to
ad
d
r
ess
th
ese
am
b
ig
u
o
u
s
ca
s
es a
n
d
p
r
o
v
id
e
in
s
ig
h
ts
in
to
p
r
ed
ictin
g
an
d
u
n
d
er
s
tan
d
in
g
th
e
tr
a
n
s
f
o
r
m
atio
n
s
th
at
o
cc
u
r
d
u
r
in
g
th
e
tr
ea
tm
en
t
o
f
ad
d
icts
,
s
h
ed
d
in
g
lig
h
t
o
n
th
e
u
n
d
er
ly
in
g
f
ac
to
r
s
b
eh
in
d
th
ese
s
h
if
ts
in
m
en
tal
h
ea
lth
.
B
y
lev
er
ag
in
g
t
h
e
p
o
we
r
o
f
a
r
t
if
i
cial
in
tellig
en
ce
an
d
m
ac
h
in
e
l
e
a
r
n
i
n
g
,
t
h
i
s
r
e
s
ea
r
c
h
e
n
d
e
a
v
o
r
s
t
o
c
o
n
t
r
i
b
u
t
e
t
o
i
m
p
r
o
v
e
d
a
d
d
i
c
t
i
o
n
t
r
e
a
t
m
e
n
t
s
t
r
a
t
e
g
i
es
a
n
d
p
a
t
i
e
n
t
o
u
t
c
o
m
e
s
.
T
h
e
r
em
ain
in
g
s
ec
tio
n
s
o
f
th
e
ar
ticle
will
b
e
o
r
g
an
ized
as
f
o
llo
ws:
Sectio
n
2
will
b
e
d
ed
icate
d
to
p
r
ev
io
u
s
wo
r
k
in
th
e
d
o
m
ain
o
f
d
is
ea
s
e
p
r
ed
ictio
n
.
I
n
s
ec
tio
n
3
,
we
will
p
r
esen
t
an
o
v
e
r
v
iew
o
f
wo
r
d
em
b
e
d
d
in
g
(
W
E
)
m
eth
o
d
s
u
s
ed
in
NL
P
th
r
o
u
g
h
o
u
t
th
is
s
tu
d
y
.
Sectio
n
4
will
o
u
tlin
e
th
e
p
r
o
p
o
s
ed
p
ip
elin
e,
w
h
ile
ex
p
er
im
en
ts
will
b
e
p
r
esen
ted
in
s
ec
tio
n
5
,
co
n
clu
d
in
g
in
s
ec
tio
n
6
with
p
r
o
p
o
s
ed
f
u
tu
r
e
ex
t
en
s
io
n
s
to
th
e
wo
r
k
.
2.
P
RE
VIOU
S WO
RK
S
I
n
th
is
s
ec
tio
n
,
s
in
ce
n
o
p
r
ev
i
o
u
s
wo
r
k
s
h
a
v
e
ad
d
r
ess
ed
th
i
s
d
is
ea
s
e
y
et,
we
will
p
r
esen
t
a
liter
atu
r
e
r
ev
iew
o
f
d
is
ea
s
e
p
r
e
d
ictio
n
in
g
e
n
er
al.
L
iter
at
u
r
e
ca
n
b
e
o
r
g
an
ized
in
v
ar
io
u
s
way
s
.
Fo
r
ex
am
p
le,
we
ca
n
ca
teg
o
r
ize
th
em
b
ased
o
n
t
h
e
d
is
ea
s
e
th
ey
ad
d
r
e
s
s
,
s
u
ch
as h
ea
r
t a
ttack
,
an
d
co
r
o
n
a
v
ir
u
s
,
o
r
b
ased
o
n
th
e
ty
p
e
o
f
m
eth
o
d
u
s
ed
,
s
u
ch
as
tr
a
d
itio
n
al
m
ac
h
in
e
lear
n
in
g
-
b
ased
o
r
d
ee
p
lear
n
in
g
-
b
ased
m
eth
o
d
s
.
I
n
o
u
r
wo
r
k
,
w
e
will a
d
o
p
t th
e
s
ec
o
n
d
ty
p
e
o
f
o
r
g
an
izatio
n
.
2
.
1
.
T
ra
ditio
na
l
m
a
chine
-
lea
rning
-
ba
s
ed
m
et
ho
ds
A
f
i
r
s
t
t
y
p
e
o
f
m
e
t
h
o
d
s
i
n
t
h
i
s
c
a
t
e
g
o
r
y
i
s
t
h
e
o
n
e
p
r
o
p
o
s
e
d
b
y
M
a
l
l
e
t
a
l
.
[
1
]
w
h
o
d
e
v
el
o
p
e
d
a
s
y
s
t
e
m
t
o
p
r
o
c
e
s
s
X
-
r
a
y
i
m
a
g
es
f
o
r
b
o
n
e
f
r
a
c
t
u
r
e
d
e
t
ec
t
i
o
n
,
e
m
p
l
o
y
i
n
g
t
h
e
t
r
a
d
i
ti
o
n
a
l
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
m
e
t
h
o
d
s
.
I
n
i
ti
a
l
l
y
,
t
h
e
y
a
n
a
l
y
z
e
d
a
n
d
p
r
o
c
ess
e
d
t
h
e
i
m
a
g
es
,
e
m
p
h
a
s
i
z
i
n
g
f
ea
t
u
r
e
e
x
t
r
a
c
t
i
o
n
u
s
i
n
g
t
h
e
g
r
a
y
l
e
v
el
co
-
o
c
c
u
r
r
e
n
c
e
m
a
t
r
i
x
(
G
L
C
M
)
m
e
t
h
o
d
.
S
u
b
s
e
q
u
e
n
tly
,
t
h
e
y
u
t
i
l
i
z
e
d
a
v
a
r
i
e
t
y
o
f
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
m
e
t
h
o
d
s
,
w
it
h
t
h
e
m
o
s
t
o
p
t
i
m
a
l
p
e
r
f
o
r
m
a
n
c
e
o
b
s
e
r
v
e
d
w
i
t
h
t
h
e
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
ti
o
n
(
R
B
F
)
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
(
S
V
M
)
m
e
t
h
o
d
.
Sy
m
o
n
s
et
a
l
ex
p
lo
r
es
in
[
2
]
.
T
h
e
p
r
ed
ictio
n
o
f
alco
h
o
l
ad
d
i
ctio
n
tr
ea
tm
en
t
u
s
in
g
a
co
m
p
a
r
ativ
e
s
tu
d
y
b
etwe
en
clin
ical
p
s
y
c
h
o
th
er
a
p
is
ts
an
d
m
ac
h
in
e
lear
n
in
g
.
On
e
n
o
tab
le
f
in
d
i
n
g
i
n
th
eir
wo
r
k
is
th
at
tr
ea
tm
en
t
is
co
n
s
id
er
ed
s
u
cc
ess
f
u
l
if
th
e
p
atien
t
co
m
p
letes
a
m
in
im
u
m
o
f
7
8
o
u
t
o
f
8
4
d
ay
s
.
W
h
ile
n
eu
r
al
n
etwo
r
k
s
h
av
e
s
h
o
wn
s
ig
n
if
ican
t
p
r
o
g
r
ess
,
th
ey
wer
e
n
o
t
u
tili
ze
d
d
u
e
to
ch
allen
g
es
in
h
an
d
lin
g
d
if
f
er
e
n
tly
o
r
g
a
n
ized
an
d
u
n
ex
p
lain
a
b
le
d
ata
as
s
tated
i
n
th
eir
wo
r
k
.
T
h
e
r
esear
ch
e
r
s
d
id
n
o
t
d
el
v
e
in
to
ex
p
lo
r
in
g
a
n
d
wo
r
k
in
g
o
n
th
is
is
s
u
e
.
I
n
s
tead
,
th
ey
em
p
lo
y
ed
a
tr
ad
itio
n
al
m
ac
h
in
e
lear
n
in
g
s
y
s
tem
,
s
p
ec
if
ically
r
a
n
d
o
m
f
o
r
est
an
d
lo
g
is
tic
r
eg
r
ess
io
n
m
eth
o
d
s
.
B
o
th
m
eth
o
d
s
y
ield
e
d
alm
o
s
t
e
q
u
al
a
n
d
h
ig
h
r
esu
lts
,
alth
o
u
g
h
th
e
p
r
ec
is
e
d
etails
o
f
th
e
d
ata
p
r
o
ce
s
s
in
g
m
eth
o
d
wer
e
n
o
t e
lab
o
r
ate
d
u
p
o
n
in
th
e
ar
ti
cle.
S
V
M
a
n
d
k
-
m
e
a
n
s
w
e
r
e
u
s
e
d
b
y
S
i
n
h
a
a
n
d
S
h
a
r
m
a
[
3
]
.
M
o
r
e
p
r
e
c
i
s
e
l
y
,
t
h
e
y
u
s
e
d
S
V
M
a
s
w
e
l
l
a
s
m
o
d
i
f
i
e
d
K
-
m
e
a
n
s
,
w
i
t
h
s
o
m
e
m
i
n
o
r
d
i
f
f
e
r
e
n
c
e
s
r
e
p
r
e
s
e
n
t
e
d
b
y
a
d
d
i
n
g
a
n
R
B
F
k
e
r
n
e
l
t
o
w
o
r
k
o
n
s
e
l
e
c
t
i
n
g
t
h
e
g
o
o
d
f
e
a
t
u
r
e
t
o
u
s
e
a
n
d
y
i
e
l
d
s
t
o
s
a
t
i
s
f
a
c
t
o
r
y
r
e
s
u
l
t
s
i
n
t
h
e
d
e
t
e
c
t
i
o
n
o
f
c
o
r
o
n
a
r
y
a
r
t
e
r
y
d
i
s
e
a
s
e
.
A
t
t
h
e
t
i
m
e
w
h
e
n
m
o
d
i
f
i
e
d
K
-
M
e
a
n
s
p
r
o
v
e
d
t
o
b
e
e
f
f
e
c
t
i
v
e
i
n
t
h
e
w
o
r
k
,
s
i
n
c
e
t
h
e
d
a
t
a
w
a
s
n
o
n
-
l
i
n
e
a
r
,
t
h
i
s
r
e
s
u
l
t
i
n
d
i
c
a
t
e
s
t
h
e
v
a
l
i
d
i
t
y
o
f
t
h
i
s
m
e
t
h
o
d
w
i
t
h
u
n
s
t
r
u
c
t
u
r
e
d
d
a
t
a
.
S
V
M
,
k
-
m
e
a
n
s
,
a
n
d
7
o
t
h
e
r
c
l
a
s
s
i
f
i
e
r
s
w
e
r
e
u
s
e
d
i
n
[
4
]
i
n
o
r
d
e
r
t
o
p
r
e
d
i
c
t
t
h
e
c
o
r
o
n
a
r
y
a
r
t
e
r
y
d
i
s
e
a
s
e
o
f
p
a
t
i
e
n
t
s
b
a
s
e
d
o
n
t
h
r
e
e
t
y
p
e
s
o
f
r
e
p
r
e
s
e
n
t
a
t
i
o
n
s
o
f
m
e
d
i
c
a
l
h
i
s
t
o
r
y
o
f
p
a
t
i
e
n
t
s
w
i
t
h
t
h
e
h
e
l
p
o
f
B
i
o
B
e
r
t
.
T
h
e
r
e
s
u
l
t
s
o
b
t
a
i
n
e
d
s
h
o
w
e
d
t
h
e
p
o
w
e
r
o
f
B
i
o
B
e
r
t
t
o
r
e
p
r
e
s
e
n
t
m
e
d
i
c
a
l
t
e
x
t
h
i
s
t
o
r
y
o
f
p
a
t
i
e
n
t
s
.
B
y
an
aly
zin
g
an
elec
tr
o
en
ce
p
h
alo
g
r
am
(
E
E
G
)
with
th
e
h
elp
o
f
r
an
d
o
m
f
o
r
est
class
if
ier
,
Min
et
a
l.
[
5
]
p
r
esen
ted
a
s
tu
d
y
to
p
r
ed
ict
in
d
iv
id
u
al
r
esp
o
n
s
es
to
elec
tr
o
c
o
n
v
u
ls
iv
e
th
e
r
ap
y
(
E
C
T
)
in
s
c
h
izo
p
h
r
e
n
ia
p
atien
ts
.
Usi
n
g
tr
an
s
f
er
en
tr
o
p
y
(
T
E
)
ex
tr
ac
ted
f
r
o
m
E
E
G
d
ata,
a
r
an
d
o
m
f
o
r
est
class
if
ier
with
f
ea
tu
r
e
s
elec
tio
n
ac
cu
r
ately
class
if
ied
E
C
T
r
esp
o
n
d
er
s
an
d
n
o
n
-
r
esp
o
n
d
er
s
,
ac
h
iev
in
g
8
5
.
3
%
b
alan
ce
d
ac
cu
r
ac
y
.
Hig
h
e
r
ef
f
ec
tiv
e
co
n
n
ec
tiv
ity
in
f
r
o
n
tal
ar
ea
s
m
ay
in
d
icate
a
f
av
o
r
ab
le
E
C
T
r
esp
o
n
s
e,
o
f
f
er
in
g
p
o
ten
tial
f
o
r
p
er
s
o
n
alize
d
E
C
T
d
ec
is
io
n
s
in
clin
ical
p
r
ac
tice.
Sleep
a
p
n
ea
d
etec
tio
n
p
r
o
b
le
m
was
ad
d
r
ess
ed
b
y
J
ez
zin
i
et
a
l
.
[
6
]
.
I
n
th
is
s
tu
d
y
,
a
u
to
m
at
ed
m
eth
o
d
s
f
o
r
d
etec
tin
g
s
leep
ap
n
ea
u
s
in
g
elec
tr
o
ca
r
d
io
g
r
am
(
E
C
G
)
,
aim
in
g
to
o
v
er
co
m
e
th
e
lim
itatio
n
s
o
f
p
o
ly
s
o
m
n
o
g
r
ap
h
y
(
PS
G)
.
B
y
co
m
p
ar
in
g
ex
is
tin
g
ap
p
r
o
ac
h
e
s
an
d
ex
p
lo
r
in
g
v
ar
io
u
s
class
i
f
ier
s
,
th
e
r
esear
ch
ev
alu
ates
th
e
ef
f
icac
y
o
f
a
r
tific
ial
in
tellig
en
ce
alg
o
r
ith
m
s
in
s
leep
ap
n
ea
d
etec
tio
n
.
R
esu
lts
d
em
o
n
s
tr
ate
th
at
th
e
K
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN
)
c
lass
if
ier
ac
h
iev
es
a
r
em
ar
k
ab
l
e
ac
cu
r
ac
y
o
f
9
8
.
7
%,
s
u
r
p
ass
in
g
o
th
er
class
if
ier
s
p
r
ev
io
u
s
ly
u
tili
ze
d
in
liter
atu
r
e.
T
h
e
KNN
class
if
ier
,
alo
n
g
with
o
th
er
class
if
ier
s
,
was
ev
alu
ated
f
o
r
p
r
ed
ictin
g
t
h
e
r
is
k
o
f
d
ev
elo
p
in
g
co
r
o
n
a
r
y
h
ea
r
t
d
is
ea
s
e
(
C
H
D)
in
th
e
n
ex
t
ten
y
ea
r
s
in
th
e
s
tu
d
y
co
n
d
u
cted
b
y
Min
o
u
et
a
l
[
7
]
.
Data
p
r
ep
r
o
ce
s
s
in
g
in
clu
d
ed
clea
n
i
n
g
th
e
d
ata
b
y
r
em
o
v
in
g
m
is
s
in
g
v
alu
es,
ad
d
r
ess
in
g
im
b
al
an
ce
d
class
es,
an
d
ap
p
ly
in
g
th
e
s
y
n
th
etic
m
in
o
r
it
y
o
v
e
r
s
am
p
lin
g
tech
n
iq
u
e
(
SMOT
E
)
to
p
r
e
v
en
t
o
v
er
f
itti
n
g
an
d
d
ata
l
o
s
s
.
