T
E
L
KO
M
N
I
KA
T
e
lec
om
m
u
n
icat
ion
,
Com
p
u
t
i
n
g,
E
lec
t
r
on
ics
an
d
Cont
r
ol
Vol.
18
,
No.
3
,
J
une
2020
,
pp.
1
433
~
14
38
I
S
S
N:
1693
-
6930,
a
c
c
r
e
dit
e
d
F
ir
s
t
G
r
a
de
by
Ke
me
nr
is
tekdikti
,
De
c
r
e
e
No:
21/E
/KP
T
/2018
DO
I
:
10.
12928/
T
E
L
KO
M
NI
KA
.
v18i3.
14837
1433
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
Prediction
schizophrenia
using
random
forest
Z
u
h
e
r
m
an
Rus
t
am
,
Gl
or
i
S
t
e
p
h
a
n
i
S
ar
agih
D
ep
ar
t
men
t
o
f
Mat
h
emat
i
cs
,
Facu
l
t
y
o
f
Mat
h
emat
i
cs
a
n
d
Sci
en
ce
,
U
n
i
v
ers
i
t
a
s
In
d
o
n
es
i
a
,
In
d
o
n
e
s
i
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Aug
15
,
2019
R
e
vis
e
d
F
e
b
12
,
2020
Ac
c
e
pted
F
e
b
21
,
2020
S
c
h
i
z
o
p
h
r
e
n
i
a
i
s
a
me
n
t
a
l
i
l
l
n
e
s
s
w
i
t
h
a
v
er
y
b
a
d
i
m
p
a
c
t
o
n
s
u
ff
er
er
s
,
a
t
t
ac
k
i
n
g
t
h
e
p
a
r
t
o
f
h
u
m
a
n
b
r
a
i
n
t
h
a
t
d
i
s
a
b
l
e
s
t
h
e
a
b
i
l
i
t
y
t
o
t
h
i
n
k
c
l
e
ar
l
y
.
I
n
2018
,
R
u
s
t
am
a
n
d
R
am
p
i
s
e
l
a
c
l
a
s
s
i
f
i
e
d
S
c
h
i
z
o
p
h
re
n
i
a
b
y
u
s
i
n
g
N
o
r
t
h
w
e
s
t
e
r
n
U
n
i
v
e
r
s
i
t
y
Sc
h
i
z
o
p
h
r
e
n
i
a
D
a
t
a,
b
as
e
d
o
n
6
6
v
a
r
i
a
b
l
e
s
c
o
n
s
i
s
t
i
n
g
o
f
g
r
o
u
p
,
d
e
m
o
g
ra
p
h
i
c,
a
n
d
q
u
e
s
t
i
o
n
n
a
i
re
s
s
t
a
t
i
s
t
i
c
s
,
b
a
s
e
d
o
n
t
h
e
s
c
a
l
e
f
o
r
t
h
e
a
s
s
e
s
s
me
n
t
o
f
n
e
g
a
t
i
v
e
s
y
m
p
t
o
m
s
(
S
A
N
S
),
a
n
d
s
c
a
l
e
fo
r
t
h
e
a
s
s
e
s
s
me
n
t
o
f
positive
sym
pto
m
s
(S
A
S),
a
n
d
t
h
e
n
c
l
a
s
s
i
f
i
er
s
t
h
a
t
u
s
e
d
ar
e
S
V
M
w
i
t
h
G
a
u
s
s
i
a
n
k
e
r
n
e
l
a
n
d
T
w
i
n
S
V
M
w
i
t
h
l
i
n
e
ar
a
n
d
G
a
u
s
s
i
an
k
e
r
n
e
l
.
F
u
r
t
h
er
m
o
r
e,
t
h
i
s
r
e
s
e
ar
c
h
i
s
n
o
v
e
l
b
a
s
e
d
o
n
t
h
e
u
s
e
o
f
ra
n
d
o
m
f
o
r
e
s
t
as
a
c
l
a
s
s
i
f
i
er,
i
n
o
r
d
er
t
o
p
r
e
d
i
c
t
S
c
h
i
z
o
p
h
re
n
i
a.
T
h
e
re
s
u
l
t
o
b
t
a
i
n
e
d
i
s
re
p
o
rt
e
d
i
n
p
e
rc
e
n
t
a
g
e
o
f
a
c
c
u
r
ac
y
,
b
o
t
h
i
n
t
r
a
i
n
i
n
g
a
n
d
t
e
s
t
i
n
g
o
f
ra
n
d
o
m
f
o
re
s
t
,
w
h
i
c
h
w
a
s
1
0
0
%
.
T
h
i
s
c
l
a
s
s
i
f
i
c
a
t
i
o
n
,
t
h
er
ef
o
re
,
s
h
o
w
s
t
h
e
b
e
s
t
v
a
l
u
e
i
n
c
o
n
t
ra
s
t
w
i
t
h
p
r
i
o
r
m
e
t
h
o
d
s
,
e
v
e
n
t
h
o
u
g
h
o
n
l
y
4
0
%
o
f
t
r
a
i
n
i
n
g
d
a
t
a
s
e
t
w
a
s
u
s
e
d
.
T
h
i
s
i
s
v
e
r
y
i
m
p
o
r
t
a
n
t
,
e
s
p
e
c
i
a
l
l
y
i
n
t
h
e
ca
s
e
s
o
f
r
a
re
d
i
s
ea
s
e,
i
n
c
l
u
d
i
n
g
s
c
h
i
z
o
p
h
re
n
i
a.
K
e
y
w
o
r
d
s
:
C
las
s
if
ica
ti
on
M
a
c
hine
lea
r
ning
R
a
ndom
f
or
e
s
t
S
c
hizophr
e
nia
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e
.
C
or
r
e
s
pon
din
g
A
u
th
or
:
Z
uhe
r
man
R
us
tam
,
De
pa
r
tm
e
nt
of
M
a
thema
ti
c
s
,
F
a
c
ult
y
of
M
a
thema
ti
c
s
a
nd
S
c
ienc
e
,
Unive
r
s
it
a
s
I
ndone
s
ia,
M
a
r
gonda
R
a
ya
S
t.
