I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
24
,
p
p
.
6
1
2
9
~
6
1
3
9
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/
ijece
.
v
14
i
6
.
pp
6
1
2
9
-
6
1
3
9
6129
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
A
no
v
el
a
ppro
a
ch
for
imba
la
nced
i
nsta
nce
ha
ndling
towa
rd
bett
e
r
pre
term
bi
rth
cla
ss
ifica
tion
H
im
a
ni
Deshp
a
nd
e
1
,
L
ee
na
Ra
g
ha
2
1
D
e
p
a
r
t
me
n
t
of
A
r
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
D
a
t
a
S
c
i
e
n
c
e
,
T
h
a
d
o
ma
l
S
h
a
h
a
n
i
En
g
i
n
e
e
r
i
n
g
C
o
l
l
e
g
e
,
M
u
m
b
a
i
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
of
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
B
LD
EA
's V
.
P
.
D
r
.
P
.
G
.
H
a
l
a
k
a
t
t
i
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
Te
c
h
n
o
l
o
g
y
,
V
i
j
a
y
a
p
u
r
,
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
Ap
r
1
7
,
2
0
2
4
R
ev
is
ed
J
u
l 1
4
,
2
0
2
4
Acc
ep
ted
Au
g
6
,
2
0
2
4
P
re
term
b
irt
h
(P
TB)
is
a
m
a
jo
r
c
a
u
se
of
c
h
il
d
a
n
d
m
o
th
e
r
m
o
rtali
ty
,
a
P
TB
c
las
sifica
ti
o
n
m
o
d
e
l
can
a
ss
ist
in
a
ss
e
ss
in
g
th
e
h
e
a
lt
h
c
o
n
d
it
io
n
a
h
e
a
d
of
ti
m
e
a
n
d
h
e
lp
a
v
o
id
c
o
m
p
li
c
a
ti
o
n
s
d
u
rin
g
c
h
il
d
b
irt
h
.
M
o
th
e
r’s
s
ig
n
ifi
c
a
n
t
fe
a
tu
re
(M
S
F
)
d
a
tas
e
t
c
re
a
ted
fo
r
t
h
is
stu
d
y
h
a
s
fe
a
tu
re
s
d
e
ri
v
e
d
fr
o
m
m
o
th
e
r’s
p
h
y
sic
a
l,
li
fe
sty
le,
s
o
c
ial
a
n
d
stre
ss
a
tt
rib
u
tes
.
M
S
F
d
a
tas
e
t
c
o
n
sists
of
1
1
9
fe
a
tu
re
s
of
1
,
0
0
0
m
o
th
e
rs
with
172
p
re
term
a
n
d
8
2
8
fu
ll
-
term
d
e
li
v
e
ries
,
re
su
lt
i
n
g
in
issu
e
s
of
d
a
tas
e
t
imb
a
lan
c
e
n
a
m
e
ly
c
las
s
in
se
p
a
ra
b
il
it
y
a
n
d
c
las
sifica
ti
o
n
b
ias
.
To
o
v
e
rc
o
m
e
t
h
e
imb
a
lan
c
e
issu
e
,
a
n
o
v
e
l
a
l
g
o
rit
h
m
n
a
m
e
d
m
a
jo
rit
y
p
e
n
a
li
z
in
g
m
in
o
ri
ty
u
p
sa
m
p
li
n
g
(
M
P
M
U)
is
p
ro
p
o
se
d
.
M
P
M
U
f
o
rm
s
c
lu
st
e
rs
lo
o
k
in
g
in
t
o
th
e
d
e
g
re
e
of
d
a
tas
e
t
imb
a
lan
c
e
,
it
a
n
a
ly
se
s
th
e
c
o
m
p
o
siti
o
n
of
each
c
lu
ste
r
in
d
iv
i
d
u
a
ll
y
a
n
d
c
o
m
p
u
tes
th
e
v
a
rie
d
p
e
n
a
lt
y
fo
r
m
a
jo
rit
y
c
las
s
in
sta
n
c
e
s.
It
fu
rt
h
e
r
b
a
lan
c
e
s
d
a
tas
e
t
c
o
m
p
o
siti
o
n
by
o
v
e
rsa
m
p
li
n
g
m
in
o
rit
y
c
las
s
in
sta
n
c
e
s.
MPMU
p
ro
c
e
ss
e
d
d
a
tas
e
t
is
fu
rth
e
r
u
se
d
to
train
t
h
e
p
r
o
p
o
se
d
6L
-
AN
N
n
e
two
rk
wh
ich
fi
n
d
s
t
h
e
p
r
o
b
a
b
il
it
y
of
o
c
c
u
rre
n
c
e
of
P
TB
.
Th
e
p
ro
p
o
se
d
m
o
d
e
l
h
a
s
sh
o
wn
e
fficie
n
t
re
su
lt
s
on
M
S
F
su
b
-
d
a
tas
e
ts
with
p
re
c
isio
n
v
a
lu
e
s
ra
n
g
in
g
fro
m
0
.
9
0
to
0
.
9
7
,
a
re
a
u
n
d
e
r
t
h
e
c
u
rv
e
(
AUC
)
b
e
twe
e
n
0
.
8
6
to
0
.
9
9
,
a
n
d
p
re
d
ictio
n
a
c
c
u
ra
c
y
ra
n
g
i
n
g
fr
o
m
9
3
.
0
4
%
to
9
9
.
4
7
%
.
Ex
p
e
rime
n
t
re
su
lt
s
sh
o
w
th
a
t
a
m
o
t
h
e
r’s
li
fe
sty
le
a
n
d
stre
ss
fe
a
tu
re
s
h
a
v
e
a
stro
n
g
in
flu
e
n
c
e
on
th
e
c
h
il
d
b
irt
h
o
u
tc
o
m
e
.
K
ey
w
o
r
d
s
:
C
las
s
im
b
alan
ce
C
lu
s
ter
in
g
Ma
jo
r
ity
class
Min
o
r
ity
class
Ov
er
s
am
p
lin
g
Pen
alty
T
h
is
is
an
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r
th
e
CC
BY
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Him
an
i
Desh
p
an
d
e
Dep
ar
tm
en
t
of
Ar
tific
ial
I
n
tellig
en
ce
an
d
Data
Scien
ce
,
T
h
a
d
o
m
al
Sh
ah
a
n
i
E
n
g
in
ee
r
in
g
C
o
lleg
e
B
an
d
r
a
(
w)
,
Mu
m
b
ai,
Ma
h
a
r
a
s
h
tr
a,
I
n
d
ia
E
m
ail:
h
im
an
i.d
esh
p
a
n
d
e@
th
ad
o
m
al.
o
r
g
1.
I
NT
RO
D
UCT
I
O
N
Ma
ter
n
al
h
ea
lth
an
d
ch
ild
b
ir
th
ex
p
er
ie
n
ce
h
a
v
e
alwa
y
s
b
e
en
a
p
r
io
r
ity
co
n
ce
r
n
ac
r
o
s
s
th
e
g
lo
b
e.
T
h
er
e
a
r
e
m
a
n
y
co
m
p
licatio
n
s
th
at
m
o
th
er
an
d
ch
ild
can
go
th
r
o
u
g
h
d
u
r
in
g
c
h
ild
b
ir
th
,
o
n
e
s
u
ch
c
o
m
p
licatio
n
is
wh
en
th
e
ch
ild
b
ir
th
tak
es
p
lace
b
ef
o
r
e
37
wee
k
s
of
p
r
e
g
n
a
n
cy
wh
ich
is
ter
m
ed
p
r
eter
m
b
ir
th
(
PTB).
PTB
is
a
lead
in
g
ca
u
s
e
of
u
n
d
er
d
ev
e
lo
p
ed
b
a
b
y
o
r
g
an
s
,
m
ater
n
al
m
o
r
tality
an
d
c
h
ild
m
o
r
tality
[
1
]
.
In
to
d
a
y
’
s
er
a
with
r
ap
id
s
o
cieta
l
an
d
life
s
ty
le
ch
an
g
es,
th
e
PTB
ca
s
es
ar
e
in
cr
ea
s
in
g
r
ap
i
d
ly
.
A
PTB
class
if
icatio
n
m
o
d
el
can
h
elp
to
ass
ess
th
e
h
ea
lth
c
o
n
d
itio
n
of
a
wo
m
a
n
ah
ea
d
of
tim
e
to
p
r
ev
e
n
t
an
y
f
u
tu
r
e
h
e
alth
co
m
p
licatio
n
s
.
Pre
g
n
an
cy
o
u
tco
m
es
ar
e
clo
s
ely
ass
o
ciate
d
with
th
e
h
ea
lth
of
th
e
m
o
th
er
[
2
]
.
A
lar
g
e
n
u
m
b
er
of
h
ea
lth
is
s
u
es
th
at
wo
m
en
f
ac
e
is
ass
o
ciate
d
with
th
eir
life
s
ty
le,
r
elatio
n
s
h
ip
with
f
am
ily
,
s
o
cial
s
u
r
r
o
u
n
d
in
g
s
,
s
tr
ess
lev
el,
an
d
p
h
y
s
ical
h
ea
lth
.
All
th
ese
f
ac
to
r
s
in
f
l
u
en
ce
t
h
eir
r
e
p
r
o
d
u
ctiv
e
h
ea
lth
as
well
[
3
]
an
d
h
en
ce
in
f
o
r
m
atio
n
ass
o
ciate
d
with
th
e
m
o
th
er
’
s
m
en
tal
an
d
p
h
y
s
ical
h
ea
lth
can
be
u
s
ed
f
o
r
th
e
p
r
o
g
n
o
s
is
of
PTB.
E
x
p
lo
r
in
g
m
ac
h
in
e
lear
n
in
g
m
et
h
o
d
s
an
d
to
o
ls
on
th
e
m
o
th
er
’
s
h
ea
lth
d
ataset
can
be
of
g
r
ea
t
h
el
p
to
war
d
s
f
in
d
in
g
u
s
ef
u
l
p
atter
n
s
a
n
d
f
ac
ts
f
o
r
a
b
etter
u
n
d
er
s
tan
d
in
g
of
PTB.
C
o
n
s
id
er
in
g
th
e
s
ize
of
t
h
e
p
o
p
u
latio
n
,
m
ed
ica
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
1
2
9
-
6
1
3
9
6130
d
o
m
ain
d
atasets
ar
e
f
o
u
n
d
to
be
h
ig
h
ly
im
b
ala
n
ce
d
with
f
ewe
r
in
s
tan
ce
s
of
d
is
ea
s
ed
p
atien
ts
[
4
]
.
L
ac
k
of
d
en
s
ity
in
im
b
alan
ce
d
d
atasets
f
r
eq
u
en
tly
r
esu
lts
in
lear
n
in
g
m
o
d
els
f
ailin
g
to
d
is
co
v
e
r
u
n
u
s
u
al
p
atter
n
s
an
d
h
av
in
g
class
if
icatio
n
b
ias
to
war
d
s
th
e
m
ajo
r
ity
class
,
th
u
s
r
esu
ltin
g
in
m
in
o
r
ity
clas
s
m
is
clas
s
if
icatio
n
[
5
]
,
[
6
]
.
T
h
is
m
ak
es
im
b
ala
n
c
ed
d
ata
h
a
n
d
lin
g
a
h
i
g
h
-
p
r
io
r
i
ty
is
s
u
e
th
at
n
ee
d
s
to
be
ad
d
r
ess
ed
f
o
r
ef
f
icien
t
class
if
icatio
n
of
h
ea
lth
co
n
d
itio
n
s
.
Ob
s
er
v
in
g
t
h
e
r
ec
en
t
wo
r
k
in
th
e
m
ater
n
al,
ch
ild
b
ir
t
h
d
o
m
ain
an
d
s
p
ec
if
ically
in
t
h
e
PT
B
d
o
m
ain
[
7
]
,
f
ir
s
tly
it
was
o
b
s
er
v
e
d
th
at
m
ajo
r
ly
t
h
e
r
esear
ch
e
r
s
an
d
ex
p
e
r
ts
ar
e
f
r
o
m
th
e
m
e
d
ical
d
o
m
ain
an
d
th
e
y
h
av
e
f
o
cu
s
ed
on
PTB
class
if
i
ca
tio
n
m
o
d
els
by
u
s
in
g
ex
is
ti
n
g
s
tatis
tical
m
eth
o
d
o
lo
g
ies
an
d
h
av
e
n
o
t
f
o
cu
s
ed
on
d
ataset
en
h
an
ce
m
en
t
tech
n
iq
u
es
an
d
m
o
d
er
n
h
e
u
r
is
tic
class
if
icatio
n
m
eth
o
d
s
.
Seco
n
d
ly
,
th
er
e
is
no
s
tan
d
ar
d
d
ataset
an
d
r
esear
ch
e
r
s
ar
e
r
ely
in
g
on
s
m
all
lo
ca
l
d
atasets
w
ith
m
ajo
r
ity
d
ata
co
n
s
is
tin
g
of
o
b
s
tetr
ic
f
ea
tu
r
es
an
d
test
r
esu
lts
an
d
h
av
e
not
wo
r
k
e
d
on
th
e
in
f
lu
e
n
cin
g
f
ac
to
r
s
th
at
af
f
ec
t
th
ese
test
r
esu
lt
s
.
To
f
ill
th
is
g
ap
a
d
ataset
titl
ed
m
o
t
h
er
’
s
s
ig
n
if
ican
t
f
ea
tu
r
e
(
MS
F)
is
cr
ea
ted
f
o
r
th
is
s
tu
d
y
,
th
r
o
u
g
h
r
ig
o
r
o
u
s
in
ter
ac
tio
n
with
g
y
n
ec
o
lo
g
is
t
s
with
th
e
aim
to
d
er
iv
e
th
e
in
f
lu
en
cin
g
f
ea
tu
r
es
d
ep
ictin
g
m
o
th
er
’
s
p
h
y
s
ical
h
ea
lth
,
life
s
ty
le,
s
o
cial
s
u
r
r
o
u
n
d
in
g
s
an
d
s
tr
ess
.
L
o
o
k
in
g
in
to
th
e
im
b
alan
ce
d
d
ataset
h
an
d
lin
g
m
eth
o
d
s
ac
r
o
s
s
d
if
f
er
en
t
d
o
m
ain
s
,
it
is
o
b
s
er
v
ed
th
at
r
esear
ch
er
s
h
av
e
co
m
e
up
wit
h
d
if
f
e
r
en
t
tec
h
n
iq
u
es
w
h
ich
ar
e
b
r
o
ad
ly
d
iv
id
e
d
in
to
th
r
ee
ca
teg
o
r
ies
n
a
m
ely
p
r
e
-
p
r
o
ce
s
s
in
g
,
co
s
t
-
s
en
s
itiv
e
lear
n
in
g
a
n
d
e
n
s
em
b
le
lear
n
in
g
m
eth
o
d
s
[
8
]
,
[
9
]
.
Pre
-
p
r
o
ce
s
s
in
g
m
eth
o
d
s
wo
r
k
on
s
tr
en
g
th
en
in
g
th
e
m
in
o
r
ity
an
d
s
u
p
p
r
ess
in
g
t
h
e
m
aj
o
r
ity
s
am
p
les
b
ef
o
r
e
class
if
icatio
n
.
T
h
e
d
ataset
ca
n
be
p
r
e
-
p
r
o
ce
s
s
ed
b
ased
on
in
s
tan
ce
s
or
f
ea
tu
r
es.
I
n
s
tan
ce
-
b
ase
d
m
eth
o
d
s
eith
er
o
v
e
r
-
s
am
p
le
m
in
o
r
ity
i
n
s
tan
ce
s
or
u
n
d
er
-
s
am
p
le
m
ajo
r
ity
in
s
tan
ce
s
or
b
o
th
[
1
0
]
–
[
1
2
]
.