R
esu
lts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Ma
ch
in
e
lea
r
n
in
g
f
o
r
men
ta
l
h
ea
lth
:
p
r
ed
ictin
g
tr
a
n
s
itio
n
s
fr
o
m
a
d
d
ictio
n
to
illn
ess
(
A
li A
lkh
a
z
r
a
ji)
387
in
d
icate
d
th
at
th
e
d
ec
is
io
n
tr
ee
alg
o
r
ith
m
p
er
f
o
r
m
e
d
b
etter
th
an
m
o
s
t
class
if
ier
s
,
ex
ce
p
t
f
o
r
n
aiv
e
B
ay
es,
KNN,
an
d
SVM,
wh
ich
also
d
em
o
n
s
tr
ated
ac
ce
p
tab
le
p
e
r
f
o
r
m
an
ce
.
KNN
also
s
h
o
wed
its
ef
f
ec
tiv
e
n
ess
to
p
r
ed
ict
h
ea
r
t
d
is
ea
s
e
in
[
8
]
.
T
h
e
r
esear
ch
e
r
s
an
aly
ze
v
a
s
t
m
ed
ical
d
ata
to
p
r
ed
ict
h
ea
r
t
d
is
ea
s
e,
u
tili
zin
g
s
u
p
er
v
is
ed
lear
n
in
g
alg
o
r
ith
m
s
lik
e
n
aïv
e
B
ay
es,
d
ec
is
io
n
tr
ee
,
KNN
,
an
d
r
an
d
o
m
f
o
r
est.
Usi
n
g
a
d
ataset
f
r
o
m
th
e
C
lev
elan
d
d
atab
ase
co
m
p
r
is
in
g
3
0
3
in
s
tan
c
es
an
d
7
6
attr
ib
u
tes,
th
is
s
tu
d
y
f
o
cu
s
es
o
n
1
4
k
e
y
attr
ib
u
tes
to
ass
ess
alg
o
r
ith
m
p
er
f
o
r
m
an
ce
.
KNN
d
e
m
o
n
s
tr
ates
th
e
h
ig
h
est
ac
cu
r
ac
y
s
co
r
e,
o
f
f
e
r
in
g
i
n
s
ig
h
ts
in
to
h
ea
r
t d
is
ea
s
e
d
ev
elo
p
m
en
t p
r
o
b
ab
ilit
ies.
R
ep
ak
a
et
a
l
.
[
9
]
d
ev
el
o
p
ed
a
m
o
b
ile
ap
p
licatio
n
ca
lled
s
m
ar
t
h
ea
r
t
d
is
ea
s
e
p
r
ed
ictio
n
(
S
HDP
)
th
at
u
s
es
m
ac
h
in
e
lear
n
in
g
tech
n
iq
u
es,
s
p
ec
if
ically
th
e
n
aiv
e
B
ay
es
class
if
ier
,
to
p
r
ed
ict
h
ea
r
t
d
i
s
ea
s
e
b
ased
o
n
p
r
io
r
p
atien
t d
ata.
T
h
e
a
p
p
licatio
n
a
im
s
to
id
en
tify
r
is
k
f
ac
t
o
r
s
f
o
r
h
ea
r
t d
is
ea
s
e
b
y
ca
teg
o
r
izin
g
u
s
er
-
s
u
b
m
itted
d
at
a
u
s
in
g
tr
ad
itio
n
al
m
ac
h
in
e
lea
r
n
in
g
m
eth
o
d
s
.
T
h
e
team
o
b
tain
ed
b
asic
d
ata
f
r
o
m
U
C
I
wh
ich
was
d
er
iv
ed
f
r
o
m
in
f
o
r
m
atio
n
o
n
p
r
e
v
io
u
s
p
atie
n
ts
.
T
h
e
ev
alu
atio
n
o
f
th
e
s
y
s
tem
'
s
ac
cu
r
ac
y
r
ate
tak
es
less
t
h
an
0
.
0
1
s
ec
o
n
d
s
an
d
y
ield
s
a
r
esu
lt o
f
8
8
.
7
7
%
[
9
]
.
Du
r
in
g
th
e
co
r
o
n
av
ir
u
s
d
is
ea
s
e
2
0
1
9
(
C
OVI
D
-
19
)
p
an
d
em
ic,
wh
er
e
in
ten
s
iv
e
ca
r
e
u
n
it
(
I
C
U
)
r
eso
u
r
ce
s
ar
e
s
tr
ain
ed
,
a
m
ac
h
i
n
e
lear
n
i
n
g
-
b
ased
r
is
k
p
r
io
r
itiz
atio
n
to
o
l
was
d
ev
el
o
p
ed
to
f
o
r
ec
ast
I
C
U
tr
an
s
f
er
s
with
in
2
4
h
o
u
r
s
,
aid
in
g
f
r
o
n
tlin
e
h
ea
lth
ca
r
e
wo
r
k
er
s
in
ef
f
icien
t
r
eso
u
r
ce
allo
ca
tio
n
an
d
h
o
s
p
ital
f
lo
w
m
an
ag
em
en
t
[
1
0
]
.
T
h
e
wo
r
k
u
s
es
tim
e
s
er
ies
d
ata
f
r
o
m
n
o
n
-
I
C
U
C
OVI
D
-
1
9
ad
m
is
s
io
n
s
,
a
r
an
d
o
m
f
o
r
est
m
o
d
el
th
at
was
tr
ain
ed
an
d
ev
alu
ated
o
n
a
r
etr
o
s
p
ec
tiv
e
co
h
o
r
t
o
f
1
9
8
7
p
atien
ts
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
d
em
o
n
s
tr
ated
7
2
.
8
%
s
en
s
itiv
ity
,
7
6
.
3
%
s
p
ec
if
icity
,
7
6
.
2
%
ac
cu
r
ac
y
,
a
n
d
7
9
.
9
%
ar
ea
u
n
d
er
th
e
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
ter
is
tics
cu
r
v
e,
o
f
f
er
in
g
a
v
alu
ab
le
s
cr
ee
n
in
g
to
o
l
f
o
r
im
m
in
en
t
I
C
U
tr
an
s
f
er
s
an
d
en
h
an
cin
g
h
o
s
p
ital
r
eso
u
r
ce
allo
ca
tio
n
a
n
d
p
atien
t
th
r
o
u
g
h
p
u
t p
lan
n
i
n
g
d
u
r
in
g
th
e
C
OVI
D
-
1
9
cr
is
is
.
C
h
en
g
et
a
l
.
[
1
0
]
co
n
d
u
cted
r
e
s
ea
r
ch
to
p
r
e
d
ict
a
p
atien
t'
s
n
ee
d
to
b
e
ad
m
itted
to
th
e
with
i
n
2
4
h
o
u
r
s
b
ased
o
n
th
e
h
ea
lth
in
f
o
r
m
atio
n
s
to
r
ed
in
th
eir
h
ea
lth
f
ile.
T
h
e
d
ataset
in
clu
d
es
d
ata
f
o
r
1
9
7
8
p
atien
ts
.
Usi
n
g
th
e
tr
ad
itio
n
al
m
ac
h
i
n
e
lear
n
i
n
g
a
p
p
r
o
ac
h
an
d
th
e
r
an
d
o
m
f
o
r
est
class
if
ier
,
th
e
ev
alu
atio
n
r
esu
lts
s
h
o
wed
an
ac
cu
r
ac
y
r
ate
o
f
6
7
.
2
%.
T
h
e
a
im
o
f
th
e
r
esear
ch
was
to
en
ab
le
tim
ely
in
ter
v
en
tio
n
an
d
c
o
n
tr
o
l
o
f
th
e
p
atien
t'
s
co
n
d
itio
n
b
ef
o
r
e
it wo
r
s
en
s
.
A
d
if
f
er
en
t
ty
p
e
o
f
m
ac
h
in
e
l
ea
r
n
in
g
m
eth
o
d
(
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
(
ML
P)
)
was
u
s
ed
with
o
th
er
alg
o
r
ith
m
s
to
p
r
ed
ict
co
v
i
d
-
1
9
o
u
tb
r
ea
k
s
.
Ar
d
ab
ili
et
a
l.
[
1
1
]
ex
p
l
o
r
es
th
e
lim
itatio
n
s
o
f
s
tan
d
ar
d
ep
id
em
io
lo
g
ical
m
o
d
els
f
o
r
p
r
ed
ictin
g
C
OVI
D
-
1
9
o
u
tb
r
e
ak
s
an
d
p
r
o
p
o
s
es
m
ac
h
i
n
e
lear
n
in
g
an
d
s
o
f
t
co
m
p
u
tin
g
m
o
d
els
as
alter
n
at
iv
es.
Am
o
n
g
v
ar
io
u
s
m
o
d
els
ex
am
in
ed
,
th
e
ML
P
a
n
d
a
d
ap
tiv
e
n
etwo
r
k
-
b
ased
f
u
zz
y
in
f
e
r
en
ce
s
y
s
tem
(
ANFI
S)
s
h
o
w
p
r
o
m
is
e
in
p
r
e
d
ictin
g
o
u
t
b
r
ea
k
s
.
T
h
e
s
tu
d
y
ad
v
o
ca
tes
f
o
r
m
ac
h
in
e
lear
n
in
g
'
s
ef
f
ec
tiv
en
ess
in
m
o
d
elin
g
th
e
co
m
p
lex
an
d
v
ar
ie
d
n
atu
r
e
o
f
C
OVI
D
-
1
9
o
u
t
b
r
e
ak
s
,
s
u
g
g
esti
n
g
its
in
teg
r
atio
n
with
tr
a
d
itio
n
al
SE
I
R
m
o
d
els f
o
r
e
n
h
an
ce
d
p
r
e
d
i
ctiv
e
ca
p
ab
il
ities
.
L
a
s
t
l
y
,
P
r
i
y
a
a
n
d
J
i
n
n
y
[
1
2
]
p
r
e
s
e
n
t
s
a
n
o
t
h
e
r
t
y
p
e
o
f
a
p
p
l
i
c
a
t
i
o
n
s
o
f
t
h
e
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m
s
.
T
h
i
s
w
o
r
k
e
v
a
l
u
a
t
e
s
m
a
c
h
i
n
e
-
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
,
i
n
c
l
u
d
i
n
g
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
s
,
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
,
e
x
t
r
e
m
e
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
,
a
n
d
r
a
n
d
o
m
f
o
r
e
s
t
,
t
o
p
r
e
d
i
c
t
i
n
v
i
t
r
o
f
e
r
t
i
l
i
z
a
t
i
o
n
(
I
V
F
)
p
r
e
g
n
a
n
c
y
o
u
t
c
o
m
e
s
u
s
i
n
g
D
o
p
p
l
e
r
a
n
d
c
l
i
n
i
c
a
l
p
a
r
a
m
e
t
e
r
s
.
A
m
o
n
g
t
h
e
s
e
t
e
c
h
n
i
q
u
e
s
,
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
c
o
m
b
i
n
e
d
w
i
t
h
r
a
n
d
o
m
f
o
r
e
s
t
i
m
p
o
r
t
a
n
c
e
f
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
d
e
m
o
n
s
t
r
a
t
e
d
t
h
e
h
i
g
h
e
s
t
p
e
r
f
o
r
m
a
n
c
e
,
a
c
h
i
e
v
i
n
g
a
n
a
c
c
u
r
a
c
y
o
f
8
2
.
3
%
.
T
h
e
r
e
s
u
l
t
s
h
i
g
h
l
i
g
h
t
t
h
e
s
i
g
n
i
f
i
c
a
n
c
e
o
f
u
l
t
r
a
s
o
u
n
d
m
e
a
s
u
r
e
m
e
n
t
p
a
r
a
m
e
t
e
r
s
,
p
a
r
t
i
c
u
l
a
r
l
y
D
o
p
p
l
e
r
p
a
r
a
m
e
t
e
r
s
,
i
n
i
n
f
l
u
e
n
c
i
n
g
I
V
F
o
u
t
c
o
m
e
s
.
2
.
2
.
Deep
-
lea
rning
-
ba
s
ed
met
ho
ds
Deep
-
lear
n
in
g
m
eth
o
d
s
s
tar
te
d
to
en
ter
in
th
e
d
o
m
ain
o
f
h
e
alth
ca
r
e
esp
ec
ially
f
o
r
d
is
ea
s
e
p
r
e
d
ictio
n
.
Z
h
an
g
et
a
l.
[
1
3
]
in
tr
o
d
u
ce
a
m
eth
o
d
f
o
r
d
is
ea
s
e
p
r
ed
ictio
n
an
d
ea
r
ly
i
n
ter
v
en
tio
n
u
s
in
g
a
s
en
ten
ce
s
im
ilar
ity
m
o
d
el.
T
h
e
ap
p
r
o
ac
h
in
co
r
p
o
r
ates
WE
an
d
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
etwo
r
k
s
(
C
NN)
to
ex
tr
a
ct
s
en
ten
ce
v
ec
to
r
s
f
r
o
m
p
atien
t
d
ata,
f
o
cu
s
in
g
o
n
s
y
m
p
to
m
s
im
ilar
ity
a
n
aly
s
is
.
T
h
e
m
o
d
el
i
n
teg
r
ates
s
y
n
ta
ctic
tr
ee
an
d
n
e
u
r
al
n
etwo
r
k
co
m
p
u
tatio
n
s
to
en
h
an
ce
ac
cu
r
ac
y
.
Utilizin
g
th
e
Mic
r
o
s
o
f
t
R
esear
ch
Par
ap
h
r
ase
I
d
en
tific
atio
n
(
MSR
P)
d
ataset,
th
e
m
o
d
el
ac
h
iev
ed
8
3
.
9
%
F1
s
co
r
e
a
n
d
ac
cu
r
ac
y
,
d
em
o
n
s
tr
atin
g
its
ef
f
ec
tiv
en
ess
in
p
r
ed
ict
in
g
d
is
ea
s
es
an
d
en
a
b
lin
g
ea
r
ly
in
ter
v
en
tio
n
.
Ad
d
itio
n
ally
,
ex
p
e
r
im
en
ts
o
n
t
h
e
s
em
an
tic
tex
tu
al
s
im
ilar
ity
t
ask
s
h
o
wed
p
r
o
m
i
s
in
g
r
esu
lts
,
in
d
icatin
g
th
e
m
o
d
el'
s
ca
p
ab
ilit
y
to
ex
tr
ac
t
k
ey
in
f
o
r
m
ati
o
n
f
r
o
m
s
en
ten
ce
s
f
o
r
d
is
ea
s
e
p
r
ed
ictio
n
an
d
ea
r
ly
in
ter
v
en
tio
n
.
C
N
N
wa
s
a
ls
o
u
s
e
d
i
n
[
1
4
]
w
h
i
c
h
p
r
e
s
e
n
t
s
a
m
e
t
h
o
d
l
e
v
e
r
a
g
in
g
r
e
a
l
-
l
i
f
e
i
n
t
e
r
n
a
t
i
o
n
a
l
c
l
ass
if
i
c
a
t
i
o
n
o
f
d
i
s
e
as
e
s
(
I
C
D
)
-
c
o
d
e
d
e
le
c
t
r
o
n
i
c
m
e
d
i
c
a
l
r
ec
o
r
d
s
(
E
MR
)
f
r
o
m
A
f
r
i
c
a
(
2
0
1
8
-
2
0
1
9
)
t
o
d
e
v
e
l
o
p
a
d
i
s
e
as
e
r
is
k
p
r
e
d
i
c
t
i
o
n
m
o
d
e
l
.