,
P
ondok
C
ina,
B
e
ji
,
De
pok,
J
a
wa
B
a
r
a
t
16424
,
I
ndone
s
ia
.
E
mail:
r
us
tam@ui.
a
c
.
id
1.
I
NT
RODU
C
T
I
ON
S
c
hizophr
e
nia
is
a
menta
l
il
lnes
s
that
ha
s
a
ve
r
y
ba
d
im
pa
c
t
on
s
uf
f
e
r
e
r
s
,
ba
s
e
d
on
it
s
a
bil
it
y
to
a
tt
a
c
k
pa
r
ts
o
f
the
human
b
r
a
in,
thus
dis
a
bli
ng
t
he
pe
r
s
ons
a
bil
it
y
to
thi
nk
c
lea
r
ly
[
1]
.
Ge
ne
r
a
ll
y,
pa
ti
e
nts
e
xpe
r
ienc
e
a
c
ha
nge
e
it
he
r
be
ha
vior
a
ll
y
or
on
t
h
e
mi
nd,
whic
h
s
ubs
e
que
ntl
y
a
f
f
e
c
ts
r
e
a
li
ty.
T
his
dis
e
a
s
e
pos
s
e
s
s
e
s
the
pr
ope
ns
it
y
to
a
tt
a
c
k
e
ve
r
yone
a
t
a
ny
a
ge
,
the
a
ve
r
a
ge
a
tt
a
c
ks
s
tar
ti
ng
a
t
the
a
ge
of
20s
in
men,
while
in
wome
n,
it
wa
s
obs
e
r
ve
d
a
t
the
e
nd
of
20s
[
1]
.
I
t
is
,
ther
e
f
or
e
,
im
por
tant
to
p
r
of
icie
ntl
y
identif
y
the
s
ympt
oms
on
ti
me,
in
or
de
r
to
pr
e
ve
nt
the
dis
e
a
s
e
f
r
om
oc
c
ur
ing
in
e
a
r
ne
s
t.
S
c
hizophr
e
nia
is
ge
ne
r
a
ll
y
divi
de
d
with
thr
e
e
s
ympt
oms
,
including
1
)
the
po
s
it
ive,
pr
e
s
e
nted
with
unne
c
e
s
s
it
a
ted
e
xtr
a
b
r
a
in
a
c
ti
vit
ies
,
including
ha
ll
uc
i
na
ti
ons
,
2)
the
ne
ga
ti
ve
,
indi
c
a
te
d
by
the
los
s
of
br
a
in
a
c
ti
vit
ies
,
3
)
the
c
ognit
ive,
whic
h
is
e
xhibi
ted
a
s
c
ha
ll
e
nge
s
with
a
bil
it
y
to
r
e
membe
r
a
nd
thi
nk
.
T
he
s
e
a
r
e
s
e
ve
r
e
f
a
c
ts
that
c
onf
e
r
potential
ne
ga
ti
ve
e
f
f
e
c
ts
on
the
li
f
e
o
f
s
uf
f
e
r
e
r
s
,
a
nd
thei
r
mo
r
talit
y
r
a
tes
a
r
e
2
to
2.
5
ti
mes
highe
r
than
the
ge
ne
r
a
l
population
[
2]
,
10
%
c
omm
it
e
d
s
uicide
a
nd
20
-
40
%
a
tt
e
mpt
e
d
s
uicide
a
t
lea
s
t
onc
e
[
3]
.
T
he
n,
the
c
a
us
e
of
s
c
hizophr
e
nia
ha
s
not
be
e
n
de
ter
mi
ne
d
f
o
r
s
ur
e
[
4]
a
nd
thi
s
il
lnes
s
ha
s
be
e
n
e
va
luate
d
to
ge
t
w
o
r
s
e
d
ue
to
the
incompe
tenc
e
of
e
a
r
ly
de
tec
ti
on,
he
nc
e
,
the
ne
e
d
t
o
identif
y
other
a
ppr
oa
c
he
s
f
or
de
tec
ti
on,
a
nd
one
of
whic
h
invol
ve
s
the
us
e
of
c
omput
a
ti
ona
l
methods
,
c
ompr
i
s
ing
of
mac
hine
lea
r
ning
.
How
e
ve
r
,
s
e
ve
r
a
l
pa
pe
r
s
ha
ve
a
tt
e
mpt
e
d
uti
li
z
ing
thi
s
tec
hnique
in
the
diagnos
is
of
s
c
hizophr
e
nia
,
including
S
VM
with
Ga
us
s
ian
ke
r
ne
l,
T
win
S
V
M
with
li
ne
a
r
a
nd
Ga
us
s
ian
ke
r
ne
l
[
5]
,
l
inea
r
dis
c
r
im
inant
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1
433
-
14
38
1434
a
na
lys
is
a
nd
k
-
Ne
a
r
e
s
t
n
e
ighbor
[
6]
,
f
is
he
r
li
ne
a
r
dis
c
r
im
inant
a
na
lys
is
[
7]
,
E
las
ti
c
Ne
t,
a
s
we
ll
a
s
lea
s
t
a
bs
olut
e
s
hr
inkage
a
nd
s
e
lec
ti
on
op
e
r
a
tor
[
8
]
.
T
hi
s
c
ur
r
e
nt
r
e
s
e
a
r
c
h
invol
ve
d
the
us
e
of
r
a
ndom
f
o
r
e
s
t
a
s
a
c
las
s
if
y,
a
lt
hough
it
ha
s
wide
ly
be
e
n
us
e
d
in
va
r
ious
s
tudi
e
s
,
including
the
pr
e
diction
of
ba
nk
f
inanc
ial
f
a
il
ur
e
s
,
wi
th
a
c
c
ur
a
c
y
of
93%
[
9
]
,
diabe
tes
melli
tus
a
t
80
.
8%
[
10
]
,
a
utom
a
ted
d
iagnos
is
of
he
a
r
t
d
is
e
a
s
e
a
t
83.