Ma
n
y
r
esear
ch
er
s
h
av
e
u
s
ed
clu
s
ter
-
b
ased
f
u
zz
y
weig
h
in
g
[
1
3
]
an
d
a
p
p
r
o
ac
h
e
s
wh
er
e
r
esam
p
lin
g
of
clu
s
ter
s
is
p
r
o
p
o
s
ed
[
1
4
]
,
[
1
5
]
,
t
h
ese
r
esear
ch
er
s
h
av
e
f
o
cu
s
ed
on
c
h
an
g
i
n
g
t
h
e
s
ize
of
clu
s
ter
s
by
o
v
er
-
s
am
p
lin
g
an
d
u
n
d
er
-
s
am
p
lin
g
a
n
d
h
av
e
not
wo
r
k
ed
on
u
n
d
er
s
tan
d
i
n
g
th
e
im
p
ac
t
of
in
s
tan
ce
s
f
allin
g
u
n
d
er
d
if
f
er
en
t
clu
s
ter
s
to
war
d
s
class
s
ep
ar
ab
ilit
y
.
Featu
r
e
s
elec
tio
n
m
eth
o
d
s
p
r
e
-
p
r
o
ce
s
s
d
atasets
by
o
b
s
er
v
i
n
g
f
ea
tu
r
es,
th
ese
m
eth
o
d
s
ar
e
b
r
o
ad
ly
d
iv
id
e
d
in
to
th
r
e
e
ca
teg
o
r
ies,
n
am
el
y
f
ilter
m
eth
o
d
s
,
wr
ap
p
er
m
eth
o
d
s
an
d
em
b
ed
d
ed
m
eth
o
d
s
[
1
6
]
,
[
1
7
]
.
T
h
ese
m
eth
o
d
s
lo
o
k
in
to
th
e
s
tr
en
g
th
of
a
f
ea
tu
r
e
in
d
iv
id
u
all
y
or
in
co
m
b
in
atio
n
with
o
th
er
f
ea
tu
r
es
to
war
d
s
ef
f
icien
t
class
if
icatio
n
.
T
h
e
is
s
u
e
wi
th
p
r
e
-
p
r
o
ce
s
s
in
g
m
eth
o
d
s
is
th
at
it
is
d
if
f
icu
lt
to
d
ec
id
e
on
th
e
ch
an
g
e
in
th
e
d
is
tr
ib
u
tio
n
of
th
e
d
atab
ase
to
s
u
it
th
e
g
iv
en
p
r
o
b
lem
[
8
]
.
C
o
s
t
-
s
en
s
it
iv
e
m
eth
o
d
s
ar
e
u
s
ed
to
s
o
lv
e
ca
teg
o
r
izatio
n
p
r
o
b
lem
s
wh
en
t
h
e
co
s
ts
of
in
co
r
r
ec
t
class
if
icatio
n
d
if
f
er
d
ep
en
d
i
n
g
on
th
e
in
s
tan
ce
s
an
d
f
ea
tu
r
es.
T
h
ese
tech
n
iq
u
es
p
an
elize
to
r
ed
u
ce
lo
s
s
es
f
ac
ed
due
to
m
is
class
if
icatio
n
[
1
8
]
.
As
th
e
co
s
t
is
a
s
s
e
s
s
ed
co
n
s
id
er
in
g
th
e
im
b
alan
ce
in
th
e
d
at
aset,
th
ese
m
eth
o
d
s
ar
e
f
o
u
n
d
to
be
ef
f
icien
t
in
h
an
d
lin
g
im
b
ala
n
ce
d
d
atasets
.
C
o
s
t
can
be
in
co
r
p
o
r
ated
in
to
th
e
class
if
ier
’
s
tr
ain
in
g
p
r
o
ce
s
s
as
a
co
s
t
m
atr
ix
or
th
r
esh
o
ld
.
To
cr
ea
te
a
co
s
t
-
s
en
s
itiv
e
cla
s
s
if
ier
,
r
esear
ch
er
s
h
av
e
f
o
llo
wed
m
eth
o
d
o
l
o
g
ies
lik
e
m
o
d
if
y
in
g
th
e
g
o
al
f
u
n
ctio
n
u
s
in
g
a
weig
h
in
g
s
tr
ateg
y
[
1
9
]
–
[
2
1
]
,
ch
a
n
g
in
g
th
e
d
ec
is
io
n
th
r
esh
o
ld
s
,
th
e
lear
n
in
g
p
r
o
ce
s
s
[
2
2
]
,
t
h
e
en
c
o
d
in
g
n
etwo
r
k
[
2
3
]
,
a
n
d
one
-
class
ap
p
r
o
ac
h
[
2
4
]
,
[
2
5
]
.
Alth
o
u
g
h
th
ese
tech
n
iq
u
es
ar
e
s
im
p
le
to
u
s
e,
o
v
er
f
itti
n
g
p
r
o
b
lem
s
co
u
ld
ar
is
e
[
8
]
.
T
h
e
in
tr
icac
y
of
co
s
t
ca
lcu
latio
n
m
ay
be
one
of
th
e
m
ain
r
ea
s
o
n
s
wh
y
co
s
t
-
s
en
s
itiv
e
ap
p
r
o
ac
h
e
s
ar
e
less
p
o
p
u
lar
th
an
r
esam
p
lin
g
m
eth
o
d
s
,
ev
en
th
o
u
g
h
th
ese
m
eth
o
d
s
ar
e
ef
f
e
ctiv
e
[
2
6
]
.
W
h
en
a
s
in
g
le
class
if
ier
f
ails
to
wo
r
k
ef
f
icien
tly
on
an
im
b
alan
ce
d
d
ataset,
m
u
ltip
le
clas
s
if
ier
s
ar
e
co
n
s
id
er
ed
wh
ile
d
ec
id
in
g
t
h
e
f
in
al
o
u
tco
m
e
[
2
7
]
.
T
h
ese
ty
p
es
of
m
eth
o
d
s
ar
e
ca
lled
en
s
em
b
le
m
eth
o
d
s
.
B
ag
g
in
g
an
d
B
o
o
s
tin
g
a
r
e
two
d
if
f
er
en
t
k
in
d
s
of
e
n
s
em
b
le
ap
p
r
o
ac
h
es
[
2
8
]
.
T
h
e
m
ajo
r
ity
of
en
s
em
b
le
m
o
d
el
s
h
av
e
tak
en
co
s
t
-
s
en
s
itiv
e
an
d
r
esam
p
lin
g
tech
n
iq
u
es
in
to
ac
co
u
n
t
[
2
6
]
.
L
iter
atu
r
e
s
u
g
g
ests
th
at
d
ataset
im
b
alan
ce
can
be
h
an
d
le
d
by
p
r
o
ce
s
s
in
g
d
atasets
or
cu
s
to
m
izin
g
p
r
ed
ictio
n
alg
o
r
ith
m
s
,
th
e
f
o
r
m
er
b
ein
g
a
more
co
m
m
o
n
ch
o
ice
am
o
n
g
r
esear
c
h
er
s
.
T
h
is
r
esear
ch
aim
s
to
o
v
er
c
o
m
e
is
s
u
es
in
th
e
m
ater
n
al
d
o
m
ain
by
cr
ea
tin
g
a
d
ataset
d
ep
icti
n
g
m
o
th
er
’
s
p
h
y
s
ical,
life
s
ty
l
e,
s
o
cial
an
d
s
tr
ess
-
r
elate
d
f
ea
tu
r
es.
Fu
r
th
er
,
a
n
o
v
el
im
b
alan
ce
h
a
n
d
lin
g
m
eth
o
d
is
p
r
o
p
o
s
ed
to
o
v
er
co
m
e
th
e
b
ias
an
d
in
s
ep
ar
ab
ilit
y
is
s
u
es
f
o
u
n
d
in
t
h
e
cr
ea
ted
d
ataset.
2.
M
E
T
H
O
D
R
esear
ch
es
in
th
e
PTB
d
o
m
ai
n
ar
e
b
ased
on
lim
ited
o
b
s
er
v
a
tio
n
s
f
r
o
m
m
ed
ical
r
ec
o
r
d
s
an
d
h
av
e
n
o
t
f
o
cu
s
ed
on
th
e
life
s
ty
le,
s
tr
ess
an
d
s
o
cial
asp
ec
ts
of
a
w
o
m
a
n
’
s
life
th
at
m
a
y
also
c
o
n
tr
ib
u
te
to
h
er
m
ater
n
al
h
ea
lth
.
Fo
r
th
e
s
am
e
r
ea
s
o
n
,
t
h
is
r
esear
ch
d
elv
es
in
to
th
e
ex
am
in
atio
n
of
b
o
th
th
e
m
e
n
tal
an
d
p
h
y
s
ical
h
ea
lth
of
m
o
th
er
s
,
lead
i
n
g
to
th
e
cr
ea
tio
n
of
a
co
r
r
esp
o
n
d
in
g
d
ataset.
Ob
s
er
v
in
g
th
e
cu
r
r
e
n
t
p
o
p
u
l
atio
n
,
it
is
r
ea
lize
d
th
at
th
e
d
ataset
m
ajo
r
l
y
co
n
s
is
t
s
of
f
u
ll
-
ter
m
b
ir
th
(
FTB)
ca
s
es
an
d
f
ew
PTB
ca
s
es,
th
u
s
r
esu
ltin
g
in
im
b
alan
ce
d
d
ataset
f
o
r
m
atio
n
.
T
o
war
d
s
th
is
,
a
n
o
v
el
clu
s
ter
in
g
-
b
ased
alg
o
r
ith
m
n
am
e
d
m
ajo
r
ity
p
en
alizin
g
m
in
o
r
ity
u
p
s
ca
lin
g
(
MPM
U)
is
p
r
o
p
o
s
ed
to
h
an
d
le
th
e
d
ata
im
b
alan
ce
by
f
o
c
u
s
in
g
on
m
i
n
im
izin
g
th
e
im
p
ac
t
of
m
ajo
r
ity
d
o
m
in
an
ce
th
at
m
ay
in
f
l
u
en
ce
th
e
class
if
icatio
n
r
esu
lts
.
Fu
r
th
e
r
,
a
n
eu
r
a
l
n
etwo
r
k
n
am
ed
6L
-
ANN
is
p
r
o
p
o
s
ed
th
at
cla
s
s
if
ies
th
e
in
s
tan
ce
o
u
tco
m
e
b
ein
g
p
r
eter
m
or
f
u
ll
-
ter
m
.
F
ig
u
r
e
1
s
h
o
ws
th
e
ar
ch
itectu
r
e
of
th
e
p
r
o
p
o
s
ed
m
o
d
el.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
imb
a
la
n
ce
d
in
s
ta
n
ce
h
a
n
d
lin
g
to
w
a
r
d
b
etter p
r
eterm
b
ir
th
…
(
Hima
n
i D
esh
p
a
n
d
e
)
6131
Fig
u
r
e
1
.
Ar
c
h
itectu
r
e
of
p
r
o
p
o
s
ed
m
o
d
el
2
.
1
.
M
o
t
her’
s
s
ig
nifica
nt
f
ea
t
ure
da
t
a
s
et
W
ith
th
e
aim
to
wo
r
k
t
o
war
d
s
th
e
b
etter
m
en
t
of
m
o
th
er
an
d
ch
ild
,
a
d
ataset
titl
ed
MSF
d
ataset
is
cr
ea
ted
.
MSF
d
ataset
g
iv
es
an
in
s
ig
h
t
i
n
to
th
e
m
o
t
h
er
’
s
m
en
tal
an
d
p
h
y
s
ical
h
ea
lth
c
o
n
d
itio
n
s
ac
r
o
s
s
th
r
e
e
p
h
ases
(
teen
ag
e,
af
ter
m
a
r
r
iag
e,
d
u
r
i
n
g
p
r
eg
n
an
cy
)
of
h
er
r
e
p
r
o
d
u
ctiv
e
ag
e
.
T
h
e
m
ajo
r
ity
of
th
e
wo
m
en
wh
o
d
eliv
er
ed
b
a
b
ies
b
etwe
en
Feb
r
u
ar
y
2
0
1
8
an
d
Sep
tem
b
er
2
0
1
9
at
D.
Y.
Patil
Ho
s
p
ital,
lo
ca
ted
in
th
e
Mu
m
b
ai
Me
tr
o
p
o
litan
R
eg
io
n
of
I
n
d
i
a,
wer
e
in
ter
v
iewe
d
.
MSF
d
ataset
co
n
s
is
t
s
of
119
f
ea
tu
r
es
of
1
,
000
w
o
m
en
,
wh
ich
ar
e
f
u
r
t
h
er
d
iv
id
ed
in
t
o
f
o
u
r
ca
teg
o
r
ies
n
am
ely
p
h
y
s
ical
(
2
2
f
ea
t
u
r
es),
s
o
cial
(
1
9
f
ea
tu
r
es),
life
s
ty
le
(
6
8
f
ea
t
u
r
es),
a
n
d
s
tr
ess
(
1
0
f
e
atu
r
es).
Ou
t
of
all
th
e
r
ec
o
r
d
s
172
wo
m
en
h
ad
Pre
ter
m
d
eliv
er
y
wh
ile
8
2
8
h
ad
f
u
ll
-
ter
m
d
eliv
er
y
.
Deta
ils
ab
o
u
t
th
e
p
lan
n
in
g
,
c
h
allen
g
es
an
d
p
r
o
ce
d
u
r
es
f
o
llo
wed
to
cr
ea
te
th
e
MSF
d
ata
s
et
ar
e
m
en
tio
n
ed
in
o
u
r
p
r
ev
i
o
u
s
wo
r
k
[
2
9
]
.
MSF
d
ataset
is
p
ar
tially
av
ailab
le
o
n
lin
e
[
3
0
]
.
Ou
r
p
r
ev
i
o
u
s
ly
p
r
o
p
o
s
ed
m
eth
o
d
s
on
t
h
e
MS
F
d
ataset
in
clu
d
e
m
o
d
els
lik
e
Var
iatio
n
a
n
d
I
n
f
o
r
m
atio
n
b
ased
r
an
d
o
m
f
o
r
est
(
VI
B
R
F)
[
3
1
]
,
[
3
2
]
wh
ich
f
o
cu
s
es
on
r
elev
an
t
f
ea
tu
r
e
s
elec
tio
n
an
d
r
a
n
d
o
m
f
o
r
est
b
ased
f
u
zz
y
f
ea
tu
r
e
weig
h
in
g
f
o
r
class
im
b
alan
ce
(
R
FF
W
C
I
)
[
3
3
]
wh
ich
weig
h
s
f
ea
tu
r
es
f
o
r
im
b
alan
ce
h
a
n
d
lin
g
.
T
h
is
s
tu
d
y
,
p
r
o
p
o
s
es
MPM
U
m
eth
o
d
wh
ich
r
ea
lizes
th
e
co
n
tr
ib
u
tio
n
of
in
s
tan
ce
s
of
th
e
d
ataset
to
war
d
s
im
b
alan
ce
is
s
u
e.
To
v
alid
ate
th
e
p
er
f
o
r
m
a
n
ce
of
th
e
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
,
s
ev
en
s
tan
d
ar
d
im
b
alan
ce
d
d
atasets
f
r
o
m
th
e
Keel
r
ep
o
s
ito
r
y
[
3
4
]
a
r
e
u
s
ed
f
o
r
ex
p
e
r
im
en
tatio
n
.
2
.
2
.
P
r
o
po
s
ed
im
ba
la
nce
ha
nd
lin
g
m
et
ho
d
As
d
is
cu
s
s
ed
in
th
e
liter
atu
r
e
u
n
d
er
s
ec
tio
n
2,
e
x
is
tin
g
c
lu
s
ter
-
b
ased
ap
p
r
o
ac
h
es
h
a
v
e
r
elied
on
r
esam
p
lin
g
a
n
d
h
av
e
lo
o
k
ed
u
p
o
n
th
e
cl
u
s
ter
s
in
d
iv
id
u
ally
a
n
d
not
in
r
elatio
n
to
th
e
d
atas
et
as
a
wh
o
le.
T
h
e
p
r
o
p
o
s
ed
im
b
ala
n
ce
h
an
d
lin
g
m
eth
o
d
MPM
U
p
r
o
v
id
es
a
n
o
v
el
co
m
p
lete
s
o
lu
tio
n
wh
er
e
c
lu
s
ter
s
ar
e
an
aly
ze
d
in
d
iv
id
u
ally
as
well
as
in
r
el
atio
n
to
th
e
o
v
er
all
d
ataset
an
d
in
s
tan
ce
s
ar
e
weig
h
ted
by
in
v
esti
g
atin
g
th
e
clu
s
ter
,
th
ey
f
all
in
.