Us
i
n
g
C
N
N
,
t
h
e
m
o
d
e
l
p
r
o
c
e
s
s
e
s
E
M
R
d
ata
i
n
t
o
m
u
l
t
i
-
s
t
e
p
ti
m
e
-
s
e
r
i
es
f
o
r
e
c
a
s
ti
n
g
d
a
t
a
a
n
d
p
r
e
d
i
c
t
s
f
u
t
u
r
e
d
is
e
a
s
e
r
i
s
k
b
as
e
d
o
n
d
e
m
o
g
r
a
p
h
i
c
a
n
d
m
e
d
i
c
a
l
h
is
t
o
r
y
.
E
x
p
e
r
i
m
e
n
t
a
l
r
esu
l
t
s
d
e
m
o
n
s
t
r
a
t
e
a
n
8
0
.
7
3
%
a
c
c
u
r
a
c
y
i
n
p
r
e
d
i
c
t
i
n
g
i
n
d
i
v
i
d
u
a
l
d
i
s
e
as
e
r
is
k
.
T
h
e
m
o
d
e
l
o
f
f
e
r
s
p
o
t
e
n
t
i
al
c
o
s
t
s
a
v
i
n
g
s
,
p
r
e
-
e
m
p
t
i
v
e
c
o
r
r
e
c
t
i
v
e
a
ct
i
o
n
s
,
a
n
d
c
a
p
a
c
it
y
p
l
a
n
n
i
n
g
i
n
s
i
g
h
t
s
f
o
r
h
o
s
p
i
t
a
ls
,
p
r
o
v
i
d
i
n
g
p
e
a
c
e
o
f
m
i
n
d
r
e
g
a
r
d
i
n
g
p
e
r
s
o
n
a
l
h
e
al
t
h
.
T
h
e
p
o
wer
o
f
d
ee
p
-
lear
n
in
g
-
b
ased
m
eth
o
d
s
was
h
ig
h
lig
h
t
ed
in
[
1
5
]
in
w
h
ich
a
co
m
p
ar
is
o
n
o
f
th
e
r
esu
lts
o
f
ap
p
ly
in
g
tr
ad
itio
n
al
m
ac
h
in
e
-
lear
n
i
n
g
-
b
ased
a
n
d
d
ee
p
-
lear
n
in
g
-
b
ased
al
g
o
r
ith
m
s
f
o
r
m
o
r
tality
esti
m
atio
n
s
was c
o
n
d
u
cted
.
T
h
e
r
esear
ch
c
o
n
d
u
cted
a
c
o
m
p
r
eh
en
s
iv
e
an
aly
s
is
o
f
C
OVI
D
-
1
9
s
p
r
ea
d
d
y
n
am
ics
an
d
m
o
r
tality
esti
m
atio
n
s
in
So
u
th
Ko
r
ea
.
Utilizin
g
d
iv
er
s
e
d
ata
s
ets
in
clu
d
in
g
d
em
o
g
r
ap
h
ic,
lo
ca
tio
n
,
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
385
-
3
9
6
388
ep
id
em
io
lo
g
ical
in
f
o
r
m
atio
n
,
th
e
s
tu
d
y
r
e
v
ea
led
s
ig
n
if
ican
t
tr
en
d
s
in
p
a
n
d
em
ic
p
r
o
g
r
ess
io
n
.
T
h
e
an
al
y
s
is
h
ig
h
lig
h
ted
a
p
ea
k
i
n
m
o
r
tality
r
ates
d
u
r
in
g
th
e
in
itial
an
d
m
id
-
p
h
ases
o
f
th
e
o
u
tb
r
ea
k
,
f
o
llo
wed
b
y
a
d
ec
lin
e
s
u
g
g
esti
n
g
in
cr
ea
s
ed
in
d
iv
i
d
u
al
r
esil
ien
ce
.
Mo
r
eo
v
er
,
th
e
s
t
u
d
y
em
p
h
asized
th
e
elev
ated
r
is
k
am
o
n
g
eld
er
l
y
in
d
iv
id
u
als
an
d
m
ales.
Fu
r
th
e
r
m
o
r
e,
t
h
e
r
esear
ch
e
m
p
lo
y
e
d
v
ar
io
u
s
ar
tific
ial
in
tellig
en
ce
-
b
ased
m
o
d
els
f
o
r
p
r
ed
ictiv
e
an
aly
s
is
o
f
q
u
ar
an
ti
n
ed
C
OVI
D
-
1
9
ca
s
es.
L
o
g
is
tic
r
eg
r
ess
io
n
,
SVM,
an
d
d
ee
p
lear
n
in
g
s
eq
u
e
n
tial
m
o
d
els
ac
h
iev
ed
h
i
g
h
ac
cu
r
a
cy
r
ates,
with
th
e
d
ee
p
lear
n
i
n
g
m
o
d
el
y
ield
in
g
th
e
h
ig
h
es
t
ac
cu
r
ac
y
o
f
9
9
%
th
r
o
u
g
h
m
eticu
lo
u
s
p
ar
am
eter
tu
n
in
g
an
d
p
r
e
v
en
tio
n
o
f
o
v
e
r
f
itti
n
g
.
T
h
ese
f
in
d
in
g
s
p
r
o
v
id
e
v
alu
ab
le
in
s
ig
h
ts
f
o
r
h
ea
lth
ca
r
e
p
r
ac
ti
tio
n
e
r
s
to
en
h
an
ce
th
eir
p
r
e
p
ar
ed
n
ess
an
d
r
esp
o
n
s
e
s
tr
ateg
ies.
T
h
e
p
o
wer
o
f
d
ee
p
-
lear
n
i
n
g
-
b
ased
alg
o
r
ith
m
s
h
as
p
r
o
v
e
n
th
eir
ef
f
icien
cy
also
b
u
t
th
is
tim
e
in
th
e
p
r
ed
ictio
n
o
f
ac
u
te
k
id
n
ey
in
j
u
r
y
(
AKI
)
[
1
6
]
.
T
h
is
s
tu
d
y
ai
m
ed
to
co
m
p
ar
e
th
e
p
r
ed
ictiv
e
ac
cu
r
ac
y
o
f
two
m
o
d
els
f
o
r
AKI
d
ev
elo
p
m
en
t
in
I
C
U
p
atie
n
ts
.
Utilizin
g
u
r
in
e
o
u
tp
u
t
tr
en
d
s
,
a
d
ee
p
l
ea
r
n
in
g
m
o
d
el
was
ev
alu
ated
ag
ain
s
t
lo
g
is
tic
r
eg
r
ess
io
n
f
o
r
AKI
p
r
e
d
ictio
n
.
R
e
s
u
lts
f
r
o
m
3
5
,
5
7
3
I
C
U
p
atien
ts
r
ev
ea
led
th
e
d
ee
p
lear
n
in
g
m
o
d
el
ac
h
iev
ed
h
ig
h
er
ac
cu
r
ac
y
(
AUC=0
.
8
9
,
s
en
s
itiv
ity
=0
.
8
,
s
p
ec
if
icity
=0
.
8
4
)
th
a
n
lo
g
is
tic
r
eg
r
ess
io
n
.
R
em
ar
k
a
b
ly
,
th
e
d
ee
p
lear
n
in
g
m
o
d
el
an
ticip
ate
d
8
8
%
o
f
AKI
ca
s
es
o
v
er
1
2
h
o
u
r
s
p
r
io
r
to
o
n
s
et,
d
em
o
n
s
tr
atin
g
its
p
o
te
n
tial f
o
r
ea
r
ly
in
ter
v
e
n
tio
n
a
n
d
p
r
ev
en
t
io
n
in
I
C
U
s
ettin
g
s
.
W
h
en
it
co
m
es
to
an
aly
s
in
g
m
ed
ical
im
ag
es
s
u
ch
as
X
-
r
ay
o
r
C
T
im
a
g
es,
h
e
r
e
co
m
es
th
e
p
o
wer
o
f
d
ee
p
C
NN
-
b
ased
n
etwo
r
k
as
u
s
ed
in
[
1
7
]
.
I
n
th
is
wo
r
k
,
au
th
o
r
s
p
r
o
v
e
th
e
u
tili
ty
o
f
d
ee
p
lear
n
in
g
-
b
ased
f
ea
t
u
r
e
ex
tr
ac
tio
n
f
r
am
ew
o
r
k
s
f
o
r
a
u
to
m
atic
C
OVI
D
-
1
9
class
if
ic
atio
n
,
g
iv
en
t
h
e
ab
s
en
ce
o
f
clin
ically
ap
p
r
o
v
e
d
tr
ea
tm
en
ts
.
Var
io
u
s
d
ee
p
C
NN
,
in
clu
d
in
g
Mo
b
ileNet,
Den
s
eNe
t,
an
d
R
esNet,
wer
e
ev
alu
ated
to
ex
tr
ac
t
f
ea
t
u
r
es
f
r
o
m
ch
est
X
-
r
ay
an
d
C
T
im
ag
es.
T
h
ese
f
ea
tu
r
es
wer
e
th
en
in
p
u
tted
in
to
m
ac
h
in
e
lea
r
n
in
g
class
if
ier
s
to
d
if
f
er
en
tiate
C
OVI
D
-
1
9
ca
s
es
f
r
o
m
c
o
n
tr
o
ls
.
No
ta
b
ly
,
th
e
Den
s
eNe
t1
2
1
f
ea
t
u
r
e
ex
tr
ac
to
r
p
a
ir
ed
with
a
B
ag
g
in
g
tr
ee
class
if
ier
d
em
o
n
s
tr
ated
th
e
h
ig
h
est
p
e
r
f
o
r
m
an
ce
,
ac
h
iev
in
g
9
9
%
class
if
icatio
n
ac
cu
r
ac
y
,
f
o
llo
wed
clo
s
ely
b
y
a
h
y
b
r
i
d
R
esNet5
0
f
ea
tu
r
e
ex
tr
ac
to
r
tr
ain
ed
with
L
ig
h
tG
B
M,
ac
h
iev
in
g
9
8
%
ac
cu
r
ac
y
.
T
h
is
ap
p
r
o
ac
h
h
o
ld
s
p
r
o
m
is
e
f
o
r
en
h
a
n
cin
g
ea
r
ly
d
etec
tio
n
an
d
m
o
n
ito
r
i
n
g
o
f
C
OVI
D
-
1
9
,
p
o
ten
tially
r
ed
u
cin
g
m
o
r
tality
r
ates.
2
.
3
.
Na
t
ura
l la
ng
ua
g
e
pro
ce
s
s
ing
Me
d
ical
d
ata
an
d
r
ec
o
r
d
s
co
n
tain
v
ar
io
u
s
ty
p
es
o
f
in
f
o
r
m
atio
n
,
in
clu
d
in
g
tex
tu
al
r
e
p
o
r
ts
an
d
d
escr
ip
tio
n
s
,
wh
ich
h
o
ld
ess
en
tial
in
f
o
r
m
atio
n
f
o
r
a
n
aly
s
is
.
Ho
wev
er
,
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
ty
p
ically
r
eq
u
ir
e
n
u
m
e
r
ic
in
p
u
t
d
ata
f
o
r
p
r
o
ce
s
s
in
g
.
T
o
b
r
id
g
e
th
is
g
a
p
,
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
(
NL
P)
tec
h
n
iq
u
es
o
f
f
er
a
s
o
l
u
tio
n
b
y
tr
an
s
f
o
r
m
in
g
tex
t
in
to
n
u
m
er
ical
v
ec
to
r
s
t
h
r
o
u
g
h
em
b
ed
d
in
g
s
tech
n
iq
u
e
s
.
T
h
ese
tech
n
iq
u
es
co
n
v
er
t
tex
t
u
al
in
f
o
r
m
atio
n
in
to
h
ig
h
-
d
im
en
s
io
n
al
v
ec
t
o
r
s
,
e
n
ab
lin
g
alg
o
r
ith
m
s
to
in
ter
p
r
e
t
an
d
an
aly
ze
tex
t
-
b
ased
d
ata
ef
f
ec
tiv
el
y
.
I
n
th
e
f
o
llo
win
g
,
we
will g
iv
e
a
n
o
v
e
r
v
iew
o
f
em
b
e
d
d
in
g
tech
n
iq
u
es in
th
e
liter
atu
r
e
:
−
W
o
r
d
2
V
e
c
[
1
8
]
:
W
o
r
d
2
V
e
c
en
c
o
d
e
s
t
h
e
m
ea
n
i
n
g
o
f
w
o
r
d
s
i
n
t
o
c
o
m
p
a
c
t
a
n
d
d
e
n
s
e
v
e
ct
o
r
s
,
c
al
l
e
d
WE
.
T
h
e
s
e
e
m
b
e
d
d
i
n
g
s
r
e
p
r
e
s
e
n
t
wo
r
d
s
i
n
a
c
o
n
t
i
n
u
o
u
s
v
e
c
t
o
r
s
p
a
c
e
,
w
h
e
r
e
w
o
r
d
s
w
i
t
h
s
i
m
il
a
r
m
e
a
n
i
n
g
s
h
a
v
e
v
e
c
t
o
r
s
t
h
at
a
r
e
cl
o
s
e
t
o
g
e
t
h
e
r
,
o
f
t
e
n
m
e
a
s
u
r
e
d
b
y
c
o
s
i
n
e
s
im
i
l
a
r
i
t
y
.
W
h
il
e
W
o
r
d
2
V
e
c
m
o
d
e
l
s
c
o
n
s
i
d
e
r
c
o
n
t
e
x
t
d
u
r
i
n
g
t
r
a
i
n
i
n
g
b
y
e
x
a
m
i
n
i
n
g
n
e
i
g
h
b
o
r
i
n
g
w
o
r
d
s
,
e
a
c
h
w
o
r
d
i
n
t
h
e
v
o
c
a
b
u
l
a
r
y
i
s
r
e
p
r
e
s
e
n
t
e
d
b
y
a
s
i
n
g
l
e
v
ec
t
o
r
.
H
o
w
e
v
e
r
,
t
h
i
s
a
p
p
r
o
a
c
h
m
a
y
n
o
t
f
u
l
l
y
c
a
p
t
u
r
e
a
ll
t
h
e
n
u
a
n
c
es
o
f
w
o
r
d
m
e
a
n
i
n
g
s
a
n
d
c
o
n
te
x
t
s
.
−
Glo
Ve
[
1
9
]
:
Glo
Ve
lev
er
ag
es
t
h
e
co
-
o
cc
u
r
r
en
ce
s
tatis
tics
o
f
wo
r
d
s
in
a
co
r
p
u
s
to
ca
p
tu
r
e
g
l
o
b
al
co
n
tex
t
u
al
in
f
o
r
m
atio
n
b
y
c
o
n
s
tr
u
ctin
g
a
wo
r
d
-
wo
r
d
co
-
o
cc
u
r
r
e
n
ce
m
atr
ix
b
ased
o
n
th
e
e
n
tire
co
r
p
u
s
.
I
t
ex
h
ib
its
n
o
tab
le
p
er
f
o
r
m
a
n
ce
in
task
s
s
u
ch
as
wo
r
d
an
alo
g
y
an
d
n
am
ed
en
tity
id
e
n
tific
atio
n
.
W
h
ile
Glo
Ve
co
m
p
etes
with
W
o
r
d
2
Vec
i
n
ce
r
tain
task
s
,
its
p
er
f
o
r
m
a
n
ce
m
ay
s
u
r
p
ass
th
at
o
f
W
o
r
d
2
Vec
in
o
th
er
s
,
u
n
d
er
s
co
r
i
n
g
its
v
er
s
atility
an
d
ef
f
ec
tiv
en
ess
in
ca
p
t
u
r
in
g
s
e
m
an
tic
r
elatio
n
s
h
ip
s
in
tex
tu
al
d
ata.