6%
,
a
pplyi
ng
the
we
ight
e
d
r
a
ndom
f
or
e
s
t
[
11
]
,
c
las
s
if
y
pr
os
tate
c
a
nc
e
r
[
12]
,
c
hr
onic
k
idney
dis
e
a
s
e
[
13]
a
nd
os
teoa
r
thr
it
is
dis
e
a
s
e
[
14]
.
He
nc
e
,
the
s
e
lec
te
d
a
ppr
oa
c
h
ha
s
pr
ove
n
the
c
a
pa
bil
it
y
of
thi
s
c
las
s
if
ier
to
a
pply
to
a
ny
pr
oblem
,
e
xhibi
ti
ng
good
mo
de
l
p
e
r
f
or
manc
e
in
the
p
r
oc
e
s
s
.
T
his
r
e
s
e
a
r
c
h
is
or
ga
nize
d
a
s
f
oll
ows
:
s
e
c
ti
on
1
p
r
ovides
ba
c
kgr
ound
of
de
tails
,
while
the
s
e
c
ond
s
pe
c
if
ies
da
ta
a
nd
r
e
s
e
a
r
c
h
meth
od
us
e
d.
I
n
a
ddit
ion,
r
e
s
ult
a
nd
a
na
lys
is
we
r
e
dis
c
us
s
e
d
in
s
e
c
ti
on
3,
a
nd
f
inally,
c
onc
lus
ions
a
r
e
include
d
in
s
e
c
ti
on
4.
2.
DA
T
A
AN
D
RE
S
E
AR
CH
M
E
T
HO
D
2.
1.
Dat
a
I
n
f
o
r
ma
t
io
n
f
r
om
th
e
da
ta
ba
s
e
of
No
r
th
we
s
te
r
n
U
n
ive
r
s
it
y
S
c
h
iz
op
h
r
e
n
ia
D
a
t
a
wa
s
us
e
d
in
th
is
s
tu
dy
[
1
]
.
F
u
r
t
he
r
mo
r
e
,
t
he
r
e
w
e
r
e
3
92
o
bs
e
r
v
a
t
io
ns
d
iv
id
e
d
i
n
to
4
g
r
ou
ps
,
wi
th
dis
t
r
i
bu
t
io
ns
a
s
s
e
e
n
in
T
a
bl
e
1
.
T
h
e
s
t
ud
y
f
ol
lo
we
d
t
he
g
r
ou
pi
ng
us
e
d
by
R
us
t
a
m
a
nd
R
a
m
pis
e
l
a
i
n
th
e
pa
pe
r
e
nt
it
le
d
“
S
u
ppo
r
t
v
e
c
t
o
r
m
a
c
hi
ne
s
a
n
d
t
wi
n
s
up
po
r
t
v
e
c
to
r
mac
hi
ne
s
i
n
t
he
c
l
a
s
s
if
ic
a
t
io
n
of
s
c
h
iz
op
h
r
e
nia
da
ta
”
[
5
]
.
T
his
wa
s
be
c
a
us
e
o
f
t
he
s
i
m
il
a
r
i
ty
in
r
e
s
e
a
r
c
h
f
oc
us
,
w
hi
c
h
wa
s
b
a
s
e
d
on
th
e
c
a
t
e
g
or
iz
a
t
io
n
in
to
s
c
h
iz
op
h
r
e
n
ics
a
nd
non
-
s
c
h
iz
op
hr
e
n
ics
o
nl
y
,
wh
ic
h
s
e
r
ve
d
a
s
t
he
g
r
o
up
va
r
i
a
b
le
[
2
]
.
No
n
-
s
c
hiz
op
h
r
e
ni
c
s
g
r
o
up
c
o
ns
is
ts
of
he
a
l
th
y
s
ib
li
ngs
of
the
pa
t
ie
n
ts
,
th
e
c
o
nt
r
ol
,
a
nd
s
i
bl
i
ngs
o
f
c
o
n
tr
o
l
.
A
to
ta
l
o
f
66
da
ta
va
r
ia
ble
s
we
r
e
c
o
l
lec
te
d
,
in
c
l
ud
i
ng
t
he
gr
ou
p
,
a
n
d
de
mo
g
r
a
ph
ics
,
c
ons
is
t
in
g
o
f
ge
nd
e
r
,
do
mi
na
nt
h
a
n
d
,
r
a
c
e
,
e
th
ni
c
i
ty
,
a
nd
a
ge
,
u
s
i
ng
q
ue
s
t
io
nna
i
r
e
s
s
ta
ti
s
t
ics
o
f
s
c
a
le
f
o
r
t
he
a
s
s
e
s
s
m
e
n
t
of
ne
ga
ti
ve
s
y
mp
to
ms
(
S
AN
S
)
[
1
5
]
a
nd
s
c
a
l
e
f
o
r
t
he
a
s
s
e
s
s
me
nt
o
f
ne
ga
ti
ve
s
ym
pt
o
ms
(
S
AP
S
)
[
1
6
]
,
a
s
s
ho
wn
in
T
a
b
le
2
.
T
a
ble
1.
Dis
tr
ibut
ion
of
g
r
oup
G
r
oup
N
umbe
r
of
obs
e
r
va
ti
ons
S
c
hi
z
ophr
e
ni
c
s
171
S
ib
li
ngs
of
S
c
hi
z
ophr
e
ni
c
s
44
C
ont
r
ol
S
ib
li
ngs
of
C
ont
r
ol
111
66
T
a
ble
2
.
T
he
va
r
iable
of
s
c
hizophr
e
nia
da
ta
th
V
a
r
ia
bl
e
D
a
ta
G
r
oup
V
a
r
ia
bl
e
D
e
s
c
r
ip
ti
on
1
-
34
Q
ue
s
ti
onna
ir
e
s
of
S
A
P
S
S
A
P
S
,
=
1
,
…
,
34
S
A
P
S
is
us
e
d
to
e
va
lu
a
te
th
e
pos
it
iv
e
s
ympt
oms
of
s
c
hi
z
ophr
e
ni
a
,
S
A
P
S
is
di
vi
de
d
in
to
4
ma
in
s
e
c
ti
ons
c
ont
a
in
in
g
34
di
f
f
e
r
e
nt
s
ympt
oms
,
na
me
ly
ha
ll
uc
in
a
ti
on,
de
lu
s
io
n,
bi
z
a
r
r
e
be
ha
vi
or
a
nd
th
ought
di
s
or
de
r
[
16]
.