MPM
U
c
o
n
s
is
ts
of
two
m
ajo
r
co
m
p
o
n
en
ts
n
am
ely
in
s
tan
ce
weig
h
in
g
(
I
W
)
an
d
r
an
d
o
m
o
v
er
s
am
p
lin
g
(
R
OS)
,
wh
ich
a
r
e
ex
p
lain
ed
f
u
r
t
h
er
in
t
h
is
s
ec
tio
n
.
2
.
2
.
1
.
M
a
j
o
rit
y
pena
lizing
min
o
rit
y
up
s
ca
lin
g
T
h
e
p
r
o
p
o
s
ed
MPM
U
alg
o
r
ith
m
aim
s
at
u
n
d
er
s
tan
d
in
g
th
e
m
ix
in
g
up
of
m
ajo
r
ity
class
in
s
tan
ce
s
with
m
in
o
r
ity
class
in
s
tan
ce
s
.
MPM
U
alg
o
r
ith
m
wo
r
k
s
on
m
ajo
r
ity
in
s
tan
ce
s
with
an
IW
m
eth
o
d
an
d
on
m
in
o
r
ity
in
s
tan
ce
s
by
r
an
d
o
m
ly
u
p
s
ca
lin
g
th
em
.
MPM
U
alg
o
r
ith
m
u
s
es
k
-
m
ea
n
s
clu
s
ter
in
g
to
g
r
o
u
p
s
im
ilar
in
s
tan
ce
s
to
g
eth
er
,
co
n
s
id
er
i
n
g
in
tr
a
-
cl
u
s
ter
v
ar
ia
n
ce
as
e
x
p
r
ess
ed
in
(
1
)
f
o
r
clu
s
ter
s
an
d
in
s
tan
ce
s
with
as
th
e
ce
n
tr
o
id
of
th
e
clu
s
ter
.
=
∑
∑
=
1
=
1
∥
(
)
−
∥
2
(
1
)
W
eig
h
ts
of
m
ajo
r
ity
in
s
tan
ce
s
ar
e
d
ec
id
ed
b
ased
on
th
e
co
m
p
o
s
itio
n
of
t
h
e
clu
s
ter
th
ey
f
all
in
,
t
h
u
s
th
e
n
u
m
b
er
of
clu
s
ter
s
to
be
f
o
r
m
ed
r
ef
lects
th
e
p
r
o
p
o
r
tio
n
of
m
ajo
r
ity
class
in
s
tan
ce
s
in
th
e
wh
o
le
d
ataset
i.e
.
p
r
o
b
ab
ilit
y
of
o
cc
u
r
r
en
ce
of
m
ajo
r
ity
class
as
s
h
o
wn
in
(
3
)
of
T
ab
le
1,
wh
er
e
(
)
is
th
e
p
r
o
b
ab
ilit
y
of
o
cc
u
r
r
e
n
ce
of
m
in
o
r
ity
in
s
ta
n
ce
an
d
(
)
is
th
e
p
r
o
b
ab
ilit
y
of
o
cc
u
r
r
e
n
ce
of
m
aj
o
r
ity
in
s
tan
ce
in
th
e
d
ataset.
is
d
ir
ec
tly
p
r
o
p
o
r
tio
n
al
to
t
h
e
d
ataset
im
b
alan
ce
,
with
m
o
r
e
im
b
alan
ce
in
th
e
d
ataset
g
r
ea
ter
n
u
m
b
er
of
clu
s
ter
s
will
be
f
o
r
m
ed
,
s
am
e
ca
n
be
s
ee
n
in
T
a
b
le
2.
A
p
er
f
ec
tly
b
alan
ce
d
d
ataset
will
h
av
e
two
clu
s
ter
s
with
v
er
y
m
in
im
al
m
ix
.
Fig
u
r
e
2
s
h
o
ws
th
e
f
l
o
wch
ar
t
f
o
r
th
e
p
r
o
p
o
s
ed
a
r
ch
itectu
r
e
u
s
in
g
MSF
d
ataset,
MPM
U
an
d
6L
-
ANN.
I
n
p
u
t
to
th
e
m
o
d
el
is
th
e
MSF
d
ataset
an
d
o
u
tp
u
t
is
a
f
u
zz
y
v
alu
e
d
ep
ictin
g
th
e
p
r
o
b
a
b
ilit
y
of
o
cc
u
r
r
en
ce
of
PTB
an
d
f
u
ll
-
ter
m
b
ir
th
.
In
Fig
u
r
e
2,
th
e
b
lu
e
c
o
lo
r
b
o
x
es
s
h
o
w
th
e
s
tep
s
in
v
o
lv
ed
f
o
r
m
ajo
r
ity
in
s
tan
c
e
h
an
d
lin
g
i.e
.
I
W
,
th
e
g
r
ee
n
co
lo
r
b
o
x
d
e
p
icts
th
e
wo
r
k
d
o
n
e
f
o
r
t
h
e
m
in
o
r
ity
class
an
d
th
e
g
r
ey
co
lo
r
b
o
x
es
d
en
o
te
th
e
p
r
o
ce
s
s
in
g
of
th
e
d
ataset
as
a
wh
o
le.
M
PMU
alg
o
r
ith
m
is
m
en
tio
n
ed
in
T
ab
le
1.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
1
2
9
-
6
1
3
9
6132
T
ab
le
1.
MPM
U
alg
o
r
ith
m
Input:
Dataset Y with ‘X’ instances and ‘F’ features
Output:
Processed majority and minority instances
1
.
Calculate dataset imbalance ‘DI’
=
(2)
2
.
Calculate the number of clusters ‘M’, to be formed.
3
.
Split dataset Y into ‘M’
clusters.
=
(
)
−
(
)
(3)
4
.
Create vector V_CMC[] of size X.
5
.
for N=1 to M (each of the M clusters)
calculate within Cluster Imbalance ‘CI
N
’
=
(4)
6
.
for i=1 to X (each of the X instances)
•
identify cluster ‘N’ within which instance ‘i’ falls.
•
identify CI
N
for cluster ‘N’
if
(CI
N
<
or CI
N
<
DI)
CMC
N
=0
else
CMC
N
=|CI
N
–
DI|
•
V_CMC[i]=CMC
N
(5)
7
.
for i=1 to X (across X instances)
for j=1 to F (across F features)
if (Minority instance)
Y[i][j]=Y[i][j]
else
Y[i][j]=Y[i][j]
–
(
*V_CMC[i])
(6)
8
.
Randomly oversample minority instances such that
A=B
T
ab
le
2
.
Data
s
ets an
aly
s
is
D
a
t
a
s
e
t
N
o
.
o
f
f
e
a
t
u
r
e
s
N
o
.
o
f
i
n
s
t
a
n
c
e
s
D
a
t
a
s
e
t
i
m
b
a
l
a
n
c
e
N
o
.
o
f
c
l
u
st
e
r
s
1
P
h
y
s
i
c
a
l
22
1
,
0
0
0
4
.
8
2
7
2
Li
f
e
st
y
l
e
68
1
,
0
0
0
4
.
8
2
7
3
S
t
r
e
ss
10
1
,
0
0
0
4
.
8
2
7
4
S
o
c
i
a
l
19
6
3
3
5
.
9
5
7
5
P
h
y
s
i
c
a
l
+
Li
f
e
s
t
y
l
e
+
S
t
r
e
ss
1
0
0
1
,
0
0
0
4
.
8
2
7
6
P
i
ma
I
n
d
i
a
7
7
6
8
1
.
8
6
4
7
v
e
h
i
c
l
e
2
13
9
8
8
2
.
8
8
6
8
seg
m
e
n
t
0
10
5
,
4
7
2
6
.
0
2
7
9
e
c
o
l
i
3
7
3
3
6
8
.
6
8
10
P
a
g
e
-
b
l
o
c
k
s
0
8
5
1
4
8
.
7
9
8
11
Y
e
a
st
2
_
v
s
_
4
8
7
6
8
9
.
0
8
8
12
v
o
w
e
l
0
19
2
,
3
0
8
9
.
9
8
8
2
.
2
.
2
.
M
a
j
o
rit
y
ins
t
a
nce
weig
hin
g
A
m
ajo
r
c
o
n
tr
ib
u
tio
n
o
f
th
e
M
PMU
alg
o
r
ith
m
is
to
p
en
alize
th
e
m
ajo
r
ity
class
in
s
tan
ce
s
th
r
o
u
g
h
th
e
I
W
m
eth
o
d
.
Fo
r
m
ed
cl
u
s
ter
s
ar
e
an
aly
ze
d
to
id
e
n
tify
th
e
m
a
jo
r
ity
class
in
s
tan
ce
s
wh
ich
ar
e
clo
s
e
to
m
in
o
r
ity
class
in
s
tan
ce
s
,
th
u
s
d
if
f
icu
lt
t
o
id
en
tify
an
d
ca
u
s
in
g
m
is
clas
s
if
icatio
n
.
T
h
e
I
W
m
eth
o
d
s
u
b
d
u
es
th
e
id
en
tifie
d
m
ajo
r
ity
class
in
s
tan
ce
s
b
y
p
en
alizin
g
th
em
.
T
h
e
p
en
alty
f
o
r
ea
ch
m
ajo
r
ity
class
in
s
tan
ce
is
d
ec
id
ed
b
y
lo
o
k
in
g
i
n
to
th
e
co
m
p
o
s
itio
n
o
f
th
e
clu
s
ter
it
f
alls
in
.
T
h
e
p
en
alty
is
n
o
t
th
e
s
am
e
f
o
r
al
l
th
e
m
ajo
r
ity
class
in
s
tan
ce
s
as
th
e
aim
i
s
to
s
u
b
s
id
e
o
n
ly
th
o
s
e
m
ajo
r
ity
class
in
s
tan
ce
s
wh
ich
ar
e
in
s
ep
ar
ab
le
f
r
o
m
m
in
o
r
ity
class
in
s
tan
ce
s
.
C
u
s
ter
im
b
ala
n
ce
is
ca
lcu
lated
f
o
r
ea
ch
o
f
th
e
‘
’
clu
s
ter
s
.
C
lu
s
ter
im
b
alan
ce
o
f
th
e
ℎ
clu
s
ter
is
ca
lcu
lated
as
s
h
o
wn
in
(
4
)
in
T
ab
le
1
,
wh
e
r
e
is
t
h
e
n
u
m
b
e
r
o
f
m
in
o
r
ity
class
i
n
s
tan
ce
s
in
th
e
ℎ
clu
s
ter
an
d
is
th
e
n
u
m
b
er
o
f
m
ajo
r
ity
class
in
s
tan
ce
s
in
th
e
ℎ
clu
s
ter
r
esp
ec
tiv
ely
.
d
ataset
im
b
alan
ce
(
DI
)
is
ca
lcu
lated
u
s
in
g
(
2
)
in
T
ab
le
1
,
w
h
er
e
is
t
h
e
n
u
m
b
e
r
o
f
m
in
o
r
ity
class
in
s
tan
ce
s
an
d
is
th
e
n
u
m
b
er
o
f
m
ajo
r
ity
class
in
s
tan
ce
s
in
th
e
d
ataset.
T
h
e
im
b
alan
ce
o
f
th
e
d
ataset
as
a
wh
o
le
p
lay
s
a
s
ig
n
if
ican
t
r
o
le
in
d
ec
id
in
g
th
e
p
en
alty
f
o
r
t
h
e
m
ajo
r
ity
class
in
s
tan
ce
s
.
I
f
th
e
im
b
alan
ce
with
in
th
e
clu
s
ter
is
g
r
ea
ter
th
a
n
th
e
im
b
alan
ce
o
f
th
e
d
ataset,
it
is
c
o
n
s
id
er
ed
t
o
b
e
d
ef
y
i
n
g
t
h
e
d
ataset
co
m
p
o
s
itio
n
an
d
th
u
s
,
a
p
en
alty
will
b
e
ap
p
lied
to
th
e
m
ajo
r
ity
in
s
tan
ce
s
o
f
s
u
ch
clu
s
ter
s
.
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
p
en
alize
s
th
e
m
ajo
r
ity
class
in
s
tan
ce
s
o
n
l
y
i
f
th
e
‘
’
v
alu
e
o
f
th
e
clu
s
ter
in
t
o
wh
ich
th
e
i
n
s
tan
ce
is
f
allin
g
,
is
g
r
ea
ter
th
an
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
imb
a
la
n
ce
d
in
s
ta
n
ce
h
a
n
d
lin
g
to
w
a
r
d
b
etter p
r
eterm
b
ir
th
…
(
Hima
n
i D
esh
p
a
n
d
e
)
6133
th
e
v
alu
e
o
f
‘
’
.
C
lu
s
ter
m
is
c
lass
if
icatio
n
co
s
t
(
C
MC)
f
o
r
ea
ch
o
f
th
e
‘
’
clu
s
ter
s
ar
e
d
if
f
er
en
t
an
d
is
d
ec
id
ed
b
ased
o
n
th
e
v
ar
iatio
n
b
etwe
en
th
e
clu
s
ter
im
b
alan
c
e
an
d
d
ataset
im
b
alan
ce
.
v
alu
e
f
o
r
th
e
ℎ
clu
s
ter
is
ca
lcu
lated
u
s
in
g
(
5
)
i
n
T
ab
le
1
.
I
f
th
e
clu
s
ter
im
b
al
an
ce
is
v
er
y
lo
w
i.e
.
b
elo
w
a
c
er
tain
th
r
esh
o
ld
i.e
.
‘
’,
it
s
h
o
ws
th
at
th
e
clu
s
ter
c
o
n
s
titu
tes
m
ain
ly
o
f
th
e
m
ajo
r
ity
class
an
d
th
e
p
r
esen
ce
o
f
th
e
m
in
o
r
ity
class
ca
n
b
e
d
u
e
to
n
o
is
e
o
r
o
u
tlier
s
.
I
n
th
is
ca
s
e,
m
ajo
r
ity
class
in
s
tan
ce
s
will
n
o
t
b
e
p
e
n
alize
d
.
Fo
r
a
d
ataset
with
‘
’
in
s
tan
ce
s
,
v
ec
t
o
r
_
[
]
o
f
s
ize
,
is
u
s
ed
to
s
to
r
e
th
e
C
MC
v
alu
es
co
r
r
esp
o
n
d
in
g
to
ea
ch
o
f
th
e
d
ataset
in
s
tan
ce
s
lo
o
k
in
g
in
to
th
e
clu
s
ter
it
f
alls
in
to
.
T
h
e
C
MC
v
alu
e
o
f
ℎ
clu
s
ter
is
ca
lcu
l
ated
u
s
in
g
(
6
)
o
f
T
ab
le
1
,
c
o
n
s
id
er
in
g
th
e
im
b
alan
ce
o
f
t
h
e
clu
s
ter
a
n
d
d
ataset.
As
s
h
o
wn
in
(
6
)
,
wh
ile
p
en
alis
in
g
th
e
ℎ
f
ea
tu
r
e
o
f
th
e
ℎ
in
s
tan
ce
,
v
alu
e
is
u
s
ed
to
n
o
r
m
alize
th
e
d
if
f
er
en
ce
b
etwe
en
th
e
v
alu
e
o
f
f
ea
tu
r
es
an
d
C
MC
v
alu
e.
Fo
r
th
e
MSF
d
ataset,
o
p
tim
al
v
alu
e
o
f
a
n
d
is
f
o
u
n
d
to
b
e
0
.
7
an
d
0
.
3
af
ter
em
p
ir
ical
an
aly
s
is
.
Fig
u
r
e
2.
Flo
wch
ar
t
of
p
r
o
p
o
s
ed
im
b
alan
ce
d
in
s
tan
ce
h
an
d
lin
g
m
eth
o
d
2
.
2
.
3
.
Clus
t
er
f
o
rm
a
t
io
n
Fig
u
r
e
3,
s
h
o
ws
an
ex
am
p
le
f
o
r
th
e
f
o
r
m
atio
n
of
MPM
U
clu
s
ter
s
an
d
how
th
e
p
o
s
itiv
e
clas
s
in
s
tan
ce
s
ar
e
p
en
alize
d
.