−
E
m
b
e
d
d
i
n
g
s
f
r
o
m
l
a
n
g
u
a
g
e
m
o
d
e
l
s
(
E
L
Mo
)
[
2
0
]
:
E
L
M
o
r
e
p
r
e
s
e
n
t
s
a
d
i
f
f
e
r
e
n
t
a
p
p
r
o
a
c
h
t
o
WE
c
o
m
p
a
r
e
d
t
o
G
l
o
V
e
a
n
d
W
o
r
d
2
V
e
c
.
I
n
s
t
e
a
d
o
f
r
e
p
r
e
s
e
n
t
i
n
g
e
a
c
h
w
o
r
d
w
i
t
h
a
s
i
n
g
l
e
v
e
c
t
o
r
r
e
g
a
r
d
l
e
s
s
o
f
i
t
s
c
o
n
t
e
x
t
,
E
L
M
o
c
a
p
t
u
r
e
s
t
h
e
c
o
n
t
e
x
t
o
f
t
h
e
w
o
r
d
w
i
t
h
i
n
t
h
e
e
n
t
i
r
e
s
e
n
t
e
n
c
e
o
r
p
h
r
a
s
e
.
I
t
a
c
h
i
e
v
e
s
t
h
i
s
b
y
p
r
o
d
u
c
i
n
g
d
i
f
f
e
r
e
n
t
e
m
b
e
d
d
i
n
g
s
f
o
r
t
h
e
s
a
m
e
w
o
r
d
u
s
e
d
i
n
d
i
f
f
e
r
e
n
t
c
o
n
t
e
x
t
s
a
c
r
o
s
s
v
a
r
i
o
u
s
s
e
n
t
e
n
c
e
s
.
T
h
i
s
c
o
n
t
e
x
t
u
a
l
i
z
e
d
w
o
r
d
r
e
p
r
e
s
e
n
t
a
t
i
o
n
e
n
a
b
l
e
s
E
L
M
o
t
o
c
a
p
t
u
r
e
n
u
a
n
c
e
s
i
n
m
e
a
n
i
n
g
a
n
d
s
y
n
t
a
c
t
i
c
s
t
r
u
c
t
u
r
e
s
m
o
r
e
e
f
f
e
c
t
i
v
e
l
y
.
−
B
id
ir
ec
tio
n
al
en
co
d
er
r
e
p
r
ese
n
tatio
n
s
f
r
o
m
tr
an
s
f
o
r
m
er
s
(
B
E
R
T
)
[
2
1
]
:
B
E
R
T
is
d
e
s
ig
n
ed
to
d
ev
el
o
p
m
o
d
els
f
o
r
s
p
ec
if
ic
task
s
s
u
ch
as
q
u
esti
o
n
-
an
s
wer
in
g
an
d
lan
g
u
ag
e
in
f
er
en
ce
.
I
t
in
co
r
p
o
r
ate
s
an
ad
d
itio
n
al
o
u
tp
u
t
lay
er
f
o
r
f
in
e
-
tu
n
in
g
,
al
lo
win
g
f
o
r
a
d
ap
tatio
n
to
d
if
f
e
r
en
t
task
s
an
d
d
atasets
.
B
E
R
T
is
co
n
s
id
er
ed
m
o
r
e
s
tr
aig
h
t
f
o
r
wa
r
d
an
d
ef
f
ec
tiv
e
co
m
p
ar
ed
to
tr
a
d
itio
n
a
l
lan
g
u
ag
e
m
o
d
elin
g
ap
p
r
o
ac
h
es
d
u
e
to
its
b
id
ir
ec
tio
n
al
ar
c
h
itectu
r
e
an
d
co
n
tex
tu
alize
d
em
b
ed
d
i
n
g
s
.
−
B
io
B
E
R
T
[
2
2
]
:
B
io
B
E
R
T
is
a
p
r
e
-
tr
ain
e
d
m
o
d
el
s
p
ec
if
i
ca
lly
d
esig
n
ed
f
o
r
r
ep
r
esen
ti
n
g
b
io
lo
g
ical
lan
g
u
ag
e
i
n
b
io
m
ed
ical
tex
t
m
in
in
g
ap
p
licatio
n
s
.
Un
lik
e
B
E
R
T
,
wh
ich
is
p
r
e
-
tr
ain
e
d
o
n
g
en
er
ic
co
r
p
o
r
a
s
u
ch
as
W
ik
ip
ed
ia
an
d
b
o
o
k
s
,
B
io
B
E
R
T
i
s
tr
ain
ed
o
n
b
io
m
ed
ical
co
r
p
o
r
a,
in
clu
d
in
g
Pu
b
Me
d
ab
s
tr
ac
ts
an
d
PMC
f
u
ll
-
tex
t
ar
ticles.
I
n
a
s
tu
d
y
b
y
L
ee
et
a
l
.
[
2
2
]
,
B
io
B
E
R
T
was
f
in
e
-
tu
n
ed
o
n
th
r
ee
b
io
m
e
d
ical
tex
t
m
in
in
g
task
s
:
n
am
ed
en
tity
r
ec
o
g
n
itio
n
(
B
io
NE
R
)
,
r
elatio
n
ex
tr
ac
tio
n
(
B
io
R
E
)
,
an
d
q
u
esti
o
n
an
s
wer
in
g
(
B
io
QA)
.
T
h
e
m
o
d
el's
weig
h
ts
ar
e
in
it
ialized
f
r
o
m
B
E
R
T
,
r
esu
ltin
g
in
n
o
tab
le
p
er
f
o
r
m
a
n
ce
im
p
r
o
v
em
e
n
ts
:
B
io
NE
R
s
aw
a
0
.
6
2
%
i
n
cr
ea
s
e
in
F1
s
co
r
e,
B
io
R
E
im
p
r
o
v
ed
b
y
2
.
8
0
%
in
F1
s
co
r
e,
a
n
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Ma
ch
in
e
lea
r
n
in
g
f
o
r
men
ta
l
h
ea
lth
:
p
r
ed
ictin
g
tr
a
n
s
itio
n
s
fr
o
m
a
d
d
ictio
n
to
illn
ess
(
A
li A
lkh
a
z
r
a
ji)
389
B
io
QA
ac
h
iev
ed
a
1
2
.
2
4
%
m
e
an
r
ec
ip
r
o
ca
l
r
an
k
(
MRR
)
im
p
r
o
v
em
en
t.
T
h
ese
r
esu
lts
h
ig
h
lig
h
t
B
io
B
E
R
T
'
s
ef
f
ec
tiv
en
ess
in
v
ar
io
u
s
b
io
m
e
d
ical
tex
t m
in
in
g
task
s
.
−
Fu
s
i
n
g
B
i
o
B
E
R
T
[
2
3
]
:
T
h
e
m
o
d
el
f
o
r
B
io
NE
R
in
te
g
r
at
es
d
ee
p
co
n
t
ex
tu
al
-
le
v
el
WE
,
c
o
m
p
r
is
i
n
g
an
atte
n
t
io
n
-
b
ase
d
b
id
ir
ec
t
io
n
al
l
o
n
g
s
h
o
r
t
-
te
r
m
m
e
m
o
r
y
(
B
i
L
S
T
M)
-
co
n
d
it
io
n
al
r
a
n
d
o
m
f
iel
d
(
C
R
F)
la
y
e
r
,
a
r
e
p
r
ese
n
t
ati
o
n
la
y
er
,
a
n
d
a
n
i
n
p
u
t
la
y
e
r
.
B
y
co
m
b
i
n
i
n
g
B
i
o
B
E
R
T
,
W
o
r
d
2
Ve
c
f
o
r
WE
,
B
i
L
ST
M
f
o
r
ch
ar
ac
t
er
e
m
b
e
d
d
in
g
,
a
n
d
c
o
n
tex
tu
al
em
b
e
d
d
i
n
g
(
E
L
M
o
)
,
n
o
ta
b
l
e
i
m
p
r
o
v
e
m
e
n
ts
w
e
r
e
o
b
s
er
v
e
d
ac
r
o
s
s
v
a
r
i
o
u
s
d
at
asets
.
S
p
ec
i
f
i
ca
l
ly
,
f
o
r
d
a
tas
ets
r
e
lat
ed
t
o
d
r
u
g
s
a
n
d
c
h
em
ica
ls
(
B
C
5
C
D
R
-
C
h
e
m
a
n
d
B
C
4
C
HE
M
D)
,
t
h
e
p
e
r
f
o
r
m
a
n
ce
r
ati
o
in
cr
ea
s
ed
b
y
0
.
7
%
a
n
d
0
.
9
5
%,
r
es
p
e
cti
v
el
y
.
S
im
i
la
r
ly
,
f
o
r
d
at
asets
lin
k
e
d
t
o
g
e
n
es
a
n
d
p
r
o
te
in
s
(
J
NL
PB
A
a
n
d
B
C
2
G
M)
,
t
h
e
m
o
d
el
e
n
h
a
n
ce
d
t
h
e
p
er
f
o
r
m
a
n
ce
r
at
io
b
y
1
.
7
%.
I
n
d
atas
ets
ass
o
c
iat
e
d
wit
h
d
is
e
ases
(
B
C
5
C
DR
-
d
is
ea
s
e
a
n
d
N
C
B
I
d
is
ea
s
e
d
atas
ets
)
,
th
e
m
o
d
el
a
ch
ie
v
e
d
a
n
im
p
r
o
v
e
d
p
er
f
o
r
m
a
n
ce
r
ati
o
o
f
2
.
0
%
an
d
2
.
1
%,
r
es
p
e
cti
v
e
ly
.
−
C
lin
ical
B
E
R
T
[
2
4
]
:
C
lin
ical
B
E
R
T
is
em
p
lo
y
ed
t
o
p
r
ed
ict
a
3
0
-
d
ay
h
o
s
p
ital r
ea
d
m
is
s
io
n
ac
r
o
s
s
v
ar
io
u
s
ad
m
is
s
io
n
tim
es.
I
n
itially
tr
ai
n
ed
o
n
clin
ical
n
o
tes,
B
E
R
T
is
th
en
f
i
n
e
-
tu
n
e
d
u
s
in
g
th
e
m
ed
ic
al
in
f
o
r
m
atio
n
m
ar
t
f
o
r
in
ten
s
iv
e
ca
r
e
I
I
I
(
MI
MI
C
-
I
I
I
)
d
ataset,
s
p
ec
if
ically
f
o
cu
s
in
g
o
n
h
o
s
p
ital
r
ea
d
m
is
s
io
n
.
T
h
is
m
o
d
el
ex
h
ib
its
v
er
s
atility
b
ey
o
n
d
r
e
ad
m
is
s
io
n
p
r
ed
ictio
n
,
p
r
o
v
in
g
u
s
ef
u
l
in
d
iag
n
o
s
in
g
co
n
d
itio
n
s
,
esti
m
atin
g
m
o
r
tality
r
is
k
,
an
d
ass
ess
in
g
th
e
d
u
r
atio
n
o
f
s
tay
.
I
ts
ef
f
icac
y
e
x
ten
d
s
to
en
h
a
n
cin
g
p
er
f
o
r
m
an
ce
ac
r
o
s
s
m
u
ltip
le
d
atasets
,
in
clu
d
in
g
d
is
ea
s
es
s
u
ch
as
B
C
5
C
D
R
-
d
is
ea
s
e
an
d
NC
B
I
d
is
ea
s
e
(
2
%
an
d
2
.
1
%,
r
esp
ec
tiv
ely
)
,
d
r
u
g
d
atasets
lik
e
B
C
5
C
D
R
-
C
h
em
an
d
B
C
4
C
HE
MD
(
0
.
7
%
an
d
0
.
9
5
%,
r
esp
ec
tiv
ely
)
,
a
n
d
g
en
e/p
r
o
tein
d
atasets
lik
e
J
NL
PB
A
an
d
B
C
2
GM
(
1
.
7
6
% a
n
d
1
.
9
6
%,
r
esp
ec
tiv
el
y
)
.
−
B
io
Alb
er
t
[
2
5
]
:
T
h
e
B
io
Alb
er
t
m
o
d
el
em
p
lo
y
s
two
-
p
ar
a
m
eter
r
ed
u
ctio
n
tech
n
iq
u
es
to
ad
d
r
ess
th
e
ch
allen
g
es
ass
o
ciate
d
with
s
ca
lin
g
p
r
e
-
tr
ain
ed
m
o
d
els
lik
e
B
E
R
T
,
wh
ich
ten
d
to
h
a
v
e
a
lar
g
e
n
u
m
b
er
o
f
p
ar
am
eter
s
,
r
esu
ltin
g
in
p
r
o
lo
n
g
ed
tr
ai
n
in
g
tim
es.
B
io
Alb
er
t,
p
r
e
-
t
r
ain
ed
o
n
b
io
m
e
d
ical
d
ata,
u
tili
ze
s
in
itial
weig
h
ts
f
r
o
m
Alb
e
r
t,
l
ev
er
ag
in
g
Pu
b
Me
d
,
an
d
PMC
d
atasets
.
Du
r
in
g
th
e
B
io
N
E
R
f
in
e
-
tu
n
in
g
p
h
ase,
th
e
m
o
d
el
d
em
o
n
s
tr
ates
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
v
ar
io
u
s
d
atasets
,
o
u
tp
er
f
o
r
m
in
g
o
th
er
ap
p
r
o
ac
h
es.
No
tab
ly
,
in
eig
h
t
d
atas
ets,
B
io
Alb
er
t
co
n
s
is
te
n
tly
d
eliv
er
s
im
p
r
ess
iv
e
r
esu
lts
,
s
u
r
p
ass
in
g
co
m
p
etin
g
m
o
d
els,
as h
ig
h
lig
h
ted
in
th
e
f
o
r
th
co
m
in
g
c
o
m
p
ar
is
o
n
.
−
B
i
L
S
T
M
+
W
E
+
c
h
a
r
a
c
t
e
r
e
m
b
e
d
d
i
n
g
(
CE
)
[
2
6
]
:
A
s
y
s
t
e
m
d
e
v
i
s
e
d
f
o
r
e
x
t
r
a
c
t
i
n
g
c
h
e
m
i
c
a
l
n
a
m
e
s
f
r
o
m
b
i
o
m
e
d
i
c
a
l
t
e
x
t
i
n
t
e
g
r
a
t
e
s
d
y
n
a
m
i
c
r
e
c
u
r
r
e
n
t
n
e
u
r
a
l
n
e
t
w
o
r
k
s
(
R
N
N
s
)
,
C
R
F
s
,
a
n
d
B
i
L
S
T
M
.
E
v
a
l
u
a
t
i
o
n
o
n
t
h
e
N
C
B
I
c
o
r
p
u
s
,
J
N
L
P
B
A
c
o
r
p
u
s
,
a
n
d
B
i
o
C
r
e
a
t
i
v
e
I
I
G
M
c
o
r
p
u
s
(
g
e
n
e
)
(
d
i
s
e
a
s
e
)
d
e
m
o
n
s
t
r
a
t
e
d
n
o
t
a
b
l
e
p
e
r
f
o
r
m
a
n
c
e
,
p
a
r
t
i
c
u
l
a
r
l
y
e
x
c
e
l
l
i
n
g
i
n
t
h
e
J
N
L
P
B
A
c
o
r
p
u
s
f
o
r
F
1
s
c
o
r
e
.
H
o
w
e
v
e
r
,
B
i
o
B
e
r
t
a
n
d
B
i
L
S
T
M
-
C
R
F
s
u
r
p
a
s
s
e
d
i
t
i
n
B
i
o
C
r
e
a
t
i
v
e
I
I
G
M
c
o
r
p
u
s
a
n
d
N
C
B
I
c
o
r
p
u
s
,
a
c
h
i
e
v
i
n
g
s
c
o
r
e
s
o
f
8
9
.
9
8
a
n
d
9
0
.
8
4
,
r
e
s
p
e
c
t
i
v
e
l
y
.