T
h
e
da
ta
is
in
s
c
a
le
(
0
.
5
)
35
-
60
Q
ue
s
ti
onna
ir
e
s
of
S
A
N
S
S
A
N
S
,
=
35
,
…
,
60
S
A
N
S
is
us
e
d
to
e
va
lu
a
te
th
e
ne
ga
ti
ve
s
ympt
oms
of
s
c
hi
z
ophr
e
ni
a
,
S
A
N
S
is
di
vi
de
d
in
to
5
ma
in
s
e
c
ti
ons
c
ont
a
in
in
g
25
di
f
f
e
r
e
nt
s
ympt
oms
,
na
me
ly
e
mot
io
na
l
r
e
a
c
ti
on
de
c
li
ne
,
a
lo
gi
a
,
a
vol
it
io
n
a
nd
a
pa
th
y,
a
nhe
doni
a
a
nd
a
s
oc
ia
li
ty
,
a
nd
a
tt
e
nt
io
n [
15]
. T
he
da
ta
i
s
i
n s
c
a
le
(
0
.
5
)
61
D
e
mogr
a
phi
c
G
e
nde
r
G
e
nde
r
i
s
di
vi
de
d i
nt
o 2 c
a
te
gor
ie
s
:
1)
M
a
le
a
nd
2)
F
e
ma
le
62
D
e
mogr
a
phi
c
D
omi
na
nt
H
a
nd
D
omi
na
nt
H
a
nd i
s
di
vi
de
d i
nt
o 2 c
a
te
gor
ie
s
:
1)
L
e
f
t
a
nd
2)
r
ig
ht
63
D
e
mogr
a
phi
c
R
a
c
e
64
D
e
mogr
a
phi
c
E
th
ni
c
it
y
E
th
ni
c
it
y
is
di
vi
de
d
in
to
3
c
a
te
gor
ie
s
:
1)
K
a
uka
s
i
a
,
2)
A
f
r
ic
a
-
A
me
r
ic
a
,
a
nd
3)
ot
he
r
s
65
D
e
mogr
a
phi
c
A
ge
I
nt
e
ge
r
(
13
.
66
)
66
G
r
oup
G
r
oup
G
r
oup is
di
vi
de
d i
nt
o 2 c
la
s
s
e
s
:
0)
N
on
-
S
c
hi
z
ophr
e
ni
c
s
a
nd
1)
S
c
hi
z
ophr
e
ni
c
s
2
.
2.
Re
s
e
ar
c
h
m
e
t
h
od
2.
2.
1
.
B
oot
s
t
r
ap
I
n
1948
,
Que
ne
ivi
ll
e
int
r
oduc
e
d
J
a
c
kknif
e
a
s
a
r
e
s
a
mpl
ing
method,
while
B
r
a
dley
E
f
r
on
ini
ti
a
ted
boots
tr
a
p
a
s
it
s
r
e
volut
ion
in
1979,
s
e
r
ving
a
s
a
r
e
s
a
mpl
ing
method
with
r
e
plac
e
ment
[
17]
.
T
his
a
l
lows
f
or
the
c
r
e
a
ti
on
o
f
ne
w
da
ta
s
e
ts
f
r
om
the
or
i
ginal,
th
r
ough
r
e
pe
a
tedly
s
a
mpl
ing
the
obs
e
r
va
ti
ons
,
a
s
thi
s
is
mo
r
e
f
e
a
s
ibl
e
,
in
c
ontr
a
s
t
with
the
method
that
r
e
qui
r
e
d
obtaining
da
ta
f
r
om
the
population
f
r
e
que
nt
ly
[
18]
.
T
his
a
ppr
oa
c
h
is
a
tr
a
ini
ng
da
ta
s
e
t,
de
noted
a
s
,
whic
h
c
ontains
obs
e
r
va
ti
ons
,
r
a
ndoml
y
s
e
lec
t
e
d
to
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
r
e
diction
s
c
hiz
ophr
e
nia
us
ing
r
andom
for
e
s
t
(
Z
u
he
r
man
R
us
tam
)
1435
pr
oduc
e
the
ne
w
s
e
t,
∗
1
,
indi
c
a
ti
ng
a
s
im
il
a
r
it
y
i
s
s
ize
with
the
o
r
igi
na
l
.
I
n
a
ddit
ion
,
s
a
mpl
ing
wa
s
pe
r
f
or
med
with
r
e
plac
e
ment,
whic
h
s
igni
f
ies
the
pr
ope
ns
it
y
of
s
a
me
obs
e
r
va
ti
ons
a
ppe
a
r
ing
mor
e
than
onc
e
[
19]
.
T
he
p
r
oc
e
s
s
is
c
onduc
ted
c
onti
nuous
ly
to
the
point
whe
r
e
ne
w
da
ta
s
e
t,
is
c
a
pa
ble
of
s
e
tt
ing
pe
r
s
ona
ll
y
(
e
.
g
=
100
)
o
r
tune
by
a
n
indepe
nde
nt
va
li
da
ti
on
da
ta
s
e
t.
2.
2.
2
.
Rand
o
m
f
or
e
s
t
I
n
2001,
B
r
e
im
a
n
int
r
oduc
e
d
r
a
ndom
f
o
r
e
s
t
a
s
a
c
a
tegor
iza
ti
on
tool
,
c
ons
is
ti
ng
of
a
c
oll
e
c
ti
on
of
tr
e
e
-
s
tr
uc
tur
e
d
c
las
s
if
ier
s
{
ℎ
(
,
)
,
=
1
,
…
}
,
whe
r
e
{
}
a
r
e
identica
l
indepe
nde
nt
dis
tr
ibut
e
d
r
a
ndom
ve
c
tor
s
,
whe
r
e
e
a
c
h
tr
e
e
c
a
s
ts
a
unit
vote
f
o
r
the
mos
t
popular
c
las
s
a
t
input
[
20]
.