In
Fig
u
r
e
3,
r
ed
cir
cles
d
e
n
o
te
m
in
o
r
ity
class
in
s
tan
ce
s
an
d
th
e
b
lack
+
(
)
s
ig
n
d
en
o
tes
m
ajo
r
ity
in
s
tan
ce
s
.
In
Fig
u
r
e
3,
clu
s
ter
1
an
d
clu
s
ter
4
h
av
e
in
s
tan
ce
s
o
n
ly
f
r
o
m
one
class
,
th
u
s
th
er
e
is
no
p
en
alty
on
th
ese
in
s
tan
ce
s
.
C
lu
s
ter
2
h
as
a
v
er
y
lo
w
im
b
alan
ce
i.e
.
b
elo
w
t
h
r
esh
o
ld
,
th
u
s
no
p
e
n
alty
on
m
ajo
r
ity
class
in
s
tan
ce
s
.
C
lu
s
ter
s
3
an
d
5
h
av
e
m
o
r
e
im
b
alan
ce
as
co
m
p
ar
ed
to
t
h
e
o
v
er
all
d
ataset,
th
e
m
ajo
r
ity
class
in
s
tan
ce
s
of
t
h
ese
clu
s
ter
s
ar
e
p
en
alize
d
by
lo
o
k
in
g
at
th
e
c
o
m
p
o
s
itio
n
of
th
e
clu
s
ter
an
d
ca
lcu
latin
g
th
e
C
MC
v
alu
e
as
ex
p
r
ess
ed
in
(
6
)
.
Dif
f
er
en
t
s
h
ad
es
of
g
r
ey
f
o
r
m
ajo
r
ity
cl
ass
in
s
tan
ce
s
(
+
)
in
Fig
u
r
e
3,
d
ep
icts
th
e
d
if
f
er
e
n
t
in
ten
s
ity
of
t
h
e
p
en
alty
,
lig
h
ter
s
h
ad
e
m
ea
n
s
a
h
ig
h
e
r
p
en
alty
.
MPM
U
alg
o
r
ith
m
d
o
es
not
p
en
alize
m
in
o
r
ity
in
s
tan
ce
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
1
2
9
-
6
1
3
9
6134
Fig
u
r
e
3
.
C
lu
s
ter
f
o
r
m
atio
n
2
.
2
.
4
.
M
ino
rit
y
ins
t
a
nce
ha
nd
lin
g
E
v
en
a
f
ter
p
en
alizin
g
th
e
m
ajo
r
ity
class
in
s
tan
ce
s
,
th
e
co
m
p
o
s
itio
n
of
th
e
d
ataset
r
em
ain
s
in
clin
ed
to
war
d
s
th
e
m
ajo
r
ity
class
.
To
b
alan
ce
th
e
co
m
p
o
s
itio
n
of
th
e
d
ataset,
th
e
MPM
U
a
lg
o
r
ith
m
o
v
er
s
am
p
les
th
e
m
in
o
r
ity
class
in
s
tan
ce
s
in
th
e
d
ataset,
s
u
ch
th
at
th
e
n
u
m
b
er
of
m
in
o
r
ity
class
in
s
tan
ce
s
g
ets
eq
u
al
to
th
e
n
u
m
b
er
of
m
ajo
r
ity
class
in
s
ta
n
ce
s
as
s
h
o
wn
in
s
tep
8
of
T
a
b
le
1.
Fo
r
th
e
MSF
d
ataset,
s
a
m
p
les
ar
e
co
llected
by
in
ter
v
iewin
g
wo
m
en
f
r
o
m
s
im
ilar
g
eo
g
r
ap
h
ic
an
d
d
em
o
g
r
ap
h
ic
lo
ca
tio
n
s
.
Kee
p
in
g
th
e
s
am
e
in
m
in
d
,
f
o
r
o
v
er
-
s
am
p
lin
g
,
no
s
y
n
th
etic
i
n
s
tan
ce
s
ar
e
g
en
er
ated
in
s
tea
d
it
is
done
by
r
an
d
o
m
d
u
p
li
ca
tio
n
of
m
in
o
r
ity
s
am
p
les,
th
u
s
ev
en
af
ter
o
v
er
-
s
am
p
lin
g
th
e
in
s
tan
ce
,
th
e
f
ea
t
u
r
e
v
alu
e
r
em
ain
s
tr
u
e
to
th
e
o
r
ig
in
al
p
o
p
u
latio
n
of
th
e
MSF
d
ataset.
2
.
3
.
6
L
_
ANN
Var
iatio
n
s
ac
r
o
s
s
d
if
f
e
r
en
t
h
y
p
er
p
ar
am
eter
s
of
a
r
tific
ial
n
eu
r
al
n
etwo
r
k
(
ANN)
m
o
d
els
u
s
in
g
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
(
ML
P)
with
b
ac
k
p
r
o
p
ag
atio
n
ar
e
u
s
ed
in
th
e
ex
p
er
im
en
ts
.
B
ase
d
on
th
e
em
p
ir
ical
an
aly
s
is
,
6
L
_
ANN,
a
s
ix
-
lay
er
ed
ANN
ar
ch
itectu
r
e
is
s
elec
ted
f
o
r
class
if
icatio
n
wh
ich
co
n
s
is
ts
of
two
d
r
o
p
-
out
lay
e
r
s
an
d
two
ac
tiv
atio
n
lay
er
s
alo
n
g
with
an
o
u
t
p
u
t
an
d
an
in
p
u
t
lay
er
.
R
ec
tifie
d
lin
ea
r
ac
tiv
atio
n
u
n
it
(
R
e
LU
)
ac
tiv
atio
n
f
u
n
cti
o
n
is
u
s
ed
f
o
r
h
id
d
en
lay
er
s
an
d
th
e
o
u
tp
u
t
lay
er
u
s
es
th
e
s
ig
m
o
id
ac
tiv
atio
n
f
u
n
ctio
n
.
MSF
d
ataset
p
r
o
ce
s
s
ed
u
s
in
g
th
e
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
is
f
u
r
th
e
r
u
s
ed
to
t
r
ain
th
e
6
L
_
ANN
n
etwo
r
k
to
class
if
y
PTB
an
d
F
T
B
o
u
tco
m
es.
3.
RE
SU
L
T
S
AND
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
of
th
e
p
a
p
er
p
r
esen
ts
th
e
class
if
icatio
n
r
esu
lts
u
s
in
g
th
e
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
an
d
6L
-
ANN
n
etwo
r
k
.
Mu
ltip
le
e
x
p
er
im
en
ts
ar
e
p
er
f
o
r
m
e
d
to
u
n
d
er
s
tan
d
th
e
n
atu
r
e
of
th
e
p
r
o
p
o
s
ed
m
eth
o
d
s
with
d
if
f
er
en
t
class
if
ier
s
an
d
d
atasets
.
To
ch
ec
k
th
e
im
b
alan
ce
h
an
d
li
n
g
ef
f
icien
cy
of
th
e
MPM
U
alg
o
r
ith
m
,
alo
n
g
with
th
e
p
r
o
p
o
s
ed
6L
-
ANN,
f
iv
e
p
o
p
u
lar
class
if
icatio
n
alg
o
r
ith
m
s
,
n
a
m
ely
,
r
a
n
d
o
m
f
o
r
est
(
R
F),
d
ec
is
io
n
tr
ee
(
DT
)
,
g
au
s
s
ian
NB
(
GN
B
)
,
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
(
ℎ
=
3
)
,
an
d
MLP
ar
e
u
s
ed
f
o
r
ex
p
er
im
en
tatio
n
.
R
esear
ch
er
s
s
u
g
g
est
th
at
wh
ile
wo
r
k
in
g
on
an
im
b
alan
ce
d
d
ataset
ar
ea
u
n
d
er
th
e
cu
r
v
e
(
AUC
)
cu
r
v
e
,
p
r
ec
is
io
n
v
alu
e
an
d
p
r
ed
ictio
n
ac
cu
r
ac
y
t
o
g
eth
er
n
ee
d
to
be
co
n
s
id
er
ed
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
of
th
e
class
if
icatio
n
task
[
3
5
]
,
[
3
6
]
.
Pre
d
ictio
n
ac
cu
r
ac
y
ev
alu
ates
th
e
o
v
er
al
l
p
er
f
o
r
m
a
n
ce
of
a
class
if
ier
,
AU
C
v
alu
e
ass
es
th
e
ca
p
ab
ilit
y
to
d
is
tin
g
u
is
h
b
et
wee
n
th
e
class
es,
an
d
p
r
ec
is
io
n
m
ea
s
u
r
es
th
e
PTB
class
.
T
h
u
s
,
th
is
s
tu
d
y
c
o
n
s
id
er
s
p
r
ec
is
io
n
,
AUC
an
d
p
r
e
d
i
ctio
n
ac
cu
r
ac
y
f
o
r
ev
al
u
atin
g
th
e
p
er
f
o
r
m
a
n
ce
of
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
3
.
1
.
Da
t
a
s
et
s
a
na
ly
s
is
Fo
r
ex
p
er
im
en
tatio
n
,
alo
n
g
with
MSF
d
atase
t,
7
s
tan
d
ar
d
d
atasets
ar
e
u
s
ed
.
T
ab
le
2,
s
h
o
ws
th
e
d
ataset
d
etails
in
ter
m
s
of
d
at
aset
s
ize
an
d
im
b
alan
ce
.
T
h
e
f
ir
s
t
f
iv
e
r
o
ws
of
T
a
b
le
2,
s
h
o
ws
MSF
d
ataset
s
in
f
o
r
m
atio
n
an
d
th
e
r
em
ain
in
g
7
r
o
ws
h
av
e
in
f
o
r
m
ati
o
n
a
b
o
u
t
th
e
o
t
h
er
s
tan
d
ar
d
d
atasets
u
s
ed
.
T
h
er
e
ar
e
m
is
s
in
g
r
ec
o
r
d
s
u
n
d
er
s
o
cial
f
ea
tu
r
es
of
t
h
e
MSF
d
ataset,
f
o
r
th
e
s
am
e
r
ea
s
o
n
o
n
ly
633
in
s
tan
ce
s
of
it
ar
e
co
n
s
id
er
ed
d
u
r
in
g
ex
p
er
im
en
t
atio
n
.
As
s
h
o
wn
in
T
ab
le
2,
s
tan
d
ar
d
d
atasets
ar
e
s
elec
ted
with
v
ar
iatio
n
s
in
ter
m
s
of
im
b
alan
ce
r
atio
,
n
u
m
b
er
of
f
ea
t
u
r
es
an
d
n
u
m
b
er
of
in
s
tan
ce
s
so
th
at
th
e
ex
p
er
im
en
tal
r
esu
lts
v
alid
ate
th
e
v
er
s
atility
of
th
e
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
.
T
h
e
last
co
lu
m
n
o
f
T
ab
le
2
s
h
o
ws
th
e
co
u
n
t
of
clu
s
ter
s
f
o
r
m
ed
as
ex
p
r
ess
ed
as
‘
’
in
(
3
)
,
of
th
e
MPM
U
alg
o
r
ith
m
.
Fo
r
ex
p
er
im
e
n
tatio
n
,
all
th
e
d
atasets
ar
e
p
r
ep
r
o
ce
s
s
ed
f
o
r
h
an
d
lin
g
m
is
s
in
g
v
alu
es
an
d
h
av
e
b
ee
n
n
o
r
m
alize
d
,
s
u
ch
th
at
f
o
r
each
r
ec
o
r
d
,
th
e
s
u
m
of
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
imb
a
la
n
ce
d
in
s
ta
n
ce
h
a
n
d
lin
g
to
w
a
r
d
b
etter p
r
eterm
b
ir
th
…
(
Hima
n
i D
esh
p
a
n
d
e
)
6135
s
q
u
ar
e
of
n
o
r
m
alize
d
v
alu
es
is
eq
u
al
to
o
n
e.
A
to
tal
o
f
7
0
%
of
th
e
wh
o
le
d
ataset
is
tak
en
f
o
r
tr
ain
i
n
g
wh
ile
30%
is
co
n
s
id
er
ed
f
o
r
test
in
g
.
3
.
2
.
E
x
perim
ent
a
l
re
s
ults o
n
M
SF
da
t
a
s
et
Mu
ltip
le
ex
p
er
im
en
ts
ar
e
p
er
f
o
r
m
e
d
on
t
h
e
MSF
d
ataset
to
u
n
d
er
s
tan
d
th
e
ef
f
ec
t
iv
en
ess
of
p
en
alizin
g
(
I
W
)
an
d
o
v
er
s
am
p
lin
g
s
tep
s
of
th
e
MPM
U
alg
o
r
ith
m
.
Fig
u
r
e
4
s
h
o
ws
th
e
class
if
icatio
n
r
esu
lts
on
f
iv
e
d
if
f
er
en
t
MSF
s
u
b
-
d
atase
ts
as
r
ep
r
esen
ted
in
th
e
f
ir
s
t
f
iv
e
r
o
ws
of
T
ab
le
2.
E
x
p
e
r
im
en
ts
ar
e
c
o
n
d
u
cted
with
s
ev
en
co
m
b
in
atio
n
s
of
d
ata
p
r
o
ce
s
s
in
g
m
eth
o
d
s
an
d
c
lass
if
ier
s
.
Fi
g
u
r
e
4
s
h
o
ws
th
e
r
esu
lts
of
in
s
tan
ce
weig
h
t
(
IW
)
an
d
MPM
U
m
eth
o
d
in
co
m
b
in
atio
n
with
ML
P,
GNB,
KNN,
R
F,
DT
an
d
th
e
p
r
o
p
o
s
ed
6L
-
ANN
class
if
ier
.
In
Fig
u
r
e
4,
th
e
M
PMU
m
o
d
el
is
ev
alu
ated
with
6L
-
ANN
class
if
ier
.
T
h
er
e
wer
e
m
an
y
m
is
s
in
g
v
alu
es
in
th
e
s
o
cial
s
u
b
-
d
ataset
f
o
r
th
e
s
am
e
r
ea
s
o
n
it
was
not
i
n
clu
d
ed
wh
ile
co
n
d
u
cti
n
g
co
m
b
in
ed
MSF
d
ataset
ex
p
er
im
en
tatio
n
.
Fig
u
r
es
4
(
a)
to
4
(
c)
s
h
o
w
th
e
p
r
ec
is
io
n
v
alu
e,
AUC
an
d
p
r
ed
ict
io
n
ac
cu
r
ac
y
u
s
in
g
th
e
p
r
o
p
o
s
ed
IW
s
tep
an
d
MP
MU
m
eth
o
d
on
MSF
s
u
b
-
d
ata
s
ets.
(
a)
(
b
)
(
c)
Fig
u
r
e
4
.
C
lass
if
icatio
n
r
esu
lts
on
MSF
Su
b
-
d
atasets
(
p
h
y
s
ic
al,
life
s
ty
le,
s
tr
ess
,
s
o
cial
an
d
co
m
b
in
ed
MSF
f
ea
tu
r
es
(
p
h
y
s
ical,
life
s
ty
le
an
d
s
tr
ess
f
ea
tu
r
es
))
u
s
in
g
th
e
p
r
o
p
o
s
ed
IW
s
tep
an
d
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
with
class
if
ier
s
n
am
ely
ML
P,
GNB,
KNN,
R
F,
DT
an
d
p
r
o
p
o
s
ed
6L
-
ANN
(
a)
d
em
o
n
s
tr
ate
class
if
icatio
n
r
esu
lts
in
ter
m
s
o
f
p
r
ec
is
io
n
,
(
b
)
p
r
e
d
ictio
n
r
esu
lts
in
ter
m
s
of
AUC
v
alu
e
,
an
d
(
c)
p
r
ed
ictio
n
ac
cu
r
ac
y
of
class
if
icatio
n
MSF
d
ataset
is
an
im
b
alan
ce
d
d
ataset
an
d
r
esu
lts
r
ef
lect
cla
s
s
if
icatio
n
b
ias,
s
am
e
ca
n
b
e
j
u
s
tifie
d
b
y
th
e
r
esu
lts
in
Fig
u
r
e
4
.