−
C
o
m
p
ar
in
g
F1
s
co
r
es
in
th
e
NC
B
I
d
is
ea
s
e
d
ataset
am
o
n
g
f
o
u
r
m
o
d
els
—
B
io
B
er
t
,
f
u
s
io
n
o
f
B
io
B
E
R
T
,
B
iLST
M+
W
E
+CE,
an
d
B
io
A
lb
er
t
—
B
io
Alb
er
t
ac
h
iev
ed
th
e
h
ig
h
est
r
esu
lt
(
9
7
.
1
8
%)
c
o
m
p
ar
ed
to
th
e
o
th
er
s
.
T
r
ad
itio
n
al
m
o
d
els
m
en
tio
n
ed
ea
r
lier
(
W
o
r
d
2
Vec
,
E
L
Mo
,
an
d
Glo
Ve)
wer
e
o
u
t
p
er
f
o
r
m
ed
b
y
B
E
R
T
ac
co
r
d
in
g
to
th
e
s
tate
-
of
-
th
e
-
a
r
t
b
en
ch
m
a
r
k
s
m
en
tio
n
ed
.
Ad
d
itio
n
ally
,
C
lin
icalBer
t,
ap
p
lied
to
a
d
is
tin
ct
d
ataset,
s
h
o
wed
p
r
o
m
is
in
g
r
esu
lts
with
a
7
1
.
4
%
A
UR
O
C
b
en
ch
m
ar
k
.
C
o
n
s
eq
u
e
n
tly
,
B
io
Alb
er
t
em
er
g
es a
s
th
e
to
p
-
p
er
f
o
r
m
in
g
m
o
d
el
am
o
n
g
t
h
o
s
e
ev
alu
ated
.
3.
P
RO
P
O
SE
D
AP
P
RO
ACH
T
h
e
o
b
jectiv
e
o
f
o
u
r
s
tu
d
y
is
to
ascer
tain
wh
eth
er
an
ad
d
ict
m
ay
r
ec
eiv
e
a
d
is
tin
ct
d
iag
n
o
s
is
b
ased
o
n
th
eir
m
ed
ical
p
r
o
f
ile.
W
e
will
wo
r
k
with
two
d
is
tin
ct
o
r
g
an
izatio
n
s
o
f
t
h
e
s
am
e
d
ataset:
o
n
e
o
r
g
an
ize
d
"
b
y
p
atien
ts
"
an
d
th
e
o
th
er
"b
y
v
is
its
".
I
n
th
e
f
o
r
m
er
,
all
p
atien
t
v
is
its
ar
e
ag
g
r
e
g
ated
in
to
a
s
in
g
le
d
ataset
with
o
u
t
d
u
p
licatio
n
.
C
o
n
s
eq
u
en
tly
,
ea
ch
p
atien
t is r
ep
r
esen
ted
b
y
a
s
in
g
le
d
ata
f
ile
co
n
tain
in
g
th
ei
r
p
r
o
f
ile,
b
u
t n
o
t th
e
d
etails
o
f
in
d
iv
i
d
u
al
v
is
its
.
I
n
co
n
tr
ast,
th
e
latter
d
ataset
r
etain
s
a
ll
v
is
it
-
s
p
ec
if
ic
in
f
o
r
m
ati
o
n
f
o
r
ea
c
h
p
atien
t.
T
h
e
ex
p
e
r
im
en
tal
p
r
o
ce
s
s
co
m
p
r
is
es f
o
u
r
s
tep
s
as sh
o
w
n
in
F
ig
u
r
e
1
.
Fig
u
r
e
1
.
Pro
p
o
s
ed
a
p
p
r
o
ac
h
p
ip
elin
e
3
.
1
.
P
r
o
po
s
ed
a
pp
ro
a
ch
pip
eline
3
.
1
.
1
.
Da
t
a
co
llect
io
n
W
e
u
tili
ze
d
au
th
en
tic
d
ata
s
o
u
r
ce
d
f
r
o
m
th
e
I
b
n
R
o
s
h
d
Ho
s
p
ital
f
o
r
Me
n
tal
I
lln
ess
an
d
Ad
d
ictio
n
T
r
ea
tm
en
t
in
B
ag
h
d
ad
,
I
r
aq
.
T
h
e
d
ataset
co
n
tain
s
in
d
iv
i
d
u
al
p
atien
t
p
r
o
f
iles
alo
n
g
with
th
eir
r
esp
ec
tiv
e
d
iag
n
o
s
es.
T
h
r
ee
E
x
ce
l sh
ee
ts
wer
e
co
llected
:
−
Ma
in
: T
h
is
s
h
ee
t e
n
co
m
p
ass
es th
e
p
atien
t p
r
o
f
iles
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
385
-
3
9
6
390
−
Su
b
_
T
ab
le:
I
t
p
r
o
v
i
d
es d
etailed
in
f
o
r
m
atio
n
a
b
o
u
t e
ac
h
p
atie
n
t v
is
it.
−
Diag
n
o
s
is
: T
h
is
s
h
ee
t lis
t
s
th
e
d
iag
n
o
s
is
n
am
es a
n
d
t
h
eir
co
r
r
esp
o
n
d
in
g
co
d
es.
a)
Pro
f
iles
T
h
e
p
atien
ts
’
p
r
o
f
iles
f
r
o
m
af
o
r
em
en
tio
n
e
d
I
r
a
q
i
h
o
s
p
ital
co
m
p
r
is
e
a
co
llectio
n
o
f
p
e
r
s
o
n
al
an
d
m
ed
ical
d
ata
s
tr
u
ctu
r
ed
as f
o
llo
ws:
1.
PAT_
I
D:
Patien
t I
D
2.
GNDR: Ge
n
d
er
3.
B
R
T
H:
B
ir
th
Date
4.
VI
SIT
_
DAT
E
: D
ate
o
f
ea
ch
p
atien
t v
is
it
5.
L
AB
_
T
E
ST:
L
ab
test
s
co
n
d
u
c
ted
b
y
th
e
p
atien
t d
u
r
in
g
e
ac
h
v
is
it
6.
HSTRY:
Patien
t
'
s
m
ed
ical
h
is
t
o
r
y
7.
E
C
T
: I
n
d
icate
s
if
th
e
p
atien
t is u
n
d
er
g
o
in
g
elec
tr
o
s
h
o
c
k
th
er
ap
y
(
0
o
r
1
)
8.
T
R
E
AT
ME
NT
: M
ed
icatio
n
ad
m
in
is
ter
ed
to
th
e
p
atien
t d
u
r
i
n
g
ea
ch
v
is
it
9.
DI
AGNO
SIS: D
iag
n
o
s
is
a
s
s
ig
n
ed
to
th
e
p
atien
t d
u
r
in
g
e
ac
h
v
is
it
10.
C
ONDI
T
I
ON:
Pat
ien
t’
s
s
tatu
s
d
u
r
in
g
ea
ch
v
is
it
b)
Diag
n
o
s
is
E
a
c
h
d
i
a
g
n
o
s
i
s
is
a
c
c
o
m
p
a
n
i
e
d
b
y
a
c
e
r
t
i
f
i
e
d
c
o
d
e
f
r
o
m
t
h
e
W
o
r
l
d
H
e
a
l
t
h
O
r
g
a
n
i
z
a
ti
o
n
'
s
d
i
a
g
n
o
s
t
i
c
l
is
ts
o
f
d
i
s
e
as
e
s
,
k
n
o
w
n
as
(
I
C
D
9
)
,
w
h
i
c
h
s
h
o
u
l
d
b
e
u
ti
l
iz
e
d
o
n
t
h
e
m
a
i
n
s
h
e
e
t
.
F
u
r
t
h
e
r
m
o
r
e
,
t
h
e
d
i
a
g
n
o
s
es
a
r
e
d
o
c
u
m
e
n
t
e
d
i
n
A
r
a
b
i
c
.
T
h
e
p
a
ti
e
n
t
s
w
e
a
r
e
a
n
a
l
y
z
i
n
g
,
w
h
o
a
r
e
a
d
d
i
c
ts
,
a
r
e
as
s
i
g
n
e
d
c
o
d
e
s
r
a
n
g
i
n
g
f
r
o
m
F
1
0
t
o
F
1
9
.
T
a
b
l
e
1
p
r
e
s
e
n
ts
t
h
e
n
u
m
er
i
c
a
l
i
n
f
o
r
m
a
t
i
o
n
o
f
t
h
e
d
a
ta
s
et
w
h
i
l
e
T
a
b
l
e
2
l
is
t
s
s
o
m
e
s
t
a
ti
s
t
ic
s
a
b
o
u
t
t
h
e
d
a
t
as
e
t
.
T
ab
le
1
.
Sam
p
le
o
f
d
ata
ID
B
R
TH
GNDR
EC
T
LA
B
V
S
T_
D
G
N
S
_
M
D
C
N
2
7
7
5
1
9
9
4
F
0
<
2
0
1
0
-
12
-
0
4
>
F
0
0
<
2
0
1
0
-
12
-
1
9
>
G
4
0
<
2
0
1
1
-
02
-
1
3
>
G
4
0
<
2
0
2
1
-
07
-
0
4
>
F
4
4
2
7
7
7
1
9
6
8
F
0
<
2
0
1
1
-
06
-
2
0
>
F
3
2
<
2
0
1
3
-
03
-
2
8
>
F
5
1
<
2
0
1
3
-
04
-
2
9
>
F
5
1
<
2
0
1
3
-
05
-
2
8
>
F
5
1
<
2
0
1
6
-
04
-
0
3
>
F
3
4
<
2
0
2
0
-
11
-
1
5
>
F
4
1
<
2
0
2
0
-
11
-
1
9
>
F
4
1
2
7
7
8
1
9
5
5
M
0
<
2
0
1
1
-
08
-
1
4
>
F
2
0
O
l
a
n
z
a
p
i
n
e
<
2
0
1
1
-
11
-
1
5
>
F
2
0
O
l
a
n
z
a
p
i
n
e
<
2
0
1
2
-
06
-
2
6
>
F
2
0
O
l
a
n
z
a
p
i
n
e
C
i
t
a
l
o
p
r
a
m (as
H
B
r
)
<
2
0
1
3
-
09
-
1
7
>
F
2
0
O
l
a
n
z
a
p
i
n
e
<
2
0
1
4
-
06
-
0
1
>
F
2
0
Q
u
e
t
i
a
p
i
n
e
<
2
0
1
4
-
10
-
1
4
>
F
2
0
Q
u
e
t
i
a
p
i
n
e
<
2
0
1
5
-
01
-
2
2
>
F
2
0
E
c
i
t
a
l
o
p
r
a
m
<
2
0
1
6
-
03
-
2
4
>
F
2
0
O
l
a
n
z
a
p
i
n
e
C
h
l
o
r
d
i
a
z
e
p
o
x
i
d
e
<
2
0
2
0
-
0
9
0
3
>
F
2
0
C
h
l
o
r
d
i
a
z
e
p
o
x
i
d
e
O
l
a
n
z
a
p
i
n
e
2
7
7
9
1
9
7
9
M
0
<
2
0
1
1
-
08
-
1
4
>
F
1
9
F
l
u
o
x
e
t
i
n
e
H
C
l
C
h
l
o
r
d
i
a
z
e
p
o
x
i
d
e
2
7
8
0
1
9
8
2
M
0
<
2
0
1
1
-
08
-
1
4
>
F
1
3
C
a
r
b
a
ma
z
e
p
i
n
e
C
i
t
a
l
o
p
r
a
m (as
H
B
r
)
T
ab
le
2.
Statis
tical
in
f
o
r
m
atio
n
ab
o
u
t
d
ataset
P
a
t
i
e
n
t
s
9
4
0
9
P
a
t
i
e
n
t
s V
i
si
t
s
5
7
5
7
2
A
d
d
i
c
t
s
1
7
8
5
A
d
d
i
c
t
s Vis
i
t
s
8
0
8
0
S
w
a
p
p
e
d
P
a
t
i
e
n
t
s
1
0
4
S
w
a
p
p
e
d
V
i
s
i
t
s
1
1
7
4
3
.
1
.
2
.
Da
t
a
prepro
ce
s
s
ing
W
e
em
p
lo
y
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
a
d
ata
m
in
in
g
tech
n
i
q
u
e,
to
r
ef
in
e
th
e
r
aw
d
ata
f
o
r
f
u
r
th
er
p
r
o
ce
s
s
in
g
,
en
s
u
r
in
g
it
is
p
r
esen
ted
in
a
m
ea
n
in
g
f
u
l
an
d
ef
f
ec
tiv
e
f
o
r
m
at
.
Giv
en
t
h
at
th
e
u
tili
za
tio
n
o
f
n
o
is
y
d
ata
ca
n
lead
to
in
f
er
i
o
r
f
in
d
in
g
s
d
esp
ite
em
p
lo
y
in
g
r
o
b
u
s
t
alg
o
r
ith
m
s
,
th
i
s
p
r
ep
r
o
ce
s
s
in
g
s
tep
is
c
r
itical
f
o
r
ac
h
iev
in
g
m
o
r
e
p
r
ec
is
e
r
esu
lts
.
T
h
e
f
o
llo
win
g
m
ea
s
u
r
es will b
e
im
p
lem
en
ted
to
ac
h
iev
e
th
is
g
o
al.
a)
Data
clea
n
in
g
T
h
e
f
o
llo
win
g
s
tr
ateg
ies ar
e
e
m
p
lo
y
ed
to
ad
d
r
ess
th
e
ch
allen
g
es p
o
s
ed
b
y
m
is
s
in
g
an
d
n
o
i
s
y
d
ata:
−
Miss
in
g
d
ata:
T
o
m
an
ag
e
m
is
s
in
g
d
ata,
we
h
av
e
two
o
p
tio
n
s
:
we
ca
n
eith
er
ex
clu
d
e
th
e
en
tire
tu
p
le
co
n
tain
in
g
m
is
s
in
g
v
alu
es,
o
r
we
ca
n
r
ep
lace
th
e
m
is
s
in
g
v
al
u
es
with
eith
er
th
e
attr
ib
u
te
m
e
an
o
r
th
e
m
o
s
t
p
r
o
b
a
b
le
v
alu
e.
I
n
o
u
r
d
ataset,
we
h
av
e
d
ec
id
ed
to
d
is
ca
r
d
t
h
e
two
co
lu
m
n
s
lab
elled
L
ab
(
6
3
9
2
/8
0
8
0
)
an
d
C
o
n
d
itio
n
(
7
2
6
6
/8
0
8
0
)
d
u
e
to
th
e
s
ig
n
if
ican
t
n
u
m
b
er
o
f
m
is
s
in
g
v
alu
es
th
ey
co
n
tain
.
Fo
r
t
h
e
r
em
ain
in
g
co
lu
m
n
s
,
an
y
m
is
s
in
g
v
alu
es
will b
e
s
u
b
s
titu
ted
with
0
.
−
No
is
y
d
ata:
Ma
ch
i
n
es
ar
e
in
c
ap
ab
le
o
f
i
n
ter
p
r
etin
g
m
ea
n
in
g
less
d
ata,
o
f
ten
r
ef
er
r
e
d
to
a
s
n
o
is
y
d
ata.
Ho
wev
er
,
th
is
asp
ec
t w
as n
o
t
co
n
s
id
er
ed
in
o
u
r
an
aly
s
is
.
b)
Data
tr
an
s
f
o
r
m
atio
n
T
h
is
s
tag
e
aim
s
to
tr
an
s
f
o
r
m
th
e
d
ata
in
to
a
f
o
r
m
at
s
u
itab
le
f
o
r
th
e
m
in
i
n
g
p
r
o
ce
s
s
.