T
he
a
ppr
oa
c
h
us
e
d
is
a
im
e
d
a
t
im
p
r
oving
s
tabili
ty
a
nd
a
c
c
ur
a
c
y
of
the
de
c
is
ion
tr
e
e
,
thr
ough
the
c
r
e
a
ti
on
of
numer
ous
u
nit
s
f
r
om
e
xis
ti
ng
tr
a
ini
ng
da
ta,
us
ing
the
boots
tr
a
p
metho
d
[
21]
.
R
a
ndom
f
or
e
s
t
is
c
a
pa
ble
of
im
pr
oving
a
c
c
ur
a
c
y
thr
ough
r
a
ndomi
z
a
ti
on
a
nd
vot
ing
methods
,
a
nd
it
is
a
ls
o
a
ble
to
r
e
duc
e
the
c
or
r
e
lation
be
twe
e
n
tr
e
e
s
,
without
s
igni
f
ica
ntl
y
r
e
du
c
ing
the
s
tr
e
ngth
of
e
a
c
h
[
22]
.
T
he
r
e
f
or
e
,
whe
n
ove
r
f
i
tt
ing
is
obs
e
r
v
e
d
in
a
pa
r
ti
c
ular
t
r
a
ini
ng
da
ta
,
othe
r
s
do
not
be
ha
ve
in
the
s
a
me
manne
r
[
20]
.
B
a
s
e
d
on
B
r
e
i
man’
s
pa
pe
r
,
the
pr
o
c
e
s
s
buil
ding
of
numer
ous
tr
e
e
s
doe
s
not
c
r
e
a
te
a
n
ove
r
f
it
,
a
lt
hough
it
p
r
oduc
e
s
a
ge
ne
r
a
li
z
a
ti
on
e
r
r
or
that
c
onve
r
ge
s
to
a
va
lue
[
20]
.
Algor
it
hm
r
a
ndom
f
o
r
e
s
t
f
or
c
las
s
if
ica
ti
on
[
9]
.
1.
Give
n
the
tr
a
ini
ng
da
ta
s
e
t,
with
a
nd
a
s
obs
e
r
va
ti
ons
a
nd
va
r
iable
s
,
r
e
s
pe
c
ti
ve
ly
2.
F
or
=
1
to
a.
Dr
a
w
a
boots
tr
a
p
s
a
mpl
e
with
number
of
obs
e
r
va
ti
ons
f
r
om
the
tr
a
ini
ng
(
o
r
igi
na
l)
da
ta
s
e
t
b.
B
uil
d
the
de
c
is
ion
tr
e
e
f
r
om
e
a
c
h
ne
w
r
e
s
ult
de
r
ived,
whe
r
e
indi
vidual
node
s
a
r
e
c
hos
e
n
a
t
r
a
ndom.
i.
S
e
lec
t
va
r
iable
a
t
r
a
ndom
f
r
om
,
with
≤
,
whe
r
e
=
1
,
2
,
…
ℎ
√
[
23]
.
ii.
C
ho
os
e
the
be
s
t
f
e
a
tu
r
e
t
ha
t
p
r
o
v
ides
s
a
t
is
f
a
c
t
or
y
I
nf
o
r
ma
ti
on
Ga
in
or
Gi
n
i
I
nd
e
x
[
24
].
iii.
S
pli
t
the
node
c.
Gr
ow
e
a
c
h
without
p
r
uning
3.
Output
the
e
ns
e
mbl
e
of
de
c
is
ion
tr
e
e
s
{
}
1
4.
C
onduc
t
voti
ng,
i
.
e
.
,
if
̂
(
)
is
the
c
las
s
pr
e
diction
of
the
th
r
a
ndom
f
or
e
s
t
t
r
e
e
,
then
̂
(
)
=
{
̂
(
)
}
1
T
he
a
lgor
it
hm
a
bove
can
be
r
e
pr
e
s
e
ntative
with
F
i
gur
e
1.
F
igur
e
1
.
F
low
of
r
a
ndom
f
or
e
s
t
2.
2.
3
.
E
valu
at
io
n
of
m
od
e
l
p
e
r
f
or
m
an
ce
T
he
e
va
luation
of
model
pe
r
f
or
manc
e
is
im
por
t
a
nt,
due
to
it
s
a
bil
it
y
to
pr
ovides
knowle
dge
on
the
tool
s
’
e
f
f
icie
nc
y
of
c
las
s
if
ying
da
ta.
T
his
wa
s
a
s
s
e
s
s
e
d
thr
ough
the
mea
s
ur
e
ment
of
a
c
c
ur
a
c
y,
obtaine
d
f
r
om
the
r
e
s
ult
of
model
with
c
onf
us
ion
matr
ix,
whe
r
e
a
high
va
lue
ind
ica
tes
a
good
c
ondit
ion
of
the
c
las
s
if
ica
ti
on
model
[
21
]
.
B
a
s
ica
ll
y,
it
is
kno
wn
to
c
ontain
c
ompar
a
ble
in
f
or
mation
with
the
r
e
s
ult
made
by
the
model
,
including
the
bina
r
y
type
of
c
las
s
if
ica
ti
on,
whic
h
indi
c
a
tes
the
p
r
e
s
e
nc
e
of
2
ou
tp
ut
c
las
s
,
e
nc
ompas
s
ing
s
c
hi
z
ophr
e
nics
a
nd
non
-
s
c
hizophr
e
n
ics
.