W
h
ile
p
r
e
-
p
r
o
ce
s
s
in
g
th
e
d
atasets
u
s
in
g
I
W
,
d
esp
ite
h
i
g
h
er
p
r
e
d
ictio
n
ac
c
u
r
ac
y
an
d
a
d
ec
en
t
AUC
v
alu
e,
th
e
p
r
ec
i
s
io
n
v
alu
e
is
v
er
y
lo
w
f
o
r
class
if
ier
s
lik
e
GNB,
KNN
an
d
DT
f
o
r
all
MSF
d
atasets
,
th
is
s
h
o
ws
th
at
th
e
r
e
s
u
lts
ar
e
in
clin
ed
to
war
d
th
e
m
in
o
r
ity
class
.
C
o
n
s
id
er
in
g
th
e
p
r
ed
ictio
n
o
f
PTB
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
1
2
9
-
6
1
3
9
6136
as
a
p
r
io
r
ity
co
n
ce
r
n
o
v
er
t
h
e
p
r
ed
ictio
n
o
f
FTB,
th
u
s
co
u
n
tin
g
p
r
ec
is
io
n
as
th
e
m
o
s
t
cr
u
cial
e
v
alu
atio
n
p
ar
am
eter
,
it
co
u
ld
b
e
c
o
n
clu
d
ed
th
at
R
F
is
g
iv
in
g
th
e
b
est
r
esu
lts
f
o
r
all
t
h
e
I
W
p
r
e
-
p
r
o
ce
s
s
ed
MSF
d
atasets
,
an
d
th
e
s
ec
o
n
d
-
b
est
r
esu
lts
a
r
e
ac
h
iev
ed
b
y
6
L
-
ANN
n
etwo
r
k
f
o
r
all
th
e
MSF
s
u
b
-
d
atasets
.
I
t
i
s
f
u
r
th
er
o
b
s
er
v
ed
th
at
th
e
I
W
s
tep
alo
n
e
is
n
o
t
as
e
f
f
ec
tiv
e
as
th
e
MPM
U
m
eth
o
d
wh
ich
co
m
b
in
es
m
ajo
r
ity
in
s
tan
ce
weig
h
in
g
an
d
m
i
n
o
r
ity
o
v
er
s
am
p
lin
g
,
as
th
er
e
is
a
s
ig
n
if
ican
t
im
p
r
o
v
em
e
n
t
in
r
esu
lts
f
o
r
all
MSF
s
u
b
-
d
atasets
wh
ile
p
r
e
-
p
r
o
ce
s
s
in
g
u
s
in
g
th
e
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
a
n
d
f
u
r
t
h
er
u
s
in
g
6
L
-
A
NN
n
etwo
r
k
as
th
e
class
if
ier
.
E
x
p
er
im
en
tal
r
esu
lts
in
Fig
u
r
e
4,
v
alid
ate
th
e
s
tr
en
g
th
of
th
e
p
r
o
p
o
s
ed
m
o
d
el
i.e
.
M
PMU
with
6L
-
ANN
class
if
ier
,
p
r
ec
is
io
n
b
etwe
en
0
.
9
0
to
0
.
9
7
is
ac
h
iev
ed
on
MST
s
u
b
-
d
atasets
,
an
AUC
v
alu
e
r
an
g
in
g
b
etwe
en
0
.
8
6
to
0
.
9
9
is
o
b
s
er
v
ed
,
an
d
p
r
e
d
ictio
n
ac
cu
r
ac
y
of
9
3
.
0
0
%
to
9
9
.
4
7
%
is
ac
h
iev
ed
.
L
o
o
k
in
g
in
to
th
e
r
esu
lts
in
Fig
u
r
e
4,
it
is
o
b
s
er
v
ed
th
at
m
o
t
h
er
’
s
life
s
ty
le
an
d
s
tr
ess
f
ea
tu
r
es
ar
e
ac
h
iev
in
g
b
etter
class
if
icatio
n
r
esu
lts
f
o
r
PTB
p
r
e
d
ictio
n
.
T
h
ir
d
-
b
est
class
if
icatio
n
r
esu
lts
ar
e
ac
h
iev
ed
u
s
in
g
s
o
cial
f
e
atu
r
es
an
d
p
h
y
s
ical
f
ea
tu
r
es
class
if
icatio
n
r
esu
lts
ar
e
f
o
u
n
d
to
be
th
e
least
ac
cu
r
ate
o
n
es.
On
e
of
th
e
p
r
o
b
ab
le
r
ea
s
o
n
s
f
o
r
in
f
er
io
r
r
esu
lts
f
o
r
s
o
cial
f
ea
tu
r
es
wo
u
ld
be
a
s
m
aller
d
ataset
an
d
h
ig
h
er
-
class
im
b
alan
ce
.
T
h
is
s
ig
n
if
ies
th
at
th
e
m
o
th
er
’
s
m
e
n
tal
h
ea
lth
f
ea
tu
r
es
ar
e
clo
s
ely
r
elate
d
to
th
e
p
r
eg
n
an
cy
o
u
tco
m
e
in
ter
m
s
of
th
e
class
if
icatio
n
of
PTB
an
d
FTB.
L
o
o
k
in
g
in
to
t
h
e
r
esu
lts
of
th
e
p
r
o
p
o
s
ed
MP
MU
m
eth
o
d
in
T
ab
le
3,
with
s
tan
d
ar
d
class
if
ier
s
an
d
with
th
e
r
esu
lts
in
Fig
u
r
e
4
with
6L
-
ANN
class
if
ier
s
h
o
ws
th
at
am
o
n
g
all
th
e
class
if
ie
r
s
MPM
U
is
g
iv
in
g
th
e
b
est
r
esu
lts
u
s
in
g
6L
-
ANN
class
if
ier
.
T
h
e
ex
p
er
im
en
ta
l
r
esu
lts
on
d
if
f
er
en
t
s
u
b
-
d
ata
s
ets
of
MP
MU
as
s
h
o
wn
in
Fig
u
r
e
4,
s
u
p
p
o
r
t
t
h
e
h
y
p
o
th
esis
th
at
a
m
o
th
er
’
s
m
en
tal
h
ea
lth
p
lay
s
a
c
r
u
cial
r
o
le
in
p
r
e
d
ictin
g
th
e
PTB
b
ir
th
o
u
tco
m
e.
As
s
ee
n
in
Fig
u
r
e
4,
m
o
s
t
of
th
e
cl
ass
if
ier
s
ar
e
g
iv
in
g
b
etter
p
r
ed
ictio
n
s
u
s
in
g
th
e
life
s
ty
le
an
d
s
tr
ess
f
ea
tu
r
e
s
u
b
-
d
at
aset.
Usi
n
g
life
s
ty
le
f
ea
t
u
r
es
p
r
ec
is
io
n
is
ac
h
iev
ed
in
t
h
e
r
an
g
e
of
0
.
2
3
to
0
.
9
7
,
AUC
v
alu
e
is
ac
h
iev
ed
b
etwe
en
0
.
7
2
to
0
.
9
9
an
d
p
r
ed
ictio
n
ac
cu
r
ac
y
is
in
th
e
r
an
g
e
of
55
%
9
9
%.
Usi
n
g
s
tr
ess
f
ea
tu
r
es
p
r
ec
is
io
n
is
ac
h
iev
ed
in
th
e
r
an
g
e
0
.
2
5
to
0
.
9
8
,
AUC
v
alu
e
is
ac
h
iev
ed
b
etwe
en
0
.
6
2
to
0
.
9
9
an
d
p
r
e
d
ictio
n
ac
cu
r
ac
y
in
th
e
r
an
g
e
59
%
to
1
0
0
%.
To
u
n
d
er
s
tan
d
th
e
p
er
f
o
r
m
an
ce
of
MPM
U
m
eth
o
d
o
v
er
ex
is
tin
g
d
ata
im
b
alan
ce
s
o
lu
tio
n
s
,
ex
p
er
im
en
ts
ar
e
co
n
d
u
cted
u
s
in
g
alg
o
r
ith
m
s
p
r
o
p
o
s
ed
by
r
esear
ch
er
s
in
th
e
p
ast.
T
ab
le
3
d
em
o
n
s
tr
ates
th
e
r
esu
lts
f
o
r
t
h
e
s
am
e.
T
h
e
d
ata
p
r
o
ce
s
s
in
g
m
eth
o
d
s
u
s
ed
f
o
r
im
p
lem
en
tatio
n
ar
e
s
ca
led
v
al
u
es
(
SV
),
clu
s
ter
ed
SMOT
E
(
CS
)
[
3
7
]
,
r
a
n
d
o
m
o
v
er
s
am
p
lin
g
(
R
OS
)
[
3
8
]
an
d
th
e
p
r
o
p
o
s
ed
MPM
U
alg
o
r
ith
m
.
B
ein
g
im
b
alan
ce
d
MSF d
ataset
is
f
o
u
n
d
to
b
e
p
e
r
f
o
r
m
in
g
ex
t
r
em
ely
p
o
o
r
ly
in
ter
m
s
o
f
p
r
ec
is
io
n
an
d
AUC
v
alu
e,
ev
en
af
ter
s
ca
lin
g
th
e
v
alu
es
u
s
in
g
SV
m
eth
o
d
,
class
if
ier
s
f
ail
to
p
er
f
o
r
m
th
e
s
am
e
ca
n
b
e
s
ee
n
in
th
e
r
esu
lts
o
f
T
a
b
le
3
.
T
h
e
cl
u
s
ter
-
b
ased
in
s
tan
ce
weig
h
in
g
an
d
o
v
er
s
am
p
lin
g
ap
p
r
o
ac
h
o
f
MPM
U
is
ca
p
ab
le
o
f
h
an
d
lin
g
m
ajo
r
ity
b
ias
an
d
in
s
ep
ar
ab
ilit
y
is
s
u
es
o
f
MSF
d
atasets
th
e
s
am
e
ca
n
b
e
s
ee
n
in
th
e
r
esu
lts
o
f
T
ab
le
3
.
T
h
er
e
is
a
s
ig
n
if
ica
n
t
im
p
r
o
v
em
e
n
t
in
th
e
r
esu
lts
o
f
MPM
U
m
eth
o
d
with
all
t
h
e
u
s
ed
class
if
ier
s
wh
en
co
m
p
ar
e
d
to
SV,
C
S,
a
n
d
R
OS
im
b
alan
ce
h
a
n
d
lin
g
m
eth
o
d
s
.
On
av
e
r
ag
e
p
r
ec
is
io
n
v
alu
e
is
im
p
r
o
v
e
d
b
y
4
8
%,
AUC
b
y
2
6
%
an
d
p
r
ed
ictio
n
ac
cu
r
ac
y
b
y
1
6
%,
w
h
en
u
s
in
g
MPM
U
ag
ain
s
t
SV.
Pre
cisi
o
n
v
alu
e
is
im
p
r
o
v
e
d
b
y
5
2
%,
AUC
is
i
m
p
r
o
v
e
d
b
y
3
2
%
an
d
p
r
e
d
ictio
n
ac
c
u
r
ac
y
is
im
p
r
o
v
ed
b
y
2
6
%,
co
m
p
ar
in
g
th
e
r
esu
lts
o
f
MPM
U
ag
ain
s
t
C
S.
Ag
ain
s
t
R
OS
m
eth
o
d
MPM
U
r
esu
lts
ar
e
im
p
r
o
v
e
d
b
y
5
4
%,
3
1
%
a
n
d
1
6
%
f
o
r
p
r
ec
is
io
n
,
AUC an
d
p
r
ed
ictio
n
ac
cu
r
ac
y
.
Gr
ea
te
r
im
p
r
o
v
em
en
t in
p
r
ec
is
io
n
v
alu
es u
s
in
g
MPM
U
v
alid
ates th
e
s
u
cc
ess
o
f
MPM
U
in
id
e
n
tify
i
n
g
PTB
in
s
tan
ce
s
co
r
r
ec
tly
.
Am
o
n
g
th
e
u
s
ed
s
tan
d
a
r
d
class
if
ier
s
R
F
is
s
h
o
win
g
th
e
b
est r
esu
lts
f
o
r
all
th
e
s
elec
ted
im
b
alan
ce
h
a
n
d
lin
g
m
eth
o
d
s
o
n
MSF su
b
-
d
atasets
.
T
ab
le
3
.
C
o
m
p
a
r
ativ
e
r
esu
lts
o
f
d
if
f
e
r
en
t im
b
ala
n
ce
h
an
d
lin
g
m
eth
o
d
s
o
n
MST
s
u
b
-
d
atasets
SV
CS
R
O
S
M
P
M
U
SV
CS
R
O
S
M
P
M
U
SV
CS
R
O
S
M
P
M
U
SV
CS
R
O
S
M
P
M
U
SV
CS
R
O
S
M
P
M
U
P
h
y
s
i
c
a
l
P
r
e
c
i
s
i
o
n
0
.
1
8
0
.
1
9
0
.
1
9
0
.
4
2
0
.
2
2
0
.
2
1
0
.
2
0
.
8
0
.
2
0
.
2
0
.
2
0
.
2
7
0
.
2
9
0
.
2
9
0
.
2
1
0
.
4
4
0
.
2
3
0
.
2
3
0
.
2
3
0
.
5
3
AUC
0
.
5
3
0
.
4
7
0
.
5
5
0
.
7
6
0
.
6
0
.
4
5
0
.
5
9
0
.
9
3
0
.
5
2
0
.
5
1
0
.
5
2
0
.
6
6
0
.
6
4
0
.
5
2
0
.
4
9
0
.
8
2
0
.
5
2
0
.
5
0
.
5
2
0
.
7
8
A
c
c
u
r
a
c
y
7
1
.
7
7
4
.
5
7
2
.
8
8
6
.
2
8
3
.
9
8
0
.
9
8
2
.
8
9
1
.
6
7
7
.
6
6
5
.
1
77
8
3
.
2
1
8
.
4
1
8
.
3
3
1
.
3
5
8
.
7
6
6
.
5
6
6
.
5
6
6
.
5
7
9
.
2
Li
f
e
st
y
l
e
P
r
e
c
i
s
i
o
n
0
.
1
7
0
.
1
9
0
.
1
7
0
.
6
2
0
.
2
1
0
.
2
0
.
2
1
0
.
8
7
0
.
1
7
0
.
1
7
0
.
1
7
0
.
2
3
0
.
1
6
0
.
1
6
0
.
1
5
0
.
4
2
0
.
2
1
0
.
2
1
0
.
2
1
0
.
6
9
AUC
0
.
4
9
0
.
5
1
0
.
5
0
.
8
4
0
.
5
4
0
.
5
1
0
.
5
4
0
.
9
5
0
.
4
9
0
.
5
1
0
.
4
9
0
.
5
5
0
.
5
0
.
5
1
0
.
4
9
0
.
8
3
0
.
5
2
0
.
4
8
0
.
5
2
0
.
8
4
A
c
c
u
r
a
c
y
6
9
.
3
6
9
.
3
7
1
.
2
91
8
2
.
2
7
5
.
3
8
2
.
4
9
3
.
2
7
9
.
2
6
3
.
7
79
7
9
.
8
1
6
.
8
1
6
.
8
2
6
.
5
7
2
.
8
70
70
70
86
S
t
r
e
ss
P
r
e
c
i
s
i
o
n
0
.
2
0
.
2
0
.
1
9
0
.
6
9
0
.
2
1
0
.
2
1
0
.
2
1
0
.
7
6
0
.
2
0
.
1
9
0
.
2
0
.
4
1
0
.
2
3
0
.
2
3
0
.
2
0
.
2
5
0
.
2
4
0
.
2
4
0
.
2
4
0
.
3
9
AUC
0
.
5
3
0
.
5
2
0
.
5
3
0
.
8
7
0
.
5
1
0
.
4
9
0
.
5
1
0
.
8
8
0
.
5
4
0
.
5
0
.
5
3
0
.
7
1
0
.
5
8
0
.
5
2
0
.
5
1
0
.
5
9
0
.
5
6
0
.
5
2
0
.
5
6
0
.
6
9
A
c
c
u
r
a
c
y
8
1
.
1
5
8
.
6
8
0
.
3
9
1
.
1
7
9
.
5
5
8
.
8
7
9
.
1
9
2
.
1
7
7
.
3
6
2
.
1
7
4
.
7
8
3
.
8
8
0
.
4
5
9
.
9
6
8
.
7
7
8
.
7
5
9
.