T
h
e
f
o
llo
win
g
s
tep
s
will b
e
u
n
d
er
tak
en
t
o
ac
h
iev
e
th
is
:
−
No
r
m
aliza
tio
n
:
T
h
is
in
v
o
lv
es
s
ca
lin
g
th
e
d
ata
to
a
s
tan
d
ar
d
r
an
g
e
.
I
n
o
u
r
ca
s
e,
we
will
p
er
f
o
r
m
th
e
f
o
llo
win
g
u
p
d
ates:
i)
T
r
an
s
late
th
e
"Ag
e"
co
lu
m
n
f
r
o
m
g
en
d
er
(
M,
F)
to
n
u
m
er
ical
v
alu
es
(
0
,
1
)
;
an
d
ii)
No
r
m
alize
th
e
"Bi
r
th
Date
"
co
lu
m
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Ma
ch
in
e
lea
r
n
in
g
f
o
r
men
ta
l
h
ea
lth
:
p
r
ed
ictin
g
tr
a
n
s
itio
n
s
fr
o
m
a
d
d
ictio
n
to
illn
ess
(
A
li A
lkh
a
z
r
a
ji)
391
−
C
r
ea
tio
n
o
f
n
ew
attr
ib
u
tes
:
W
e
will
g
en
er
ate
n
ew
attr
ib
u
tes
b
ased
o
n
th
e
p
r
o
v
id
ed
lis
t.
Fo
r
c
o
lu
m
n
s
co
n
tain
in
g
d
i
v
er
s
e
in
f
o
r
m
atio
n
s
u
ch
as
"h
is
to
r
y
,
"
"d
iag
n
o
s
tic,
"
an
d
"th
er
ap
y
,
"
we
will
s
p
li
t
th
e
d
ata
in
to
d
is
tin
ct
co
lu
m
n
s
,
u
tili
zin
g
an
ap
p
r
o
ac
h
s
im
ilar
to
th
e
o
n
e
ap
p
lied
f
o
r
o
th
er
r
elev
an
t c
o
lu
m
n
s
.
T
h
ese
s
tep
s
will e
n
s
u
r
e
th
at
th
e
d
ata
is
ap
p
r
o
p
r
iately
p
r
ep
ar
e
d
f
o
r
t
h
e
s
u
b
s
eq
u
e
n
t m
in
in
g
p
r
o
ce
s
s
.
3
.
1
.
3
.
F
ea
t
ure
e
x
t
ra
ct
i
o
n
T
h
e
p
r
o
ce
s
s
o
f
e
x
t
r
ac
t
in
g
ess
e
n
ti
al
i
n
f
o
r
m
ati
o
n
f
r
o
m
r
aw
d
at
a
f
o
r
d
i
v
e
r
s
e
a
p
p
li
ca
t
io
n
s
is
ca
l
led
f
e
at
u
r
e
ex
t
r
ac
ti
o
n
.
Gi
v
en
t
h
at
m
o
s
t
d
at
a
p
o
s
s
ess
n
u
m
e
r
i
ca
l a
ttr
ib
u
t
es,
v
a
r
i
o
u
s
m
et
h
o
d
s
e
x
is
t
t
o
co
n
v
e
r
t
te
x
t
u
al
d
at
a
i
n
t
o
n
u
m
er
i
ca
l
v
e
ct
o
r
s
,
ali
g
n
i
n
g
wit
h
t
h
e
o
b
jec
ti
v
es
o
f
m
a
ch
i
n
e
le
ar
n
i
n
g
al
g
o
r
i
th
m
s
.
W
o
r
d
v
ec
t
o
r
iz
ati
o
n
,
N
L
P
tec
h
n
i
q
u
e,
f
a
cil
ita
tes
t
h
e
c
o
n
v
e
r
s
i
o
n
o
f
w
o
r
d
s
i
n
t
o
v
e
ct
o
r
s
.
W
o
r
d
s
ar
e
m
ap
p
ed
o
n
t
o
v
ec
to
r
s
p
a
ce
s
u
s
i
n
g
a
r
a
n
g
e
o
f
la
n
g
u
a
g
e
m
o
d
els
.
W
it
h
i
n
t
h
i
s
v
e
ct
o
r
s
p
ac
e,
ea
c
h
w
o
r
d
is
d
e
p
ic
te
d
b
y
a
v
ec
t
o
r
c
o
n
s
is
t
in
g
o
f
r
e
al
n
u
m
b
e
r
s
,
s
u
c
h
as
‘
m
an
’
a
n
d
‘
w
o
m
a
n
’
f
o
r
n
o
u
n
s
o
r
‘
w
al
k
i
n
g
’
a
n
d
‘
wa
lk
e
d
’
f
o
r
v
e
r
b
s
,
as
s
h
o
w
n
i
n
F
ig
u
r
e
2
.
M
a
le
-
F
e
m
a
le
Ve
rb
ten
se
Fig
u
r
e
2
. S
im
ilar
wo
r
d
v
ec
to
r
i
za
tio
n
Sev
er
al
m
o
d
els,
p
ar
ticu
lar
ly
th
o
s
e
tailo
r
ed
f
o
r
th
e
b
io
m
ed
ical
d
o
m
ain
,
a
r
e
task
ed
with
th
is
r
esp
o
n
s
ib
ilit
y
,
as
d
is
cu
s
s
ed
i
n
th
e
NL
P
s
ec
tio
n
.
T
o
in
co
r
p
o
r
ate
o
u
r
d
ata,
we
h
av
e
o
p
t
ed
f
o
r
B
io
B
er
t
an
d
B
io
Alb
er
t.
T
h
e
th
r
ee
attr
ib
u
te
s
u
tili
ze
d
in
th
is
ap
p
r
o
ac
h
ar
e
h
is
to
r
y
,
tr
ea
tm
en
t,
an
d
d
iag
n
o
s
is
.
T
h
e
f
o
llo
win
g
s
tep
s
will b
e
tak
en
:
−
L
o
ad
in
g
p
r
e
-
tr
ai
n
ed
m
o
d
els
:
W
e
will
lo
ad
th
e
p
r
e
-
tr
ai
n
ed
m
o
d
els
f
o
r
B
io
B
er
t
[
2
7
]
an
d
B
io
Alb
er
t
[
2
8
]
u
s
in
g
th
e
h
u
g
g
in
g
f
ac
e
a
p
p
lica
tio
n
p
r
o
g
r
am
m
i
n
g
in
ter
f
ac
e
(
API
)
.
−
Ob
tain
in
g
em
b
ed
d
in
g
s
:
W
e
w
ill
u
tili
ze
s
p
ec
if
ic
f
u
n
ctio
n
s
f
r
o
m
B
E
R
T
an
d
AL
B
E
R
T
[
2
8
]
,
[
2
9
]
to
o
b
tai
n
em
b
ed
d
in
g
s
an
d
c
h
ar
ac
ter
is
tics
.
I
t
i
s
w
o
r
t
h
n
o
t
i
n
g
t
h
a
t
t
h
e
f
e
a
t
u
r
e
c
o
u
n
t
s
e
x
t
r
a
c
t
e
d
f
r
o
m
t
h
e
t
w
o
m
o
d
e
l
s
d
i
f
f
e
r
(
s
e
e
T
a
b
l
e
3
)
.
E
a
c
h
f
e
a
t
u
r
e
i
s
i
d
e
n
t
i
f
i
e
d
b
y
i
t
s
a
s
s
o
c
i
a
t
e
d
a
t
t
r
i
b
u
t
e
n
a
m
e
a
n
d
u
n
i
q
u
e
n
u
m
b
e
r
,
s
u
c
h
a
s
"
H
S
T
R
i
,
"
"
T
R
E
A
T
M
E
N
T
i
,
"
a
n
d
"
D
G
N
S
i
"
.
Hen
ce
,
we
n
o
w
p
o
s
s
ess
th
e
f
in
alize
d
d
ata
r
ea
d
y
f
o
r
ap
p
lic
atio
n
in
th
e
s
u
b
s
eq
u
en
t
p
h
ase,
n
am
ely
class
if
icatio
n
.
T
h
e
tex
tu
al
att
r
ib
u
tes
h
av
e
b
ee
n
r
ep
lace
d
with
th
eir
co
r
r
esp
o
n
d
i
n
g
ch
a
r
ac
ter
is
tics
,
an
d
th
e
ca
teg
o
r
ical
attr
ib
u
te
h
as
b
ee
n
en
co
d
ed
as (
0
.
1
)
.
Ad
d
itio
n
ally
,
th
e
attr
ib
u
te
"Ag
e"
h
as r
ep
la
ce
d
th
e
b
ir
th
d
ate.
T
ab
le
3
.
Nu
m
b
er
o
f
ex
tr
ac
te
d
f
ea
tu
r
es
M
o
d
e
l
F
e
a
t
u
r
e
s
B
i
o
B
e
r
t
7
6
8
B
i
o
A
l
B
e
r
t
4
0
9
6
3
.
1
.
4
.
Cla
s
s
if
ica
t
io
n
I
d
en
tify
in
g
an
d
p
ar
titi
o
n
in
g
d
i
f
f
er
en
t
d
ata
class
es
an
d
co
n
ce
p
ts
th
r
o
u
g
h
a
m
o
d
el
is
a
cr
u
ci
al
s
tep
in
th
e
d
ata
m
in
in
g
p
r
o
ce
s
s
,
k
n
o
w
n
as
class
if
icat
io
n
.
T
h
e
g
o
al
is
to
ac
cu
r
ately
p
r
e
d
ict
th
e
tar
g
et
class
f
o
r
ea
ch
ca
s
e
b
ased
o
n
th
e
av
ailab
le
d
ata.
T
h
is
task
was a
cc
o
m
p
lis
h
ed
th
r
o
u
g
h
t
h
e
f
o
llo
win
g
s
tep
s
,
as d
ep
icted
in
Fig
u
r
e
3
.
Step
s
:
−
E
n
s
u
r
e
th
e
d
ata
is
p
r
ep
ar
e
d
ac
co
r
d
in
g
to
ea
r
lier
p
h
ases
.
−
Select
v
ar
io
u
s
class
if
ier
s
.
−
E
m
p
lo
y
th
e
cr
o
s
s
-
v
alid
atio
n
a
p
p
r
o
ac
h
to
tr
ain
an
d
ass
ess
th
e
m
o
d
els u
s
in
g
5
-
f
o
ld
v
alid
atio
n
.
Fig
u
r
e
3
.
C
lass
if
icatio
n
ac
tio
n
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
385
-
3
9
6
392
Af
t
er
we
tr
a
n
s
f
o
r
m
ed
te
x
t
u
a
l
d
at
a
i
n
t
o
n
u
m
e
r
ic
al
o
n
e
u
s
i
n
g
B
io
B
E
R
T
a
n
d
B
i
o
A
L
B
E
R
T
em
b
e
d
d
i
n
g
tec
h
n
i
q
u
es.
w
e
t
h
e
n
u
tili
ze
d
t
h
is
d
at
a
in
tw
o
d
is
ti
n
ct
w
a
y
s
.
First
ly
,
b
y
a
n
al
y
zi
n
g
p
ati
en
t
I
Ds
r
e
g
a
r
d
less
o
f
th
e
n
u
m
b
e
r
o
f
th
ei
r
v
is
its
a
n
d
c
h
a
n
g
es.
S
ec
o
n
d
l
y
,
we
c
o
n
s
id
e
r
e
d
t
h
e
p
ati
en
t'
s
f
r
e
q
u
e
n
t
v
is
its
a
n
d
th
e
co
r
r
es
p
o
n
d
i
n
g
d
ia
g
n
o
s
is
at
ea
ch
v
is
it.
W
e
an
aly
ze
d
t
h
e
r
es
u
l
ts
u
s
i
n
g
b
o
th
b
ala
n
c
ed
a
n
d
u
n
b
al
an
ce
d
d
at
a,
i
g
n
o
r
i
n
g
t
h
e
d
at
e
at
all
p
r
ev
io
u
s
s
t
ag
es
a
n
d
f
o
c
u
s
i
n
g
o
n
t
h
e
s
w
a
p
as
t
h
e
c
lass
f
ea
tu
r
e
.
T
h
e
t
wo
o
r
g
an
iz
ati
o
n
s
o
f
t
h
e
d
a
ta
a
r
e
as
f
o
l
lo
ws
.
a)
B
y
p
atien
ts
Un
d
er
th
is
v
er
s
io
n
,
two
s
am
p
l
es
o
f
d
ata
will
b
e
ca
teg
o
r
ized
:
o
n
e
with
b
alan
ce
d
class
es
an
d
an
o
t
h
er
with
u
n
b
alan
ce
d
class
es.
B
o
t
h
s
am
p
les
will
in
co
r
p
o
r
ate
f
ea
tu
r
es
d
er
iv
ed
f
r
o
m
B
io
B
er
t
an
d
B
io
Alb
er
t.
I
t
is
im
p
o
r
tan
t to
h
ig
h
lig
h
t th
at
th
e
class
if
ier
"stack
"
in
teg
r
ates S
VM
,
r
an
d
o
m
f
o
r
est
,
a
n
d
n
e
u
r
a
l n
etwo
r
k
s
.
i)
Un
b
alan
ce
d
class
if
icatio
n
B
y
u
s
in
g
th
e
two
m
o
d
els
o
f
WE
(
B
io
B
E
R
T
an
d
B
io
AL
B
E
R
T
)
,
d
ata
wh
e
n
th
e
ta
r
g
et
class
es
in
a
d
ataset
ar
e
u
n
e
n
ly
d
is
tr
ib
u
te
d
.
T
ab
le
4
d
elin
ea
tes
th
e
d
i
f
f
er
in
g
n
u
m
b
er
s
o
f
s
wap
p
e
d
an
d
n
o
n
-
s
wa
p
p
ed
p
atien
ts
in
th
is
s
ce
n
ar
io
.
−
B
io
B
er
t:
T
ab
le
5
r
ev
ea
ls
lo
w
p
r
ec
is
io
n
an
d
r
ec
all
v
alu
e
s
f
o
r
class
1
.
Sp
ec
if
ically
,
th
e
SVM
m
o
d
el
d
em
o
n
s
tr
ates
a
m
a
x
im
u
m
p
r
e
cisi
o
n
o
f
0
.
8
3
b
u
t
a
r
elativ
el
y
lo
w
r
ec
all
o
f
0
.
3
6
.
C
o
n
v
e
r
s
ely
,
th
e
Naiv
e
B
ay
es
m
o
d
el
ac
h
iev
es
a
m
ax
im
u
m
r
ec
all
o
f
0
.
8
7
,
b
u
t
its
p
r
e
cisi
o
n
is
ap
p
r
o
x
im
ately
ze
r
o
.
Hen
ce
,
t
h
e
f
in
d
in
g
s
o
f
th
is
ex
p
er
im
e
n
t m
a
y
n
o
t b
e
r
o
b
u
s
t.
T
ab
le
4
.
Statis
tical
in
f
o
r
m
atio
n
ab
o
u
t
u
n
b
alan
ce
d
d
ata
M
o
d
e
l
s
N
o
t
S
w
a
p
p
e
d
S
w
a
p
p
e
d
B
i
o
B
e
r
t
1
6
0
0
82
B
i
o
A
l
b
e
r
t
1
6
6
8
99
T
ab
le
5
.
B
y
p
atien
ts
-
u
n
b
alan
c
ed
-
B
io
B
er
t,
B
io
Alb
er
t r
esu
lts
p
er
class
1
No
tab
ly
,
h
ig
h
ac
cu
r
ac
y
is
o
b
s
er
v
ed
f
o
r
th
e
"
n
o
t
s
wap
p
ed
"
cl
ass
,
wh
ich
is
ex
p
ec
ted
g
iv
e
n
it
s
m
ajo
r
ity
r
ep
r
esen
tatio
n
i
n
th
e
d
ataset.