How
e
ve
r
,
ther
e
a
r
e
4
pa
r
ts
to
thi
s
c
onf
us
ion
m
a
tr
ix:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1
433
-
14
38
1436
1)
T
P
:
T
r
ue
P
os
it
ive,
obs
e
r
ve
d
on
ins
tanc
e
s
whe
r
e
s
c
hizophr
e
nia
is
de
tec
ted
a
s
S
c
hizophr
e
nics
2)
T
N:
T
r
ue
Ne
ga
ti
ve
,
is
whe
n
non
-
s
c
hizophr
e
nia
is
de
tec
ted
a
s
non
-
s
c
hizophr
e
nics
3)
F
P
:
F
a
ls
e
P
os
it
ive,
is
s
e
e
n
in
c
a
s
e
s
whe
r
e
non
-
s
c
hi
z
ophr
e
nia
is
de
tec
ted
a
s
s
c
hizo
phr
e
nics
4)
F
N:
F
a
ls
e
Ne
ga
ti
ve
,
is
whe
n
s
c
hizophr
e
nia
is
de
tec
ted
a
s
non
-
s
c
hizophr
e
nic
s
T
a
ble
3
s
hows
the
c
onf
us
ion
mat
r
ix
us
e
d
in
thi
s
r
e
s
e
r
c
h.
T
a
ble
3
.
C
onf
us
ion
matr
ix
A
c
tu
a
l
C
la
s
s
S
c
hi
z
ophr
e
ni
c
s
N
on
-
S
c
hi
z
ophr
e
ni
c
s
P
r
e
di
c
te
d C
la
s
s
S
c
hi
z
ophr
e
ni
c
s
TP
FN
N
on
-
S
c
hi
z
ophr
e
ni
c
s
FP
TN
T
he
r
e
f
or
e
,
ba
s
e
d
on
the
va
lue
of
T
P
,
T
N
,
F
P
,
a
n
d
F
N
f
r
o
m
the
matr
ix,
i
t
is
pos
s
ibl
e
to
obtain
th
e
va
lued
a
c
c
ur
a
c
y
with
the
f
oll
owing
f
o
r
mul
a
:
A
ccu
r
a
cy
=
TP
+
TN
TP
+
TN
+
FP
+
FN
(
1)
3.
RE
S
UL
T
S
A
ND
AN
AL
YSI
S
3.
1
.
P
r
e
p
r
oc
e
s
s
in
g
d
at
a
T
he
r
e
66
va
r
iable
s
in
s
c
hizophr
e
nia
Da
ta,
ther
e
f
o
r
e
,
f
e
a
tur
e
s
e
lec
ti
on
wa
s
c
onduc
ted
be
f
or
e
f
it
ti
ng
the
model,
in
or
de
r
to
im
p
r
ove
a
c
c
ur
a
c
y.
T
his
wa
s
ba
s
e
d
on
the
pe
r
c
e
ntage
o
f
m
is
s
ing
da
ta
f
r
om
v
a
r
iable
s
les
s
than
10%
,
th
us
,
the
f
e
a
tur
e
is
s
e
lec
ted.
T
he
n,
I
f
the
indi
vidual
va
r
ianc
e
is
mor
e
than
that
in
the
da
ta
c
oll
e
c
ted
f
r
om
g
r
oup,
then
c
hoice
s
a
r
e
made
ba
s
e
d
on
thos
e
c
ondit
ions
,
with
60
us
e
d
in
the
model.
Anothe
r
mea
ns
of
pr
o
mot
ing
a
c
c
ur
a
c
y
is
by
tuni
ng
the
h
ype
r
pa
r
a
mete
r
s
,
whic
h
include
thos
e
that
a
f
f
e
c
t
t
he
model
s
tr
uc
tur
e
a
nd
a
ls
o
the
r
e
s
ult
output
,
ther
e
f
or
e
,
the
r
e
is
ne
e
d
to
identif
y
it
s
opti
mal
s
e
t
[
25]
.
T
his
is
ob
taine
d
by
lea
r
ning
va
r
ious
a
lgor
it
hms
with
dif
f
e
r
e
nt
s
e
ts
,
a
nd
s
ubs
e
que
ntl
y
c
ompar
ing
the
r
e
s
ult
s
of
e
a
c
hs
’
pe
r
f
or
manc
e
,
a
ls
o
known
a
s
tun
ing
the
model.
F
u
r
ther
mor
e
,
the
pa
pe
r
by
S
a
r
a
gih
a
nd
R
us
tam
wa
s
f
oll
owe
d
f
or
the
us
e
c
r
it
e
r
ion
(
Gin
i
a
nd
E
ntr
opy
)
[
9]
.
T
he
number
o
f
e
s
ti
mator
s
/t
r
e
e
s
s
e
r
ve
s
a
s
hype
r
pa
r
a
mete
r
s
to
im
pr
ove
the
a
c
c
ur
a
c
y,
whic
h
c
oll
e
c
ti
ve
ly
with
the
c
r
it
e
r
ion
is
a
f
unc
ti
on
that
mea
s
ur
e
s
the
e
qua
li
ty
of
a
s
pli
t,
e
ntr
opy
f
or
the
I
nf
or
mation
Ga
in
a
nd
Gini
index
[
9
]
.
3.
2
.
Re
s
u
lt
a
n
d
an
a
lys
is
A
s
tudy
[
5]
a
ppli
e
d
4
types
of
S
VM
on
t
he
s
a
m
e
s
c
hizophr
e
nia
da
ta,
the
r
e
f
or
e
,
the
main
goa
l
in
thi
s
r
e
s
e
a
r
c
h
is
nove
l,
thr
ough
the
us
e
of
r
a
ndo
m
f
or
e
s
t,
in
or
de
r
to
e
nha
nc
e
pr
e
dicta
bil
it
y.
T
his
r
e
s
e
a
r
c
h
r
e
quir
e
d
that
the
a
lgor
it
hm
wa
s
r
un
10
t
im
e
s
,
a
nd
t
he
r
e
pe
ti
ti
on
wa
s
pe
r
f
or
med
due
to
the
p
r
e
s
e
nc
e
of
e
leme
nt
r
a
ndom
in
thi
s
e
xpe
r
im
e
nt.
As
mention
in
s
e
c
t
ion
3.