5
5
9
.
5
5
9
.
5
6
2
.
5
S
o
c
i
a
l
P
r
e
c
i
s
i
o
n
0
.
1
7
0
.
1
7
0
.
1
6
0
.
5
2
0
.
1
8
0
.
1
8
0
.
1
7
0
.
8
1
0
.
1
5
0
.
1
4
0
.
1
5
0
.
1
8
0
.
1
7
0
.
1
8
0
.
1
5
0
.
2
6
0
.
1
7
0
.
1
7
0
.
1
7
0
.
4
AUC
0
.
5
4
0
.
4
8
0
.
5
2
0
.
8
2
0
.
5
1
0
.
4
8
0
.
5
1
0
.
9
4
0
.
5
3
0
.
5
3
0
.
5
3
0
.
5
4
0
.
5
1
0
.
4
8
0
.
4
3
0
.
6
6
0
.
5
1
0
.
5
4
0
.
5
1
0
.
7
5
A
c
c
u
r
a
c
y
7
5
.
6
7
3
.
5
7
4
.
7
8
9
.
7
8
1
.
9
7
6
.
7
8
0
.
9
9
3
.
3
8
2
.
7
69
8
0
.
8
8
0
.
1
5
9
.
7
3
4
.
2
6
2
.
3
7
5
.
6
6
6
.
8
6
6
.
8
6
6
.
8
7
6
.
4
Li
f
e
st
y
l
e
+
P
h
y
s
i
c
a
l
+
S
t
r
e
ss
P
r
e
c
i
s
i
o
n
0
.
2
0
.
2
0
.
1
9
0
.
2
4
0
.
2
5
0
.
2
2
0
.
2
6
0
.
5
3
0
.
1
7
0
.
1
6
0
.
1
7
0
.
2
1
0
.
2
4
0
.
2
4
0
.
1
8
0
.
2
6
0
.
2
1
0
.
2
1
0
.
2
1
0
.
7
6
AUC
0
.
5
3
0
.
5
0
.
5
3
0
.
6
2
0
.
5
8
0
.
5
2
0
.
5
9
0
.
7
7
0
.
5
0
.
4
8
0
.
5
0
.
5
3
0
.
5
8
0
.
4
7
0
.
4
6
0
.
6
6
0
.
5
3
0
.
4
9
0
.
5
3
0
.
8
8
A
c
c
u
r
a
c
y
7
2
.
2
7
2
.
2
7
0
.
3
7
7
.
3
8
1
.
8
8
0
.
1
8
1
.
7
8
4
.
6
79
6
2
.
9
7
8
.
4
7
8
.
6
2
4
.
7
25
5
2
.
6
3
7
.
6
7
2
.
3
7
2
.
3
7
2
.
3
8
9
.
5
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
n
o
ve
l a
p
p
r
o
a
c
h
fo
r
imb
a
la
n
ce
d
in
s
ta
n
ce
h
a
n
d
lin
g
to
w
a
r
d
b
etter p
r
eterm
b
ir
th
…
(
Hima
n
i D
esh
p
a
n
d
e
)
6137
3.
3
.
E
x
perim
ent
a
l r
esu
lt
s
o
n sta
nd
a
rd
da
t
a
s
et
s
Fo
r
v
alid
atio
n
o
f
th
e
e
f
f
icien
cy
o
f
th
e
p
r
o
p
o
s
ed
MPM
U
al
g
o
r
ith
m
,
ex
p
e
r
im
en
ts
ar
e
co
n
d
u
cted
o
n
s
ev
en
s
tan
d
ar
d
d
atasets
u
s
in
g
f
iv
e
d
if
f
er
e
n
t
class
if
ier
s
.
T
h
ese
d
atasets
ar
e
co
m
m
o
n
ly
u
s
ed
b
y
r
esear
ch
e
r
s
in
th
e
im
b
alan
ce
d
h
an
d
lin
g
liter
a
tu
r
e.
T
h
e
n
u
m
b
er
o
f
clu
s
ter
s
f
o
r
m
ed
u
s
in
g
MPM
U,
f
o
r
ea
ch
o
f
th
e
d
atasets
ar
e
as
s
h
o
wn
in
T
ab
le
2
.
I
t
is
o
b
s
er
v
ed
th
at
th
e
MPM
U
alg
o
r
ith
m
is
o
v
er
s
h
in
in
g
o
th
e
r
alg
o
r
it
h
m
s
f
o
r
m
o
s
t
o
f
th
e
d
atasets
an
d
class
if
ier
co
m
b
in
atio
n
s
.
R
esu
lts
in
T
ab
le
4
,
d
ep
ict
th
at
th
e
p
r
o
p
o
s
ed
MPM
U
m
eth
o
d
wo
r
k
s
well
with
lar
g
e
s
ize
d
atasets
lik
e
v
o
wel0
an
d
s
eg
m
en
t0
.
Fo
r
s
m
aller
d
atasets
lik
e
Pima
I
n
d
ia
a
n
d
ec
o
li3
th
e
f
in
al
r
esu
lts
ar
e
n
o
t
v
er
y
g
o
o
d
b
u
t
t
h
er
e
is
a
r
em
ar
k
a
b
le
im
p
r
o
v
e
m
en
t
in
r
esu
lts
wh
en
co
m
p
ar
e
d
with
SV,
C
S
an
d
R
OS
im
b
alan
ce
h
an
d
lin
g
m
e
th
o
d
s
.
W
h
ile
co
m
p
ar
in
g
th
e
r
esu
lts
o
f
SV
o
v
er
th
e
r
es
u
lts
o
f
th
e
MPM
U
alg
o
r
ith
m
,
im
p
r
o
v
e
m
en
t
in
r
es
u
lts
h
as
b
ee
n
o
b
s
er
v
ed
b
etwe
e
n
0
%
to
2
1
%
f
o
r
p
r
ec
is
io
n
,
b
e
twee
n
0
%
to
1
0
%
f
o
r
AUC
an
d
b
etwe
en
0
%
to
1
7
%
f
o
r
p
r
ed
ictio
n
ac
cu
r
ac
y
.
T
h
e
r
esu
lts
in
th
e
tab
le
co
n
v
e
y
th
at
alo
n
g
with
th
e
MSF
d
ataset,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
s
wo
r
k
eq
u
ally
well
f
o
r
o
th
er
im
b
alan
ce
d
d
atasets
wit
h
v
ar
ied
s
izes
an
d
im
b
alan
ce
s
.
T
ab
le
4.
C
o
m
p
a
r
ativ
e
r
esu
lts
of
d
if
f
e
r
en
t
im
b
ala
n
ce
h
an
d
lin
g
m
eth
o
d
s
on
s
tan
d
ar
d
im
b
ala
n
ce
d
d
atasets
D
a
t
a
s
e
t
s
D
ec
i
s
i
o
n
T
re
e
Ran
d
o
m
F
o
re
s
t
KNN
G
au
s
s
i
a
n
N
B
ML
P
SV
CS
RO
S
MPM
U
SV
CS
RO
S
MPM
U
SV
CS
RO
S
MPM
U
SV
CS
RO
S
MPM
U
SV
CS
RO
S
MPM
U
Prec
i
s
i
o
n
eco
l
i
3
0
.
4
7
0
.
3
2
0
.
3
9
0
.
6
8
0
.
7
2
0
.
6
8
0
.
6
5
0
.
8
6
0
.
4
6
0
.
4
0
.
3
7
0
.
6
6
0
.
6
3
0
.
6
0
.
6
5
0
.
8
0
.
6
6
0
.
6
6
0
.
6
6
0
.
8
1
Pi
m
a
I
n
d
i
a
0
.
4
7
0
.
4
9
0
.
4
6
0
.
4
7
0
.
7
1
0
.
7
0
.
7
0
.
7
3
0
.
5
6
0
.
5
7
0
.
5
6
0
.
5
8
0
.
6
7
0
.
6
6
0
.
6
7
0
.
7
2
0
.
7
0
.
7
0
.
7
0
.
7
8
y
ea
s
t
-
2
_
v
s
_
4
0
.
6
5
0
.
5
7
0
.
5
6
0
.
6
7
0
.
9
2
0
.
9
3
0
.
9
2
0
.
9
4
0
.
6
3
0
.
6
4
0
.
5
9
0
.
8
8
0
.
7
2
0
.
7
6
0
.
7
2
0
.
8
3
0
.
8
4
0
.
8
4
0
.
8
4
0
.
8
8
p
a
g
e
-
b
l
o
ck
s
0
0
.
7
2
0
.
7
1
0
.
7
0
.
9
2
0
.
9
4
0
.
9
4
0
.
9
3
0
.
9
9
0
.
7
8
0
.
7
8
0
.
7
2
0
.
8
0
.
5
8
0
.
5
4
0
.
5
8
0
.
5
5
0
.
8
9
0
.
8
9
0
.
8
9
0
.
9
2
s
e
g
me
n
t
0
0
.
9
5
0
.
9
4
0
.
9
5
0
.
9
5
1
1
1
1
0
.
9
9
0
.
9
8
0
.
9
7
1
0
.
9
2
0
.
8
0
.
9
2
0
.
9
9
1
1
1
1
v
o
w
el
0
0
.
8
5
0
.
7
9
0
.
8
2
0
.
9
0
.
9
9
1
0
.
9
9
1
1
1
0
.
9
9
1
0
.
8
8
0
.
5
8
0
.
8
7
0
.
9
2
1
1
1
1
v
e
h
i
cl
e2
0
.
8
5
0
.
8
6
0
.
8
5
0
.
8
9
1
0
.
9
9
1
1
0
.
9
3
0
.
9
4
0
.
8
8
0
.
9
6
0
.
6
2
0
.
5
4
0
.
6
1
0
.
7
6
0
.
9
9
0
.
9
9
0
.
9
9
1
AUC
eco
l
i
3
0
.
7
9
0
.
7
1
0
.
4
9
0
.
8
9
0
.
9
5
0
.
9
4
0
.
5
3
0
.
9
5
0
.
8
7
0
.
8
6
0
.
5
0
.
8
7
0
.
9
0
.
9
0
.
5
0
.
9
1
0
.
9
5
0
.
9
5
0
.
5
1
0
.
9
8
Pi
m
a
I
n
d
i
a
0
.
6
7
0
.
6
8
0
.
5
0
.
6
8
0
.
8
2
0
.
8
3
0
.
5
0
.
8
5
0
.
7
4
0
.
7
4
0
.
4
9
0
.
7
7
0
.
8
1
0
.
8
0
.
5
1
0
.
8
4
0
.
8
3
0
.
8
3
0
.
4
9
0
.
8
7
y
ea
s
t
-
2
_
v
s
_
4
0
.
8
8
0
.
8
6
0
.
5
2
0
.
9
2
1
1
0
.
5
1
0
.
9
9
0
.
8
6
0
.
8
6
0
.
5
1
0
.
9
5
0
.
9
0
.
8
8
0
.
5
0
.
9
0
.
9
4
0
.
9
4
0
.
4
3
0
.
9
6
p
a
g
e
-
b
l
o
ck
s
0
0
.
9
1
0
.
9
1
0
.
4
9
0
.
9
8
0
.
9
9
0
.
9
9
0
.
4
9
1
0
.
9
3
0
.
9
3
0
.
5
1
0
.
9
3
0
.
9
3
0
.
8
9
0
.
5
0
.
9
1
0
.
9
8
0
.
9
8
0
.
4
9
0
.
9
9
s
e
g
me
n
t
0
0
.
9
8
0
.
9
8
0
.
5
0
.
9
9
1
1
0
.
5
1
0
.
9
9
0
.
9
9
0
.
5
1
0
.
9
8
0
.
9
6
0
.
4
8
1
1
1
0
.
5
1
1
v
o
w
el
0
0
.
9
7
0
.
9
5
0
.
5
1
0
.
9
7
1
1
0
.
5
1
1
1
0
.
5
1
1
0
.
9
8
0
.
9
0
.
4
7
0
.
9
9
1
1
0
.
4
9
1
v
e
h
i
cl
e2
0
.
9
4
0
.
9
5
0
.
4
9
0
.
9
6
1
1
0
.
5
3
1
0
.
9
8
0
.
9
8
0
.
5
3
0
.
9
9
0
.
8
2
0
.
7
1
0
.
4
7
0
.
9
1
1
0
.
5
1
A
ccu
r
ac
y
eco
l
i
3
9
1
.
1
8
7
.
3
8
9
.
9
4
9
5
.
8
8
9
4
.
1
9
2
.
9
9
1
.
8
9
5
.
8
8
9
1
.
8
8
9
.
8
8
8
.
8
9
3
.
2
4
7
5
.
1
7
5
.
7
7
3
.
4
8
1
.
6
9
8
9
.
8
8
9
.
8
8
9
.
8
9
1
.
5
9
Pi
m
a
I
n
d
i
a
6
9
.
5
7
0
.
6
6
8
.
5
5
6
9
.
8
4
7
5
.
3
7
6
.
1
7
5
.
5
7
8
.
2
1
7
1
.
1
71
6
9
.
2
7
4
.
4
6
7
4
.
4
7
3
.
3
7
2
.
8
7
7
.
4
2
7
4
.
8
7
4
.
8
7
4
.
8
7
9
.
4
4
y
ea
s
t
-
2
_
v
s
_
4
9
5
.
5
9
4
.
1
9
4
.
0
9
9
5
.
9
1
9
5
.
9
9
6
.
1
9
5
.
9
9
7
.
5
3
9
5
.
3
9
5
.
2
94
9
7
.
7
4
3
1
.
3
49
2
7
.
7
3
3
.
9
8
9
3
.
1
9
3
.
1
9
3
.
1
9
1
.
6
1
p
a
g
e
-
b
l
o
ck
s
0
9
6
.
6
9
6
.
6
9
6
.
4
8
9
9
.
1
1
9
7
.
6
9
7
.
6
9
7
.
6
9
9
.
4
1
9
6
.
1
9
6
.
1
9
5
.
3
9
6
.
0
2
8
9
.
7
90
8
9
.
1
8
9
.
3
1
9
4
.
8
9
4
.
8
9
4
.
8
9
4
.
8
1
s
e
g
me
n
t
0
9
9
.
2
9
9
.
1
9
9
.
2
1
9
9
.
2
6
9
9
.
7
9
9
.
8
9
9
.
7
9
9
.
7
6
9
9
.
4
9
9
.
3
9
9
.
3
9
9
.
8
8
8
2
.
2
8
8
.
8
7
9
.
9
9
8
.
9
2
9
9
.
7
9
9
.
8
9
9
.
7
9
8
.
5
7
v
o
w
el
0
9
8
.
5
9
7
.
8
9
8
.
1
5
9
8
.
9
9
9
9
.
4
9
9
.
6
9
9
.
6
9
9
.
6
1
9
9
.
9
9
9
.
9
9
9
.
8
9
9
.
7
2
9
2
.
7
8
9
.
2
8
6
.
8
9
5
.
7
4
9
9
.
8
9
9
.
8
9
9
.
8
9
9
.
7
7
4.
CO
NCLU
SI
O
N
T
h
is
wo
r
k
p
r
o
p
o
s
es
a
n
o
v
el
im
b
alan
ce
h
a
n
d
lin
g
m
o
d
el
t
o
war
d
s
th
e
p
r
e
d
ictio
n
o
f
p
r
e
-
t
er
m
b
ir
th
u
s
in
g
th
e
cr
ea
ted
MSF
d
atas
et,
wh
ich
co
n
s
is
ts
o
f
l
if
esty
l
e,
s
o
cial,
p
h
y
s
ical
an
d
s
tr
es
s
f
ea
tu
r
es
of
1
,
000
m
o
th
er
s
.
T
h
e
p
r
o
p
o
s
ed
MPM
U
im
b
alan
ce
h
an
d
lin
g
m
eth
o
d
wo
r
k
s
on
class
in
s
ep
ar
ab
ilit
y
an
d
class
if
icatio
n
b
ias
is
s
u
es
by
s
u
b
d
u
in
g
th
e
m
ajo
r
ity
class
in
s
tan
ce
s
f
o
llo
win
g
a
clu
s
ter
-
b
ased
p
en
alty
ap
p
r
o
ac
h
an
d
s
tr
en
g
th
en
in
g
th
e
m
in
o
r
ity
i
n
s
tan
ce
s
by
r
an
d
o
m
o
v
er
s
am
p
li
n
g
.