Mo
s
t m
o
d
els ex
h
i
b
it e
x
ce
llen
t
o
u
tco
m
es,
with
p
r
ec
is
io
n
an
d
r
ec
all
s
co
r
es a
r
o
u
n
d
0
.
9
7
a
n
d
0
.
9
8
,
r
esp
ec
tiv
ely
.
−
B
io
Alb
er
t:
I
n
c
o
n
tr
ast,
t
h
e
r
esu
lts
f
o
r
B
io
Alb
er
t
m
o
d
els
in
d
ica
te
r
elativ
ely
p
o
o
r
p
r
ec
is
io
n
an
d
r
ec
all
v
alu
es
ac
r
o
s
s
th
e
b
o
ar
d
.
Fo
r
in
s
tan
ce
,
th
e
r
an
d
o
m
f
o
r
est
m
o
d
el
a
ch
iev
es
a
m
ax
im
u
m
p
r
ec
is
io
n
o
f
0
.
6
1
a
n
d
a
r
ec
all
o
f
0
.
2
2
.
Similar
ly
,
n
aiv
e
B
ay
es
ac
h
iev
es
a
m
ax
im
u
m
r
ec
all
o
f
0
.
8
7
b
u
t
m
in
i
m
al
p
r
ec
is
io
n
.
Nev
er
th
eless
,
th
e
in
clu
s
io
n
o
f
class
0
co
n
tr
ib
u
tes to
h
ig
h
ac
c
u
r
ac
y
.
−
C
o
m
p
ar
is
o
n
:
Desp
ite
em
p
l
o
y
i
n
g
v
ar
io
u
s
class
if
icatio
n
s
tr
ate
g
ies,
b
o
th
th
e
B
io
B
er
t
an
d
B
io
Alb
er
t
m
o
d
els
y
ield
ed
p
o
o
r
r
esu
lts
o
v
er
all.
Ho
wev
er
,
th
e
B
io
B
er
t
m
o
d
el
d
em
o
n
s
tr
ated
s
lig
h
tly
h
i
g
h
er
ac
cu
r
ac
y
co
m
p
ar
ed
t
o
th
e
B
io
Alb
er
t m
o
d
el,
with
a
m
ar
g
in
al
d
i
f
f
er
en
c
e
o
f
0
.
0
2
in
m
a
x
im
u
m
ac
cu
r
ac
y
b
etwe
en
th
e
two
m
o
d
els.
−
Dis
cu
s
s
io
n
: T
h
e
p
o
o
r
p
er
f
o
r
m
an
ce
o
f
b
o
th
m
o
d
els ca
n
b
e
attr
ib
u
ted
to
t
h
e
im
b
alan
ce
d
is
tr
ib
u
tio
n
o
f
ea
c
h
class
with
in
th
e
d
ataset.
T
h
e
alg
o
r
ith
m
s
d
em
o
n
s
tr
ate
a
te
n
d
en
cy
to
war
d
s
f
r
eq
u
en
t
clas
s
,
with
lo
wer
p
er
f
o
r
m
an
ce
o
n
t
h
e
m
in
o
r
ity
class
.
E
s
s
en
tially
,
th
e
a
lg
o
r
ith
m
s
p
r
io
r
itize
p
r
ed
ictin
g
th
e
m
o
s
t
f
r
eq
u
en
t
class
,
wh
ich
lead
s
to
ex
ce
llen
t
o
v
er
all
r
esu
lts
b
u
t
at
th
e
ex
p
en
s
e
o
f
m
is
class
if
y
in
g
m
i
n
o
r
ity
in
s
tan
ce
s
.
Ho
wev
er
,
it'
s
im
p
o
r
tan
t
to
n
o
t
e
th
at
th
e
m
is
clas
s
if
icatio
n
s
ar
e
n
o
t
n
ec
ess
ar
ily
in
d
icativ
e
o
f
th
e
alg
o
r
ith
m
wea
k
n
ess
es,
b
u
t r
ath
er
s
tem
f
r
o
m
th
e
r
a
r
ity
o
f
i
n
s
tan
ce
s
in
th
e
m
in
o
r
ity
class
.
I
n
o
u
r
e
x
p
er
im
e
n
tatio
n
,
we
in
itially
u
tili
ze
d
7
m
o
d
els
in
th
e
u
n
b
ala
n
ce
d
class
if
icatio
n
s
ettin
g
.
Su
b
s
eq
u
en
tly
,
we
au
g
m
en
ted
th
e
n
u
m
b
er
o
f
m
o
d
els in
th
e
b
ala
n
ce
d
class
if
icatio
n
s
ettin
g
.
ii)
B
alan
ce
d
class
if
icatio
n
I
n
th
is
ex
p
er
im
en
t,
th
e
class
d
is
tr
ib
u
tio
n
is
eq
u
al
(
8
2
s
wap
p
e
d
an
d
8
2
n
o
t
s
wap
p
ed
p
atien
ts
)
,
u
tili
z
in
g
(
B
io
B
er
t a
n
d
B
io
Alb
er
t)
with
r
esu
lts
in
T
ab
le
6
.
B
i
o
A
l
b
e
r
t
B
i
o
B
e
r
t
M
o
d
e
l
A
c
c
u
r
a
c
y
(
o
v
e
r
t
h
e
t
w
o
c
l
a
sses)
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
M
o
d
e
l
A
c
c
u
r
a
c
y
(
o
v
e
r
t
h
e
t
w
o
c
l
a
sses)
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
A
d
a
B
o
o
st
0
.
9
3
0
.
3
8
0
.
3
5
A
d
a
B
o
o
st
0
.
9
5
0
.
4
8
0
.
5
4
G
r
a
d
i
e
n
t
B
o
o
st
i
n
g
0
.
9
4
0
.
5
9
0
.
2
8
G
r
a
d
i
e
n
t
B
o
o
st
i
n
g
0
.
9
6
0
.
6
1
0
.
5
3
Lo
g
i
s
t
i
c
R
e
g
r
e
ssi
o
n
0
.
9
4
0
.
5
5
0
.
2
9
Lo
g
i
s
t
i
c
R
e
g
r
e
ssi
o
n
0
.
9
6
0
.
6
6
0
.
5
1
N
a
i
v
e
B
a
y
e
s
0
.
4
8
0
.
0
8
0
.
8
7
N
a
i
v
e
B
a
y
e
s
0
.
5
2
0
.
0
8
0
.
8
7
N
e
u
r
a
l
N
e
t
w
o
r
k
0
.
9
4
0
.
4
4
0
.
2
6
N
e
u
r
a
l
N
e
t
w
o
r
k
0
.
9
6
0
.
6
8
0
.
5
1
R
a
n
d
o
m F
o
r
e
s
t
0
.
9
4
0
.
6
1
0
.
2
2
R
a
n
d
o
m F
o
r
e
s
t
0
.
9
6
0
.
6
7
0
.
5
3
S
V
M
S
V
M
0
.
9
6
0
.
8
3
0
.
3
6
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Ma
ch
in
e
lea
r
n
in
g
f
o
r
men
ta
l
h
ea
lth
:
p
r
ed
ictin
g
tr
a
n
s
itio
n
s
fr
o
m
a
d
d
ictio
n
to
illn
ess
(
A
li A
lkh
a
z
r
a
ji)
393
−
B
io
B
er
t:
F
o
r
clas
s
1
,
th
e
g
r
ad
i
en
t
b
o
o
s
tin
g
m
o
d
el
ac
h
iev
es
t
h
e
h
ig
h
est
ac
cu
r
ac
y
o
f
0
.
8
5
an
d
p
r
ec
is
io
n
o
f
0
.
9
0
.
Alth
o
u
g
h
t
h
e
r
ec
all
(
0
.
8
0
)
is
s
lig
h
tly
lo
wer
,
it
s
til
l
co
m
p
ar
es
f
av
o
r
ab
ly
t
o
o
th
er
m
o
d
els,
m
ak
in
g
th
is
m
o
d
el
th
e
s
tan
d
o
u
t
p
e
r
f
o
r
m
er
in
th
e
ex
p
er
im
en
t.
Fo
r
class
0
,
g
r
a
d
ien
t
b
o
o
s
tin
g
also
ex
ce
ls
ac
r
o
s
s
all
m
etr
ics with
Acc
u
r
ac
y
,
Pre
cisi
o
n
,
An
d
R
ec
all
v
alu
es o
f
0
.
8
5
,
0
.
8
2
,
an
d
0
.
9
1
r
esp
ec
tiv
ely
.
−
B
io
Alb
er
t:
F
o
r
class
1
,
Gr
a
d
ien
t
B
o
o
s
tin
g
s
im
ilar
ly
d
e
m
o
n
s
tr
ates
th
e
h
ig
h
est
ac
cu
r
ac
y
(
0
.
8
5
)
an
d
p
r
ec
is
io
n
(
0
.
8
8
)
,
with
r
ec
all
(
0
.
8
0
)
b
ein
g
clo
s
e
b
eh
in
d
.
Ag
ain
,
g
r
ad
ien
t
b
o
o
s
tin
g
em
er
g
es
as
th
e
to
p
-
p
er
f
o
r
m
in
g
m
o
d
el
f
o
r
th
is
clas
s
.
Fo
r
clas
s
0
,
g
r
ad
ien
t b
o
o
s
tin
g
ac
h
iev
es m
ax
im
u
m
v
alu
es with
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
an
d
r
ec
all
o
f
0
.
8
5
,
0
.
8
2
,
an
d
0
.
8
9
r
esp
ec
tiv
ely
.
T
ab
le
6
.
B
y
p
ate
n
ts
-
b
alan
ce
d
-
B
io
b
er
t,
B
io
Alb
er
t r
esu
lts
p
er
class
1
B
i
o
b
e
r
t
B
i
o
A
l
b
e
r
t
M
o
d
e
l
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
M
o
d
e
l
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
A
d
a
B
o
o
st
0
.
7
8
0
.
7
9
0
.
7
6
A
d
a
B
o
o
st
0
.
7
2
0
.
7
5
0
.
6
7
K
N
N
0
.
5
2
0
.
5
2
0
.
5
3
K
N
N
0
.
4
9
0
.
4
9
0
.
5
0
G
r
a
d
i
e
n
t
B
o
o
st
i
n
g
0
.
8
5
0
.
9
0
0
.
8
0
G
r
a
d
i
e
n
t
B
o
o
st
i
n
g
0
.
8
5
0
.
8
8
0
.
8
0
Lo
g
i
s
t
i
c
R
e
g
r
e
ssi
o
n
0
.
8
1
0
.
8
1
0
.
8
1
Lo
g
i
s
t
i
c
R
e
g
r
e
ssi
o
n
0
.
8
1
0
.
8
2
0
.
7
8
N
a
i
v
e
B
a
y
e
s
0
.
7
2
0
.
6
7
0
.
8
5
N
a
i
v
e
B
a
y
e
s
0
.
7
1
0
.
6
7
0
.
8
2
N
e
u
r
a
l
N
e
t
w
o
r
k
0
.
8
2
0
.
8
4
0
.
8
0
N
e
u
r
a
l
N
e
t
w
o
r
k
0
.
7
3
0
.
7
2
0
.
7
6
R
a
n
d
o
m F
o
r
e
s
t
0
.
8
4
0
.
8
7
0
.
8
1
R
a
n
d
o
m F
o
r
e
s
t
0
.
7
9
0
.
8
1
0
.
7
5
S
V
M
0
.
7
9
0
.
8
0
0
.
7
9
S
V
M
0
.
7
4
0
.
7
3
0
.
7
7
Tr
e
e
0
.
7
8
0
.
7
9
0
.
7
8
Tr
e
e
0
.
7
4
0
.
7
6
0
.
7
0
C
o
m
p
ar
is
o
n
:
B
o
th
B
io
B
er
t
an
d
B
io
Alb
er
t
m
o
d
els
ac
h
iev
e
th
eir
h
ig
h
est
p
er
f
o
r
m
a
n
ce
with
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
class
if
ier
.
W
ith
a
s
lig
h
t
d
if
f
er
en
ce
o
f
0
.
0
2
in
p
r
ec
is
io
n
f
o
r
B
io
B
er
t.
Ho
wev
e
r
,
o
th
er
m
o
d
els
s
h
o
w
lar
g
er
d
is
cr
ep
a
n
cies,
with
d
if
f
er
en
ce
s
o
f
ar
o
u
n
d
0
.
1
.
Dis
cu
s
s
io
n
:
W
h
ile
th
e
r
esu
lts
in
th
is
p
h
ase
ar
e
s
o
m
ewh
at
ac
ce
p
tab
le,
th
ey
a
r
e
n
o
t
o
p
tim
al
.
T
h
is
ca
n
b
e
attr
ib
u
ted
to
th
e
s
m
all
s
ize
o
f
th
e
d
ataset
. W
ith
o
n
ly
8
2
p
atien
ts
f
o
r
ea
ch
class
,
wh
ich
m
ay
n
o
t b
e
s
u
f
f
icien
t
f
o
r
tr
ain
in
g
m
o
d
els to
ac
h
iev
e
h
ig
h
ac
cu
r
ac
y
.
Sectio
n
d
is
cu
s
s
io
n
:
C
o
m
p
ar
in
g
p
r
ec
is
io
n
an
d
r
ec
all
b
etwe
en
th
e
im
b
ala
n
ce
d
a
n
d
b
alan
ce
d
class
if
icatio
n
s
r
ev
ea
ls
a
s
ig
n
if
i
ca
n
t
ad
v
an
tag
e
f
o
r
th
e
b
alan
ce
d
class
if
icatio
n
.
I
n
m
o
s
t
m
o
d
els,
th
e
d
if
f
er
en
ce
is
s
u
b
s
tan
tial,
ex
ce
p
t
f
o
r
SVM
an
d
n
aiv
e
B
ay
es,
wh
er
e
th
e
u
n
b
alan
ce
d
class
if
icatio
n
o
u
tp
e
r
f
o
r
m
s
in
p
r
ec
is
io
n
an
d
r
ec
all
r
esp
ec
tiv
ely
.
T
h
is
d
is
cr
ep
an
cy
is
p
r
im
ar
ily
d
u
e
to
th
e
s
u
b
s
tan
tial
cla
s
s
im
b
al
an
ce
in
th
e
d
ataset,
wh
ich
im
p
ac
ts
th
e
ac
cu
r
ac
y
o
f
th
e
m
o
d
els,
as d
is
cu
s
s
ed
p
r
ev
io
u
s
ly
.
b)
B
y
v
is
its
Her
e,
in
th
is
ex
p
er
im
en
t,
we
u
t
ilized
th
e
s
ec
o
n
d
d
er
iv
ed
d
ataset
o
r
g
an
ized
b
y
v
is
its
.
C
la
s
s
if
i
ca
tio
n
was
co
n
d
u
cte
d
s
o
lely
o
n
b
alan
ce
d
d
ata,
in
co
r
p
o
r
atin
g
f
ea
tu
r
es
e
x
tr
ac
ted
f
r
o
m
B
io
B
er
t
an
d
B
io
Alb
er
t.
T
h
e
n
u
m
b
er
o
f
v
is
its
in
ea
ch
f
ile
is
d
ep
icte
d
in
T
ab
le
7.
‒
B
io
B
er
t:
I
n
T
ab
le
8
,
f
o
r
class
1
,
g
r
ad
ie
n
t
b
o
o
s
tin
g
o
u
tp
er
f
o
r
m
s
all
o
th
er
m
o
d
els
b
y
ac
h
iev
in
g
m
ax
im
u
m
v
alu
es a
cr
o
s
s
all
m
etr
ics
: 0
.
9
8
in
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
an
d
r
e
ca
ll.
Similar
ly
,
f
o
r
class
0
,
g
r
ad
ien
t b
o
o
s
tin
g
attain
s
th
e
b
est r
esu
lts
with
m
ax
im
u
m
v
al
u
es
ac
r
o
s
s
all
m
etr
ics:
0
.