1
,
the
model
tuni
ng
wa
s
c
onduc
ted
with
the
us
e
of
2
hype
r
pa
r
a
mete
r
c
ombi
na
ti
ons
,
a
nd
T
a
ble
4
pr
ovides
the
r
e
s
ult
of
c
las
s
if
ica
ti
on,
us
ing
r
a
ndom
f
o
r
e
s
t
with
e
ntr
opy,
while
gini
wa
s
the
c
r
it
e
r
ion
in
T
a
ble
5.
I
n
thi
s
r
e
s
e
a
r
c
h,
we
us
e
d
s
c
i
k
it
-
lear
n
li
br
a
r
y.
F
r
o
m
T
a
ble
4,
r
a
ndom
f
o
r
e
s
t
with
e
ntr
opy
in
a
c
c
ur
a
c
y
of
t
r
a
ini
ng
da
ta
wa
s
a
ble
to
c
or
r
e
c
tl
y
c
las
s
if
y
s
c
hizophr
e
nia
da
ta,
with
a
100
%
a
c
c
ur
a
c
y
leve
l
f
or
a
ll
c
ompos
it
ions
of
t
r
a
ini
ng
da
ta
s
e
t,
a
nd
number
o
f
t
r
e
e
s
.
T
his
oc
c
ur
s
on
ins
tanc
e
s
whe
r
e
the
number
of
tr
e
e
s
is
50
a
nd
100
f
or
50
-
80%
o
f
the
c
ompos
it
ion
tr
a
ini
ng
da
t
a
.
F
r
om
T
a
ble
5,
the
gin
i
in
a
c
c
ur
a
c
y
of
tr
a
ini
ng
da
ta
f
o
r
r
a
ndom
f
o
r
e
s
t,
pr
ov
ides
the
s
a
me
r
e
s
ult
a
s
e
ntr
op
y
in
the
c
las
s
if
ica
ti
on
of
s
c
hizophr
e
nia
da
ta,
whic
h
is
100%
f
or
a
ll
da
ta
s
e
t,
a
nd
number
of
tr
e
e
.
T
he
r
e
f
or
e
,
if
th
is
pe
r
c
e
ntage
wa
s
obtaina
ble
with
e
ntr
opy
f
or
tes
ti
n
g
in
50
-
80%
t
r
a
ini
ng
da
ta
s
e
t,
then
the
r
e
s
ult
of
gi
ni
in
40
-
80%
,
with
the
numbe
r
of
tr
e
e
s
is
100
,
is
c
o
r
r
e
c
t.
T
a
ble
4
.
Ac
c
ur
a
c
y
of
s
c
hizophr
e
nia
da
ta
c
las
s
if
ica
ti
on
us
ing
r
a
ndom
f
or
e
s
t
with
e
ntr
opy
a
s
c
r
it
e
r
ion
P
e
r
c
e
nt
a
ge
of
D
a
ta
T
r
a
in
in
g
(%)
N
umbe
r
of
T
r
e
e
10
50
100
10
50
100
A
c
c
ur
a
c
y of
T
e
s
ti
ng D
a
ta
(
%
)
A
c
c
ur
a
c
y of
T
r
a
in
in
g D
a
ta
(
%
)
10
94
95
98
100
100
100
20
95
96
98
100
100
100
30
97
98
99
100
100
100
40
96
99
99
100
100
100
50
97
100
100
100
100
100
60
99
100
100
100
100
100
70
100
100
100
100
100
100
80
100
100
100
100
100
100
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
r
e
diction
s
c
hiz
ophr
e
nia
us
ing
r
andom
for
e
s
t
(
Z
u
he
r
man
R
us
tam
)
1437
T
a
ble
5
.
Ac
c
ur
a
c
y
of
s
c
hizophr
e
nia
da
ta
c
las
s
if
ica
ti
on
us
ing
r
a
ndom
f
or
e
s
t
with
gini
a
s
c
r
it
e
r
ion
P
e
r
c
e
nt
a
ge
of
D
a
ta
T
r
a
in
in
g
(%)
N
umbe
r
of
T
r
e
e
10
50
100
10
50
100
A
c
c
ur
a
c
y of
T
e
s
ti
ng D
a
ta
(
%
)
A
c
c
ur
a
c
y of
T
r
a
in
in
g D
a
ta
(
%
)
10
93
96
98
100
100
100
20
95
95
98
100
100
100
30
96
98
98
100
100
100
40
96
99
100
100
100
100
50
96
99
100
100
100
100
60
97
100
100
100
100
100
70
100
100
100
100
100
100
80
100
100
100
100
100
100
B
a
s
e
d
on
T
a
ble
4
a
nd
T
a
ble
5
,
it
is
s
e
e
n
that
a
gr
e
a
ter
c
ompos
it
ion
us
e
d
is
a
ble
to
pr
oduc
e
higher
a
c
c
ur
a
c
y,
due
to
the
pr
e
s
e
nc
e
of
mor
e
da
ta
lea
r
ne
d
by
the
model.
He
nc
e
,
tuni
ng
indi
c
a
tes
the
be
s
t
a
c
c
ur
a
c
y
on
ins
tanc
e
s
whe
r
e
the
c
r
i
ter
ion
e
ntr
opy
is
with
number
o
f
e
s
ti
mator
s
a
s
100
.
How
e
ve
r
,
it
is
s
e
e
n
that
the
r
e
s
ult
s
a
r
e
s
im
il
a
r
f
or
e
a
c
h
c
ompos
it
ion
tr
a
ini
ng
da
ta
a
nd
number
of
e
s
ti
mator
s
.
As
s
a
id
in
S
e
c
ti
on
2,
s
c
hizophr
e
nia
da
ta
wa
s
f
oll
owe
d
that
s
a
me
wa
y
a
s
in
the
pa
pe
r
wr
it
ten
by
R
us
tam
a
nd
R
a
mpi
s
e
la
[
5
]
,
due
to
a
de
s
ir
e
to
c
ompar
e
pe
r
f
or
manc
e
r
e
s
ult
s
(
a
c
c
ur
a
c
y)
of
the
di
f
f
e
r
e
nt
methods
us
e
d.