It
h
as
b
ee
n
o
b
s
er
v
ed
th
at
th
e
p
r
o
p
o
s
ed
m
o
d
el
co
n
s
is
tin
g
of
MPM
U
an
d
6L
-
ANN
class
if
ier
,
is
h
ig
h
ly
ef
f
icien
t
in
class
if
y
in
g
PTB
an
d
FTB
o
u
tco
m
es
on
MSF
d
ataset
s
with
p
r
ec
is
io
n
r
an
g
in
g
f
r
o
m
0
.
9
0
to
0
.
9
7
,
AU
C
b
etwe
en
0
.
8
6
to
0
.
9
9
,
an
d
p
r
ed
ictio
n
ac
cu
r
ac
y
f
allin
g
u
n
d
e
r
93
%
to
9
9
.
4
7
%.
MPM
U
im
b
alan
ce
h
an
d
lin
g
m
eth
o
d
h
as
b
ee
n
v
alid
ated
on
o
th
er
im
b
alan
ce
d
d
atasets
to
o
.
MPM
U
s
h
o
ws
a
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
wh
e
n
c
o
m
p
ar
ed
to
o
th
er
I
m
b
alan
ce
h
a
n
d
lin
g
m
eth
o
d
s
on
MSF
as
we
ll
as
o
th
er
d
ataset
s
.
T
h
u
s
,
th
is
r
esear
ch
wo
r
k
s
u
cc
ee
d
s
in
p
r
o
p
o
s
in
g
an
e
f
f
i
cien
t
clu
s
ter
-
b
ased
im
b
alan
ce
d
in
s
tan
ce
h
an
d
lin
g
m
eth
o
d
.
E
x
p
er
im
e
n
tal
r
esu
lts
on
th
e
m
en
tal
an
d
p
h
y
s
ical
h
ea
lth
-
r
elate
d
f
ea
tu
r
es
of
th
e
m
o
th
er
,
s
h
o
w
th
at
th
e
life
s
ty
le
o
f
th
e
m
o
th
er
an
d
Stre
s
s
-
r
elate
d
f
ea
tu
r
es
ar
e
m
ajo
r
co
n
tr
ib
u
t
o
r
s
to
war
d
s
an
aly
zin
g
h
e
r
m
ater
n
al
h
ea
lth
wh
ich
d
ef
in
es
h
er
m
en
tal
h
ea
lth
.
T
h
u
s
,
th
is
s
tu
d
y
,
co
n
clu
d
es
th
at
a
m
o
th
er
’
s
m
en
tal
h
ea
lth
is
a
m
ajo
r
co
n
tr
i
b
u
to
r
t
o
war
d
s
u
n
d
er
s
tan
d
in
g
th
e
p
r
e
-
ter
m
b
ir
th
p
o
s
s
ib
ilit
ie
s
.
RE
F
E
R
E
NC
E
S
[
1
]
T.
S
t
r
u
n
k
,
A
.
C
u
r
r
i
e
,
P
.
R
i
c
h
m
o
n
d
,
K
.
S
i
mm
e
r
,
a
n
d
D
.
B
u
r
g
n
e
r
,
“
I
n
n
a
t
e
i
m
mu
n
i
t
y
i
n
h
u
ma
n
n
e
w
b
o
r
n
i
n
f
a
n
t
s
:
p
r
e
mat
u
r
i
t
y
me
a
n
s
mo
r
e
t
h
a
n
i
mm
a
t
u
r
i
t
y
,
”
T
h
e
J
o
u
rn
a
l
o
f
M
a
t
e
rn
a
l
-
Fe
t
a
l
a
n
d
N
e
o
n
a
t
a
l
Me
d
i
c
i
n
e
,
v
o
l
.
2
4
,
n
o
.
1
,
p
p
.
2
5
–
3
1
,
Ja
n
.
2
0
1
1
,
d
o
i
:
1
0
.
3
1
0
9
/
1
4
7
6
7
0
5
8
.
2
0
1
0
.
4
8
2
6
0
5
.
[
2
]
M
.
A
.
A
r
f
a
n
d
i
,
A
.
A
n
sari
a
d
i
,
R
.
A
m
i
r
u
d
d
i
n
,
W
.
W
a
h
i
d
u
d
d
i
n
,
A
.
U
.
S
a
l
m
a
h
,
a
n
d
A
.
S
a
l
a
m,
“
P
r
e
t
e
r
m
b
i
r
t
h
r
i
s
k
i
n
m
o
t
h
e
r
w
i
t
h
h
y
p
e
r
t
e
n
si
v
e
d
i
s
o
r
d
e
r
s
o
f
p
r
e
g
n
a
n
c
y
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
P
u
b
l
i
c
H
e
a
l
t
h
S
c
i
e
n
c
e
,
v
o
l
.
1
2
,
n
o
.
2
,
p
p
.
5
9
0
–
5
9
7
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
p
h
s.
v
1
2
i
2
.
2
2
5
9
9
.
[
3
]
G
.
H
o
ma
n
,
J.
L
i
t
t
,
a
n
d
R
.
J
.
N
o
r
man
,
“
T
h
e
F
A
S
T
st
u
d
y
:
F
e
r
t
i
l
i
t
y
a
ssessme
n
t
a
n
d
a
d
v
i
c
e
t
a
r
g
e
t
i
n
g
l
i
f
e
st
y
l
e
c
h
o
i
c
e
s
a
n
d
b
e
h
a
v
i
o
u
r
s:
A
p
i
l
o
t
st
u
d
y
,
”
H
u
m
a
n
R
e
p
ro
d
u
c
t
i
o
n
,
v
o
l
.
2
7
,
n
o
.
8
,
p
p
.
2
3
9
6
–
2
4
0
4
,
2
0
1
2
,
d
o
i
:
1
0
.
1
0
9
3
/
h
u
mr
e
p
/
d
e
s
1
7
6
.
[
4
]
S
.
S
.
K
a
d
d
i
a
n
d
M
.
M
.
P
a
t
i
l
,
“
E
n
s
e
mb
l
e
l
e
a
r
n
i
n
g
b
a
s
e
d
h
e
a
l
t
h
c
a
r
e
c
l
a
i
m
f
r
a
u
d
d
e
t
e
c
t
i
o
n
i
n
a
n
i
m
b
a
l
a
n
c
e
d
a
t
a
e
n
v
i
r
o
n
me
n
t
,
”
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
e
r
20
24
:
6
1
2
9
-
6
1
3
9
6138
I
n
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
E
l
e
c
t
r
i
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
3
2
,
n
o
.
3
,
p
p
.
1
6
8
6
–
1
6
9
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s.
v
3
2
.
i
3
.
p
p
1
6
8
6
-
1
6
9
4
.
[
5
]
G
.
N
a
i
k
,
D
.
S
i
r
o
y
a
,
M
.
N
i
s
a
r
,
B
.
S
h
a
h
,
a
n
d
H
.
D
e
s
h
p
a
n
d
e
,
“
A
n
a
l
y
s
i
s
o
n
t
h
e
e
f
f
i
c
a
c
y
o
f
A
N
N
o
n
sm
a
l
l
i
m
b
a
l
a
n
c
e
d
d
a
t
a
s
e
t
s
,
”
S
p
r
i
n
g
e
r P
r
o
c
e
e
d
i
n
g
s
i
n
M
a
t
h
e
m
a
t
i
c
s
a
n
d
S
t
a
t
i
s
t
i
c
s
,
v
o
l
.
4
0
3
,
p
p
.
1
2
9
–
1
3
7
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
1
-
1
6
1
7
8
-
0
_
1
1
.
[
6
]
M
.
H
.
B
.
M
.
R
a
z
a
l
i
,
R
.
B
i
n
S
a
i
a
n
,
Y
.
B
.
W
a
h
,
a
n
d
K
.
R
.
K
u
-
M
a
h
a
mu
d
,
“
A
c
l
a
ss
s
k
e
w
-
i
n
se
n
si
t
i
v
e
A
C
O
-
b
a
sed
d
e
c
i
si
o
n
t
r
e
e
a
l
g
o
r
i
t
h
m
f
o
r
i
m
b
a
l
a
n
c
e
d
d
a
t
a
s
e
t
s
,
”
I
n
d
o
n
e
s
i
a
n
J
o
u
rn
a
l
o
f
El
e
c
t
ri
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
2
1
,
n
o
.
1
,
p
p
.
4
1
2
–
4
1
9
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s.v
2
1
.
i
1
.
p
p
4
1
2
-
4
1
9
.
[
7
]
M
.
S
o
n
a
n
d
E
.
S
.
M
i
l
l
e
r
,
“
P
r
e
d
i
c
t
i
n
g
p
r
e
t
e
r
m
b
i
r
t
h
:
C
e
r
v
i
c
a
l
l
e
n
g
t
h
a
n
d
f
e
t
a
l
f
i
b
r
o
n
e
c
t
i
n
,
”
S
e
m
i
n
a
r
s
i
n
Pe
ri
n
a
t
o
l
o
g
y
,
v
o
l
.
4
1
,
n
o
.
8
,
p
p
.
4
4
5
–
4
5
1
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
5
3
/
j
.
se
mp
e
r
i
.
2
0
1
7
.
0
8
.
0
0
2
.
[
8
]
P
.
B
r
a
n
c
o
,
L.
T
o
r
g
o
,
a
n
d
R
.
P
.
R
i
b
e
i
r
o
,
“
A
s
u
r
v
e
y
o
f
p
r
e
d
i
c
t
i
v
e
mo
d
e
l
i
n
g
o
n
i
m
b
a
l
a
n
c
e
d
d
o
mai
n
s,
”
AC
M
C
o
m
p
u
t
i
n
g
S
u
rve
y
s
,
v
o
l
.
4
9
,
n
o
.
2
,
2
0
1
6
,
d
o
i
:
1
0
.
1
1
4
5
/
2
9
0
7
0
7
0
.
[
9
]
F
.
N
i
e
t
o
-
D
e
l
-
a
m
o
r
e
t
a
l
.
,
“
C
o
mb
i
n
a
t
i
o
n
o
f
f
e
a
t
u
r
e
se
l
e
c
t
i
o
n
a
n
d
r
e
sa
mp
l
i
n
g
m
e
t
h
o
d
s
t
o
p
r
e
d
i
c
t
p
r
e
t
e
r
m
b
i
r
t
h
b
a
se
d
o
n
e
l
e
c
t
r
o
h
y
st
e
r
o
g
r
a
p
h
i
c
si
g
n
a
l
s fr
o
m
i
m
b
a
l
a
n
c
e
d
a
t
a
,
”
S
e
n
so
rs
,
v
o
l
.
2
2
,
n
o
.
1
4
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
2
1
4
5
0
9
8
.
[
1
0
]
M
.
H
.
S
a
e
e
d
a
n
d
J.
I
.
H
a
ma
,
“
C
a
r
d
i
a
c
d
i
s
e
a
s
e
p
r
e
d
i
c
t
i
o
n
u
s
i
n
g
A
I
a
l
g
o
r
i
t
h
ms
w
i
t
h
S
e
l
e
c
t
K
B
e
st
,
”
Me
d
i
c
a
l
a
n
d
Bi
o
l
o
g
i
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
6
1
,
n
o
.
1
2
,
p
p
.
3
3
9
7
–
3
4
0
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
5
1
7
-
0
2
3
-
0
2
9
1
8
-
8.
[
1
1
]
W
.
A
.
K
u
s
u
ma
,
N
.
N
o
v
i
a
n
a
,
L.
S
.
H
a
si
b
u
a
n
,
a
n
d
M
.
N
u
r
i
l
ma
l
a
,
“
I
mp
r
o
v
i
n
g
D
N
A
b
a
r
c
o
d
e
-
b
a
se
d
f
i
s
h
i
d
e
n
t
i
f
i
c
a
t
i
o
n
sy
st
e
m
o
n
i
mb
a
l
a
n
c
e
d
d
a
t
a
u
si
n
g
S
M
O
TE,
”
T
e
l
k
o
m
n
i
k
a
(
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
C
o
m
p
u
t
i
n
g
E
l
e
c
t
r
o
n
i
c
s
a
n
d
C
o
n
t
r
o
l
)
,
v
o
l
.
1
5
,
n
o
.
3
,
p
p
.
1
2
3
0
–
1
2
3
8
,
2
0
1
7
,
d
o
i
:
1
0
.
1
2
9
2
8
/
TEL
K
O
M
N
I
K
A
.
v
1
5
i
3
.
5
0
1
1
.
[
1
2
]
B
.
Z
h
u
,
X
.
P
a
n
,
S
.
v
a
n
d
e
n
B
r
o
u
c
k
e
,
a
n
d
J.
X
i
a
o
,
“
A
G
A
N
-
b
a
s
e
d
h
y
b
r
i
d
sa
mp
l
i
n
g
m
e
t
h
o
d
f
o
r
i
mb
a
l
a
n
c
e
d
c
u
s
t
o
mer c
l
a
ssi
f
i
c
a
t
i
o
n
,
”
I
n
f
o
rm
a
t
i
o
n
S
c
i
e
n
c
e
s
,
v
o
l
.
6
0
9
,
p
p
.
1
3
9
7
–
1
4
1
1
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
n
s.
2
0
2
2
.
0
7
.
1
4
5
.
[
1
3
]
S
.
A
l
i
,
P
.
C
h
o
u
r
a
s
i
a
,
Z.
T
a
y
e
b
i
,
B
.
B
e
l
l
o
,
a
n
d
M
.
P
a
t
t
e
r
so
n
,
“
V
i
r
a
l
V
e
c
t
o
r
s:
c
o
mp
a
c
t
a
n
d
s
c
a
l
a
b
l
e
a
l
i
g
n
m
e
n
t
-
f
r
e
e
v
i
r
o
m
e
f
e
a
t
u
r
e
g
e
n
e
r
a
t
i
o
n
,
”
a
r
Xi
v
p
re
p
ri
n
t
a
r
Xi
v
:
2
3
0
4
.
0
2
8
9
1
,
A
p
r
.
2
0
2
3
.
[
1
4
]
G
.
A
n
d
r
e
s
i
n
i
,
A
.
A
p
p
i
c
e
,
a
n
d
D
.
M
a
l
e
r
b
a
,
“
D
e
a
l
i
n
g
w
i
t
h
c
l
a
ss
i
mb
a
l
a
n
c
e
i
n
a
n
d
r
o
i
d
ma
l
w
a
r
e
d
e
t
e
c
t
i
o
n
b
y
c
a
sc
a
d
i
n
g
c
l
u
s
t
e
r
i
n
g
a
n
d
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
S
t
u
d
i
e
s
i
n
C
o
m
p
u
t
a
t
i
o
n
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
8
8
0
,
p
p
.
1
7
3
–
1
8
7
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
3
6
6
1
7
-
9
_
1
1
.
[
1
5
]
R
.
C
h
o
u
d
h
a
r
y
a
n
d
S
.
S
h
u
k
l
a
,
“
A
c
l
u
s
t
e
r
i
n
g
b
a
se
d
e
n
s
e
m
b
l
e
o
f
w
e
i
g
h
t
e
d
k
e
r
n
e
l
i
z
e
d
e
x
t
r
e
me
l
e
a
r
n
i
n
g
m
a
c
h
i
n
e
f
o
r
c
l
a
ss
i
mb
a
l
a
n
c
e
l
e
a
r
n
i
n
g
,
”
Ex
p
e
r
t
S
y
s
t
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
6
4
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
2
0
.
1
1
4
0
4
1
.
[
1
6
]
H
.
H
.
H
t
u
n
,
M
.
B
i
e
h
l
,
a
n
d
N
.
P
e
t
k
o
v
,
“
S
u
r
v
e
y
o
f
f
e
a
t
u
r
e
sel
e
c
t
i
o
n
a
n
d
e
x
t
r
a
c
t
i
o
n
t
e
c
h
n
i
q
u
e
s
f
o
r
s
t
o
c
k
m
a
r
k
e
t
p
r
e
d
i
c
t
i
o
n
,
”
Fi
n
a
n
c
i
a
l
I
n
n
o
v
a
t
i
o
n
,
v
o
l
.