9
8
f
o
r
all.
‒
B
io
Alb
er
t:
Acc
o
r
d
in
g
to
th
e
r
esu
lts
in
T
ab
le
8
,
f
o
r
class
1
,
g
r
ad
ien
t
b
o
o
s
tin
g
s
u
r
p
ass
es
all
o
th
er
m
o
d
els
b
y
ac
h
iev
in
g
m
a
x
im
u
m
v
alu
es
in
all
m
etr
ics
:
0
.
9
9
in
ac
cu
r
ac
y
an
d
p
r
ec
is
io
n
,
an
d
0
.
9
8
in
r
e
ca
ll.
L
ik
ewise,
f
o
r
class
0
,
g
r
ad
ie
n
t
b
o
o
s
tin
g
a
ttain
s
th
e
h
ig
h
est
v
al
u
es:
0
.
9
9
i
n
ac
cu
r
ac
y
an
d
r
ec
all,
an
d
0
.
9
8
in
p
r
ec
is
io
n
.
T
ab
le
7
.
S
tatis
tical
in
f
o
r
m
atio
n
ab
o
u
t
b
y
v
is
its
d
ata
M
o
d
e
l
s
V
i
si
t
s
B
i
o
B
e
r
t
2
3
0
0
B
i
o
a
l
b
e
r
t
2
2
9
8
T
ab
le
8
.
B
y
v
is
its
-
b
alan
ce
d
-
B
i
o
B
er
t,
B
io
Alb
er
t r
esu
lts
p
er
class
1
B
i
o
B
e
r
t
B
i
o
A
l
b
e
r
t
M
o
d
e
l
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
M
o
d
e
l
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
A
d
a
B
o
o
st
0
.
9
4
0
.
9
1
0
.
9
7
A
d
a
B
o
o
st
0
.
9
4
0
.
9
1
0
.
9
7
K
N
N
0
.
8
6
0
.
8
2
0
.
9
3
K
N
N
0
.
8
6
0
.
8
1
0
.
9
4
G
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
0
.
9
8
0
.
9
8
0
.
9
7
G
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
0
.
9
9
0
.
9
9
0
.
9
8
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
0
.
9
2
0
.
9
1
0
.
9
4
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
0
.
8
3
0
.
8
6
0
.
7
9
N
a
i
v
e
B
a
y
e
s
0
.
7
2
0
.
7
1
0
.
7
3
N
a
i
v
e
B
a
y
e
s
0
.
7
1
0
.
7
2
0
.
6
7
N
e
u
r
a
l
n
e
t
w
o
r
k
0
.
9
5
0
.
9
5
0
.
9
6
N
e
u
r
a
l
n
e
t
w
o
r
k
0
.
9
5
0
.
9
5
0
.
9
6
R
a
n
d
o
m
f
o
r
e
s
t
0
.
9
6
0
.
9
6
0
.
9
5
R
a
n
d
o
m
f
o
r
e
s
t
0
.
9
6
0
.
9
7
0
.
9
5
S
V
M
0
.
5
0
0
.
5
0
0
.
3
0
S
V
M
0
.
5
5
0
.
6
2
0
.
2
6
Tr
e
e
0
.
9
3
0
.
9
2
0
.
9
4
Tr
e
e
0
.
9
3
0
.
9
1
0
.
9
6
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
385
-
3
9
6
394
C
o
m
p
ar
is
o
n
:
I
n
th
is
ex
p
er
im
e
n
t,
two
m
o
d
els
n
a
m
ely
B
io
Al
b
er
t
an
d
B
io
B
er
t
wer
e
test
ed
t
o
m
ea
s
u
r
e
th
eir
ac
cu
r
ac
y
.
T
h
e
r
esu
lts
s
h
o
w
th
at
B
io
Alb
er
t
m
o
d
el
s
lig
h
tly
o
u
tp
er
f
o
r
m
s
B
io
B
er
t.
W
ith
a
m
ar
g
in
al
d
if
f
er
en
ce
o
f
0
.
0
1
b
etwe
en
th
eir
m
a
x
im
u
m
ac
cu
r
ac
ies.
Sectio
n
Dis
cu
s
s
io
n
:
T
h
e
ac
h
ie
v
ed
r
esu
lts
ar
e
v
er
y
p
r
o
m
is
in
g
.
W
ith
g
r
ad
ien
t
b
o
o
s
tin
g
atta
in
in
g
h
ig
h
ac
cu
r
ac
y
s
co
r
es
clo
s
e
to
1
.
T
h
is
s
u
cc
ess
ca
n
b
e
attr
ib
u
ted
to
th
e
b
alan
ce
d
n
atu
r
e
o
f
th
e
d
at
aset
an
d
th
e
wea
lth
o
f
p
atien
t in
f
o
r
m
atio
n
av
ailab
l
e.
4.
DIS
CU
SS
I
O
N
W
e
co
n
d
u
cted
ex
p
e
r
im
en
ts
o
n
d
atasets
o
r
g
an
ized
b
o
th
"b
y
P
atien
ts
"
an
d
"b
y
Vis
its
,
"
with
v
ar
iatio
n
s
in
class
b
alan
ce
.
W
h
ile
th
e
r
es
u
lts
f
o
r
u
n
b
alan
ce
d
d
ata
wer
e
p
o
o
r
d
u
e
to
th
e
s
ca
r
city
o
f
clas
s
1
in
s
tan
ce
s
,
th
o
s
e
f
o
r
b
alan
ce
d
d
ata
wer
e
ac
ce
p
t
ab
le
b
u
t
n
o
t
o
p
tim
al.
T
h
e
lim
ited
d
ataset
s
ize,
with
o
n
ly
8
2
p
atien
ts
p
er
class
,
m
ay
h
av
e
h
i
n
d
er
e
d
th
e
m
o
d
el
s
'
ab
ilit
y
to
ac
h
iev
e
h
ig
h
er
ac
cu
r
ac
y
.
Su
b
s
eq
u
en
tly
,
f
o
cu
s
in
g
o
n
th
e
d
etails
o
f
ea
ch
p
atien
t
v
is
it
in
th
e
"b
y
Vis
its
"
ap
p
r
o
ac
h
y
ield
ed
n
o
ta
b
ly
h
ig
h
an
d
n
ea
r
ly
p
er
f
ec
t
r
esu
lts
,
s
u
r
p
ass
in
g
th
e
"b
y
Patien
ts
"
ap
p
r
o
ac
h
.
T
h
is
em
p
h
asizes th
e
ef
f
icac
y
o
f
co
n
s
id
er
in
g
v
is
it
-
lev
el
d
etails f
o
r
p
r
ed
i
ctiv
e
p
u
r
p
o
s
es.
R
eg
ar
d
in
g
th
e
B
io
B
er
t
an
d
B
io
Alb
er
t
m
o
d
els,
b
o
th
e
x
h
ib
ited
co
m
p
ar
ab
le
p
er
f
o
r
m
an
ce
,
with
B
io
Alb
er
t
s
lig
h
tly
ed
g
i
n
g
a
h
e
ad
in
th
e
"
b
y
Vis
its
"
ap
p
r
o
ac
h
.
Ho
wev
er
,
in
p
r
e
v
io
u
s
e
x
p
er
i
m
en
ts
,
B
io
B
er
t
h
ad
a
s
lig
h
t
ad
v
an
tag
e
in
s
ev
er
al
in
s
tan
ce
s
.
Ov
er
all,
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
two
m
o
d
els
was
ap
p
r
o
x
im
ately
eq
u
al
.
Sev
er
al
class
if
ier
s
wer
e
em
p
l
o
y
ed
in
t
h
e
class
if
icatio
n
p
h
a
s
e,
with
g
r
a
d
ien
t
b
o
o
s
tin
g
co
n
s
is
ten
tly
ac
h
iev
in
g
h
ig
h
s
co
r
es
ac
r
o
s
s
m
o
s
t
ex
p
er
i
m
en
ts
.
T
h
is
s
u
g
g
ests
th
at
g
r
ad
i
en
t
b
o
o
s
tin
g
is
th
e
m
o
s
t
ef
f
icie
n
t
cla
s
s
if
ier
am
o
n
g
th
o
s
e
u
tili
ze
d
wh
en
X
ax
is
r
ep
r
esen
ts
n
o
.
o
f
p
atien
ts
an
d
Y
a
x
is
r
ep
r
esen
ts
n
o
.
o
f
r
esu
lts
as sh
o
wn
in
F
ig
u
r
e
4
.
Fig
u
r
e
4
.
Fin
al
r
esu
lts
5.
CO
NCLU
SI
O
N
AND
F
U
T
U
RE
WO
RK
S
Ou
r
r
es
ea
r
c
h
f
o
c
u
s
e
d
o
n
d
is
ea
s
e
p
r
ed
ict
io
n
wit
h
i
n
t
h
e
d
o
m
ai
n
o
f
a
d
d
i
cti
o
n
.
W
e
co
lle
ct
ed
r
ea
l
d
at
a
f
r
o
m
a
h
o
s
p
it
al
in
I
r
aq
-
B
a
g
h
d
a
d
,
co
m
p
r
is
in
g
p
r
o
f
iles
o
f
p
a
tie
n
ts
wit
h
v
a
r
i
o
u
s
a
il
m
e
n
ts
.
H
o
we
v
er
,
o
u
r
s
t
u
d
y
s
p
e
ci
f
ic
all
y
t
ar
g
ete
d
p
a
tie
n
ts
wit
h
a
d
d
ic
ti
o
n
is
s
u
e
s
.
T
h
e
o
b
ject
iv
e
w
as
t
o
p
r
e
d
ic
t
i
n
s
ta
n
c
es
o
f
r
e
la
p
s
e
u
s
i
n
g
v
a
r
i
o
u
s
NL
P
m
o
d
els
f
o
r
te
x
t
an
al
y
s
is
a
n
d
a
r
a
n
g
e
o
f
cl
ass
i
f
ier
s
f
o
r
p
r
e
d
i
cti
o
n
.
W
e
d
iv
id
e
d
t
h
e
d
a
ta
i
n
t
o
t
wo
s
am
p
les
:
"b
y
Pat
ie
n
ts
"
a
n
d
"
b
y
Vis
its
.
"
T
h
e
"
b
y
Pat
ie
n
ts
"
v
er
s
i
o
n
lac
k
ed
d
et
ails
r
e
g
a
r
d
i
n
g
in
d
i
v
i
d
u
a
l
p
at
ie
n
t
v
is
i
ts
,
u
n
li
k
e
t
h
e
l
att
e
r
.
W
e
e
x
p
l
o
r
e
d
b
o
t
h
b
al
an
ce
d
a
n
d
u
n
b
a
lan
ce
d
d
at
asets
i
n
t
h
e
"b
y
P
ati
en
ts
"
v
er
s
io
n
,
w
h
il
e
ex
cl
u
s
i
v
el
y
f
o
cu
s
in
g
o
n
b
al
a
n
c
ed
d
at
a
f
o
r
th
e
"
b
y
V
is
its
"
d
a
tas
et.
P
r
e
p
r
o
ce
s
s
i
n
g
s
ta
g
es
w
e
r
e
e
m
p
lo
y
ed
t
o
p
r
ep
ar
e
th
e
d
ata
s
ets
f
o
r
te
x
t
c
lass
i
f
i
ca
t
io
n
.
W
e
d
is
c
u
s
s
e
d
d
i
f
f
er
e
n
t
WE
te
ch
n
i
q
u
es
a
n
d
u
l
ti
m
at
el
y
s
e
lect
e
d
B
io
B
er
t
a
n
d
B
io
Al
b
er
t
f
o
r
f
ea
tu
r
e
e
x
t
r
ac
ti
o
n
.
S
u
b
s
e
q
u
e
n
tl
y
,
c
lass
i
f
i
er
m
o
d
els
w
er
e
tr
ai
n
e
d
u
s
in
g
c
r
o
s
s
-
v
al
id
ati
o
n
te
ch
n
i
q
u
e
(5
-
f
o
ld
s
)
f
o
r
t
h
e
p
r
e
d
i
cti
o
n
p
r
o
c
ess
.
I
n
th
e
"
b
y
Pat
ie
n
ts
"
d
at
aset
,
u
n
b
a
la
n
c
ed
d
a
ta
y
ie
l
d
e
d
u
n
f
av
o
r
a
b
l
e
r
esu
lts
,
wh
e
r
e
as
t
h
e
b
a
la
n
c
ed
d
a
tase
t
p
er
f
o
r
m
e
d
s
i
g
n
if
ic
an
tl
y
b
ett
er
.
Ho
w
ev
er
,
w
h
e
n
c
o
m
p
a
r
e
d
t
o
t
h
e
"b
y
V
is
its
"
d
atas
et,
th
e
l
att
er
e
x
h
i
b
it
ed
s
u
p
er
io
r
p
e
r
f
o
r
m
a
n
ce
.
B
o
t
h
B
io
B
er
t
a
n
d
B
io
Al
b
er
t
m
o
d
e
ls
ac
h
ie
v
ed
s
i
m
il
ar
r
esu
lts
.
A
m
o
n
g
th
e
cl
ass
i
f
ie
r
s
test
ed
,
g
r
a
d
i
en
t
b
o
o
s
ti
n
g
e
m
e
r
g
e
d
as
t
h
e
m
o
s
t
ef
f
ic
ie
n
t
m
o
d
el
.
Ul
ti
m
at
el
y
,
o
u
r
p
r
o
p
o
s
e
d
p
r
e
d
ic
ti
o
n
m
o
d
el
ca
n
a
id
m
e
d
i
ca
l
p
r
o
f
ess
io
n
als
i
n
a
n
t
ici
p
ati
n
g
th
e
li
k
el
i
h
o
o
d
o
f
a
n
a
d
d
ic
t
r
ela
p
s
i
n
g
.
T
h
is
a
p
p
r
o
ac
h
o
f
f
e
r
s
th
e
p
o
te
n
t
ial
t
o
i
n
te
r
v
e
n
e
a
n
d
s
u
p
p
o
r
t
i
n
d
iv
id
u
als
s
tr
u
g
g
li
n
g
w
ith
a
d
d
i
cti
o
n
,
t
h
u
s
m
it
ig
ati
n
g
t
h
e
r
is
k
o
f
d
ev
el
o
p
i
n
g
f
u
r
t
h
er
ill
n
ess
es
.
N
u
m
e
r
o
u
s
p
o
te
n
t
ial
a
v
e
n
u
es
f
o
r
f
u
t
u
r
e
r
es
ea
r
c
h
w
er
e
id
en
ti
f
ie
d
,
i
n
c
l
u
d
i
n
g
e
x
p
e
r
i
m
e
n
ts
wit
h
d
i
f
f
er
e
n
t
N
L
P
m
o
d
els
,
c
o
n
s
i
d
er
i
n
g
p
ati
e
n
t
v
is
its
as
s
e
q
u
e
n
t
ial
d
at
a,
r
e
f
i
n
i
n
g
t
h
e
p
r
o
d
u
c
e
d
e
m
b
e
d
d
in
g
s
,
a
n
d
in
c
o
r
p
o
r
a
ti
n
g
a
d
d
it
io
n
al
f
e
at
u
r
es.
Ou
r
t
h
esis
la
y
s
t
h
e
g
r
o
u
n
d
wo
r
k
f
o
r
f
u
r
t
h
e
r
ex
p
l
o
r
at
io
n
i
n
th
is
ar
ea
,
p
r
o
v
i
d
i
n
g
v
al
u
ab
le
t
o
o
ls
f
o
r
a
n
a
ly
zi
n
g
d
i
v
e
r
s
e
d
is
e
ases
u
s
in
g
r
ea
l
-
w
o
r
l
d
d
atas
ets.
Evaluation Warning : The document was created with Spire.PDF for Python.