9
,
n
o
.
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
8
6
/
s
4
0
8
5
4
-
022
-
0
0
4
4
1
-
7.
[
1
7
]
P
.
D
h
a
l
a
n
d
C
.
A
z
a
d
,
“
A
c
o
m
p
r
e
h
e
n
s
i
v
e
s
u
r
v
e
y
o
n
f
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
i
n
t
h
e
v
a
r
i
o
u
s
f
i
e
l
d
s
o
f
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
A
p
p
l
i
e
d
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
5
2
,
n
o
.
4
,
p
p
.
4
5
4
3
–
4
5
8
1
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
4
8
9
-
0
2
1
-
0
2
5
5
0
-
9.
[
1
8
]
J.
W
.
W
a
n
a
n
d
M
.
Y
a
n
g
,
“
S
u
r
v
e
y
o
n
c
o
s
t
-
se
n
si
t
i
v
e
l
e
a
r
n
i
n
g
m
e
t
h
o
d
,
”
R
u
a
n
J
i
a
n
X
u
e
Ba
o
/
J
o
u
r
n
a
l
o
f
S
o
f
t
w
a
re
,
v
o
l
.
3
1
,
n
o
.
1
,
p
p
.
1
1
3
–
1
3
6
,
2
0
2
0
,
d
o
i
:
1
0
.
1
3
3
2
8
/
j
.
c
n
k
i
.
j
o
s.
0
0
5
8
7
1
.
[
1
9
]
A
.
Te
l
i
k
a
n
i
,
A
.
H
.
G
a
n
d
o
m
i
,
K
.
K
.
R
.
C
h
o
o
,
a
n
d
J
.
S
h
e
n
,
“
A
c
o
st
-
se
n
s
i
t
i
v
e
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
a
p
p
r
o
a
c
h
f
o
r
n
e
t
w
o
r
k
t
r
a
f
f
i
c
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
I
EE
E
T
ra
n
s
a
c
t
i
o
n
s
o
n
N
e
t
w
o
rk
a
n
d
S
e
rvi
c
e
Ma
n
a
g
e
m
e
n
t
,
v
o
l
.
1
9
,
n
o
.
1
,
p
p
.
6
6
1
–
6
7
0
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
TN
S
M
.
2
0
2
1
.
3
1
1
2
2
8
3
.
[
2
0
]
F
.
C
h
e
n
g
,
J.
Zh
a
n
g
,
a
n
d
C
.
W
e
n
,
“
C
o
st
-
sen
si
t
i
v
e
l
a
r
g
e
mar
g
i
n
d
i
st
r
i
b
u
t
i
o
n
mac
h
i
n
e
f
o
r
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
i
mb
a
l
a
n
c
e
d
d
a
t
a
,
”
P
a
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
L
e
t
t
e
rs
,
v
o
l
.
8
0
,
p
p
.
1
0
7
–
1
1
2
,
2
0
1
6
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
a
t
r
e
c
.
2
0
1
6
.
0
6
.
0
0
9
.
[
2
1
]
R
.
G
u
n
a
w
a
n
,
H
.
A
.
G
h
a
n
i
,
N
.
K
h
a
mi
s,
J
.
A
l
A
mi
e
n
,
a
n
d
E.
I
sman
t
o
,
“
D
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
t
o
D
D
o
S
a
t
t
a
c
k
w
i
t
h
i
m
b
a
l
a
n
c
e
d
d
a
t
a
a
t
t
h
e
a
p
p
l
i
c
a
t
i
o
n
l
a
y
e
r
,
”
T
e
l
k
o
m
n
i
k
a
(
T
e
l
e
c
o
m
m
u
n
i
c
a
t
i
o
n
C
o
m
p
u
t
i
n
g
El
e
c
t
ro
n
i
c
s
a
n
d
C
o
n
t
r
o
l
)
,
v
o
l
.
2
1
,
n
o
.
5
,
p
p
.
1
0
6
0
–
1
0
6
7
,
2
0
2
3
,
d
o
i
:
1
0
.
1
2
9
2
8
/
TEL
K
O
M
N
I
K
A
.
v
2
1
i
5
.
2
4
8
5
7
.
[
2
2
]
G
.
C
a
s
a
ñ
o
l
a
-
M
a
r
t
i
n
e
t
a
l
.
,
“
E
x
p
l
o
r
i
n
g
d
i
f
f
e
r
e
n
t
st
r
a
t
e
g
i
e
s fo
r
i
mb
a
l
a
n
c
e
d
A
D
M
E
d
a
t
a
p
r
o
b
l
e
m:
c
a
se
s
t
u
d
y
o
n
C
a
c
o
-
2
p
e
r
mea
b
i
l
i
t
y
mo
d
e
l
i
n
g
,
”
Mo
l
e
c
u
l
a
r
d
i
v
e
rsi
t
y
,
v
o
l
.
2
0
,
n
o
.
1
,
p
p
.
9
3
–
1
0
9
,
2
0
1
6
.
[
2
3
]
J.
C
h
e
n
,
Z.
W
u
,
a
n
d
J
.
Z
h
a
n
g
,
“
D
r
i
v
i
n
g
safe
t
y
r
i
sk
p
r
e
d
i
c
t
i
o
n
u
s
i
n
g
c
o
st
-
sen
s
i
t
i
v
e
w
i
t
h
n
o
n
n
e
g
a
t
i
v
i
t
y
-
c
o
n
st
r
a
i
n
e
d
a
u
t
o
e
n
c
o
d
e
r
s
b
a
s
e
d
o
n
i
m
b
a
l
a
n
c
e
d
n
a
t
u
r
a
l
i
st
i
c
d
r
i
v
i
n
g
d
a
t
a
,
”
I
EEE
T
r
a
n
s
a
c
t
i
o
n
s
o
n
I
n
t
e
l
l
i
g
e
n
t
T
r
a
n
s
p
o
rt
a
t
i
o
n
S
y
s
t
e
m
s
,
v
o
l
.
2
0
,
n
o
.
1
2
,
p
p
.
4
4
5
0
–
4
4
6
5
,
D
e
c
.
2
0
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
TI
TS.2
0
1
8
.
2
8
8
6
2
8
0
.
[
2
4
]
M
.
G
.
M
a
d
d
e
n
a
n
d
S
.
S
.
K
h
a
n
,
“
O
n
e
-
c
l
a
ss
c
l
a
ssi
f
i
c
a
t
i
o
n
:
t
a
x
o
n
o
m
y
o
f
st
u
d
y
a
n
d
r
e
v
i
e
w
o
f
t
e
c
h
n
i
q
u
e
s,
”
K
n
o
w
l
e
d
g
e
En
g
i
n
e
e
ri
n
g
Re
v
i
e
w
,
v
o
l
.
2
0
,
n
o
.
2
,
p
p
.
1
1
7
–
1
2
5
,
2
0
0
4
.
[
2
5
]
H
.
J.
Le
e
a
n
d
S
.
C
h
o
,
“
Th
e
n
o
v
e
l
t
y
d
e
t
e
c
t
i
o
n
a
p
p
r
o
a
c
h
f
o
r
d
i
f
f
e
r
e
n
t
d
e
g
r
e
e
s
o
f
c
l
a
ss
i
m
b
a
l
a
n
c
e
,
”
L
e
c
t
u
re
N
o
t
e
s
i
n
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
i
n
c
l
u
d
i
n
g
s
u
b
s
e
ri
e
s
L
e
c
t
u
re
N
o
t
e
s
i
n
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
L
e
c
t
u
re
N
o
t
e
s
i
n
Bi
o
i
n
f
o
rm
a
t
i
c
s)
,
p
p
.
2
1
–
3
0
,
2
0
0
6
,
d
o
i
:
1
0
.
1
0
0
7
/
1
1
8
9
3
2
5
7
_
3
.
[
2
6
]
G
.
H
a
i
x
i
a
n
g
,
L.
Y
i
j
i
n
g
,
J.
S
h
a
n
g
,
G
.
M
i
n
g
y
u
n
,
H
.
Y
u
a
n
y
u
e
,
a
n
d
G
.
B
i
n
g
,
“
Le
a
r
n
i
n
g
f
r
o
m
c
l
a
ss
-
i
m
b
a
l
a
n
c
e
d
d
a
t
a
:
R
e
v
i
e
w
o
f
met
h
o
d
s
a
n
d
a
p
p
l
i
c
a
t
i
o
n
s
,
”
Ex
p
e
r
t
S
y
st
e
m
s
w
i
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
7
3
,
p
p
.
2
2
0
–
2
3
9
,
M
a
y
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
s
w
a
.
2
0
1
6
.
1
2
.
0
3
5
.
[
2
7
]
L.
R
o
k
a
c
h
,
“
En
s
e
mb
l
e
-
b
a
s
e
d
c
l
a
ss
i
f
i
e
r
s,”
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
Re
v
i
e
w
,
v
o
l
.
3
3
,
n
o
.
1
–
2
,
p
p
.
1
–
3
9
,
2
0
1
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
4
6
2
-
0
0
9
-
9
1
2
4
-
7.
[
2
8
]
N
.
H
.
A
.
M
a
l
e
k
,
W
.
F
.
W
.
Y
a
a
c
o
b
,
Y
.
B
.
W
a
h
,
S
.
A
.
M
d
N
a
s
i
r
,
N
.
S
h
a
a
d
a
n
,
a
n
d
S
.
W
.
I
n
d
r
a
t
n
o
,
“
C
o
mp
a
r
i
s
o
n
o
f
e
n
s
e
mb
l
e
h
y
b
r
i
d
samp
l
i
n
g
w
i
t
h
b
a
g
g
i
n
g
a
n
d
b
o
o
s
t
i
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
f
o
r
i
mb
a
l
a
n
c
e
d
d
a
t
a
,
”
I
n
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
2
9
,
n
o
.
1
,
p
p
.
5
9
8
–
6
0
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s.
v
2
9
.
i
1
.
p
p
5
9
8
-
6
0
8
.
[
2
9
]
H
.
D
e
sh
p
a
n
d
e
a
n
d
L.
R
a
g
h
a
,
“
R
e
a
l
i
z
i
n
g
m
o
t
h
e
r
’
s
f
e
a
t
u
r
e
s
i
n
f
l
u
e
n
t
i
a
l
o
n
c
h
i
l
d
b
i
r
t
h
e
x
p
e
r
i
e
n
c
e
,
t
o
w
a
r
d
s
c
r
e
a
t
i
o
n
o
f
a
d
a
t
a
s
e
t
,
”
D
a
t
a
S
c
i
e
n
c
e
,
p
p
.
1
4
3
–
1
6
7
,
2
0
2
2
,
d
o
i
:
1
0
.
1
2
0
1
/
9
7
8
1
0
0
3
2
8
3
2
4
9
-
1
0
.
[
3
0
]
H
.
D
e
sh
p
a
n
d
e
a
n
d
L.
R
a
g
h
a
,
“
M
o
t
h
e
r
’
s si
g
n
i
f
i
c
a
n
t
f
e
a
t
u
r
e
(
M
S
F
)
d
a
t
a
se
t
,
”
I
EEE
D
a
t
a
P
o
rt
,
2
0
2
1
.
[
3
1
]
H
.
D
e
s
h
p
a
n
d
e
a
n
d
L
.
R
a
g
h
a
,
“
M
o
t
h
e
r
’
s
l
i
f
e
st
y
l
e
f
e
a
t
u
r
e
r
e
l
e
v
a
n
c
e
f
o
r
N
I
C
U
a
n
d
p
r
e
t
e
r
m
b
i
r
t
h
p
r
e
d
i
c
t
i
o
n
,
”
I
T
M
We
b
o
f
C
o
n
f
e
re
n
c
e
s
,
v
o
l
.
4
0
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
5
1
/
i
t
mc
o
n
f
/
2
0
2
1
4
0
0
3
0
3
9
.
[
3
2
]
L.
R
a
g
h
a
a
n
d
H
.
S
.
D
e
s
h
p
a
n
d
e
,
“
A
h
y
b
r
i
d
r
a
n
d
o
m
f
o
r
e
s
t
-
b
a
se
d
f
e
a
t
u
r
e
s
e
l
e
c
t
i
o
n
m
o
d
e
l
u
s
i
n
g
m
u
t
u
a
l
i
n
f
o
r
ma
t
i
o
n
a
n
d
F
-
sco
r
e
f
o
r
p
r
e
t
e
r
m
b
i
r
t
h
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Me
d
i
c
a
l
E
n
g
i
n
e
e
ri
n
g
a
n
d
I
n
f
o
rm
a
t
i
c
s
,
v
o
l
.
1
5
,
n
o
.
1
,
2
0
2
3
,
d
o
i
:
1
0
.
1
5
0
4
/
i
j
me
i
.
2
0
2
3
.
1
0
0
5
1
2
0
7
.
[
3
3
]
H
.
D
e
sh
p
a
n
d
e
a
n
d
L.
R
a
g
h
a
,
“
R
a
n
d
o
m
f
o
r
e
st
b
a
se
d
f
u
z
z
y
f
e
a
t
u
r
e
w
e
i
g
h
i
n
g
m
o
d
e
l
f
o
r
i
m
b
a
l
a
n
c
e
c
l
a
ss
d
i
st
r
i
b
u
t
i
o
n
t
o
w
a
r
d
s
p
r
e
t
e
r
m
-
b
i
r
t
h
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
S
S
RN
El
e
c
t
r
o
n
i
c
J
o
u
rn
a
l
,
2
0
2
1
,
d
o
i
:
1
0
.
2
1
3
9
/
ssr
n
.
3
8
6
7
3
5
4
.
[
3
4
]
L.
S
.
R
a
m
o
s
e
t
a
l
.
,
“
K
EEL
d
a
t
a
-
m
i
n
i
n
g
s
o
f
t
w
a
r
e
t
o
o
l
:
D
a
t
a
se
t
r
e
p
o
s
i
t
o
r
y
,
i
n
t
e
g
r
a
t
i
o
n
o
f
a
l
g
o
r
i
t
h
ms
a
n
d
e
x
p
e
r
i
men
t
a
l
a
n
a
l
y
si
s
f
r
a
mew
o
r
k
,
”
J
o
u
r
n
a
l
o
f
M
u
l
t
i
p
l
e
-
V
a
l
u
e
d
L
o
g
i
c
a
n
d
S
o
f
t
C
o
m
p
u
t
i
n
g
,
2
0
1
1
.
[
3
5
]
H
.
H
e
a
n
d
E
.
G
a
r
c
i
a
,
“
Le
a
r
n
i
n
g
f
r
o
m
i
mb
a
l
a
n
c
e
d
d
a
t
a
,
”
I
EE
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
K
n
o
w
l
e
d
g
e
a
n
d
D
a
t
a
E
n
g
i
n
e
e
r
i
n
g
,
2
0
0
9
.
[
3
6
]
B
.
M
o
h
a
me
d
,
D
.
H
.
K
h
e
l
i
o
u
a
n
e
,
a
n
d
T.
A
.
A
l
i
t
o
u
c
h
e
,
“
E
v
a
l
u
a
t
i
o
n
mea
s
u
r
e
s
f
o
r
mo
d
e
l
s
a
ssessm
e
n
t
o
v
e
r
i
mb
a
l
a
n
c
e
d
d
a
t
a
s
e
t
s,
”
J
o
u
rn
a
l
o
f
I
n
f
o
rm
a
t
i
o
n
E
n
g
i
n
e
e
ri
n
g
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
3
,
n
o
.
1
0
,
p
p
.
2
7
–
3
8
,
2
0
1
3
.
[
3
7
]
F
.
La
st
,
G
.
D
o
u
z
a
s,
a
n
d
F
.
B
a
c
a
o
,
“
O
v
e
r
sam
p
l
i
n
g
f
o
r
i
m
b
a
l
a
n
c
e
d
l
e
a
r
n
i
n
g
b
a
s
e
d
o
n
k
-
m
e
a
n
s
a
n
d
S
M
O
TE
,
”
a
rX
i
v
p
r
e
p
r
i
n
t
a
rXi
v
:
1
7
1
1
.
0
0
8
3
7
,
N
o
v
.
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
n
s.
2
0
1
8
.
0
6
.
0
5
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.