T
a
ble
6,
ther
e
f
or
e
,
s
hows
the
c
ompar
is
on
of
e
a
c
h
methods
’
T
e
s
ti
ng
a
c
c
ur
a
c
y.
T
a
ble
6
s
hows
the
highes
t
a
c
c
ur
a
c
y
oc
c
ur
r
ing
a
t
r
a
ndom
f
or
e
s
t,
whic
h
wa
s
in
c
ontr
a
s
t
with
other
S
VM
met
hods
us
e
d
in
the
pa
pe
r
of
R
us
tam
a
nd
R
a
mpi
s
e
la
[
5]
.
T
his
wa
s
r
e
c
or
de
d
a
t
a
leve
l
of
100%
,
whe
n
the
pe
r
c
e
nt
a
ge
of
tr
a
ini
ng
da
ta
is
40
-
80%
,
whi
le
other
s
only
a
c
hieve
d
90%
a
t
60
-
80%
T
a
ble
6
.
P
e
r
f
o
r
manc
e
r
e
s
ult
s
T
r
a
in
in
g D
a
ta
(%)
L
in
e
a
r
S
V
M
(%)
G
a
us
s
ia
n S
V
M
(%)
L
in
e
a
r
T
w
in
S
V
M
(
%
)
G
a
us
s
ia
n T
w
in
S
V
M
(
%
)
R
a
ndom F
or
e
s
t
(
%
)
A
c
c
ur
a
c
y (
%
)
S
td
de
vi
a
ti
on
10
88
88
89
89
98
0.0001
20
89
89
89
89
98
0.0002
30
89
89
89
89
98
0.0001
40
89
89
89
89
100
0
50
89
89
89
89
100
0
60
90
90
90
90
100
0
70
90
90
90
90
100
0
80
90
90
90
90
100
0
4.
CONC
L
USI
ON
C
las
s
if
ica
ti
on
of
s
c
hizophr
e
nia
ha
s
pr
e
vious
ly
be
e
n
c
onduc
ted
by
R
us
tam
a
nd
R
a
mpi
s
e
la,
us
ing
S
VM
models
,
ther
e
f
or
e
,
thi
s
r
e
s
e
a
r
c
h
wa
s
nove
l
in
the
us
e
o
f
r
a
ndom
f
o
r
e
s
t
to
pr
e
dict
ba
s
e
d
on
the
inf
or
mation
c
oll
e
c
ted
f
r
om
the
da
taba
s
e
of
Nor
t
hwe
s
te
r
n
Unive
r
s
it
y
S
c
hizophr
e
nia
Da
ta,
whic
h
wa
s
a
ls
o
us
e
d
by
R
us
tam
a
nd
R
a
mpi
s
e
la.
T
he
r
e
f
or
e
,
it
wa
s
e
s
tablis
he
d
that
r
a
ndom
f
or
e
s
t
wi
th
e
nt
r
opy
a
nd
gi
ni
s
hows
s
im
il
a
r
r
e
s
ult
s
,
a
lt
hough
it
only
s
li
ghtl
y
pe
r
f
or
ms
be
tt
e
r
with
gini
a
nd
the
number
of
e
s
ti
mator
s
be
ing
100.
S
ubs
e
que
ntl
y,
thi
s
tec
hnique
wa
s
a
ls
o
a
ble
to
pr
e
dict
with
good
a
c
c
ur
a
c
y,
us
ing
40%
tr
a
ini
ng
da
ta,
whic
h
wa
s
in
c
ontr
a
s
t
with
other
methods
us
e
d
i
n
pr
ior
s
tudi
e
s
whic
h
ins
is
t
on
the
us
e
of
80%
,
in
or
de
r
to
obtain
90%
a
c
c
ur
a
c
y.
T
his
is
ve
r
y
im
po
r
t
a
nt,
e
s
pe
c
ially
in
the
p
r
e
diction
o
f
r
a
r
e
dis
e
a
s
e
,
whe
r
e
da
ta
is
dif
f
icult
to
obtain
.
I
n
c
ompar
is
on
with
pa
s
t
s
tudi
e
s
us
ing
the
s
a
me
da
ta,
r
a
ndom
f
or
e
s
t
wa
s
obs
e
r
ve
d
to
s
how
be
tt
e
r
a
c
c
ur
a
c
y,
a
t
100
%
f
or
tr
a
ini
ng
,
a
nd
a
ls
o
tes
ti
ng.
T
his
a
ppr
oa
c
h
is
,
ther
e
f
or
e
,
e
xpe
c
ted
to
be
r
e
leva
nt
in
the
medic
a
l
f
ield,
e
s
pe
c
ially
in
the
pr
e
diction
of
s
c
hi
z
ophr
e
nia
,
a
nd
s
ubs
e
que
ntl
y
in
other
dis
e
a
s
e
s
,
whic
h
is
c
ur
r
e
ntl
y
ha
r
d
in
diagnos
is
,
he
nc
e
,
e
nha
nc
ing
the
a
c
c
ur
a
c
y
of
the
medic
a
l
tea
m
in
pr
ovidi
ng
t
r
e
a
tm
e
nt.
I
t
is
s
ugge
s
ted
that
s
uc
c
e
s
s
ive
r
e
s
e
a
r
c
h
us
e
f
e
a
tur
e
s
e
lec
ti
on
to
identi
f
y
the
im
po
r
tant
f
e
a
tur
e
a
s
s
is
ts
the
medic
a
l
pr
a
c
ti
ti
one
r
s
to
f
oc
us
on
s
e
ve
r
a
l
da
ta
point
s
,
a
nd
a
l
s
o
that
the
a
ppli
c
a
ti
on
o
f
r
a
ndom
f
or
e
s
t
is
a
dopted
in
other
da
tas
e
t
with
upda
ted
a
nd
s
upe
r
ior
d
im
e
ns
ion.
AC
KNOWL
E
DGM
E
N
T
T
his
r
e
s
e
a
r
c
h
s
uppor
ted
f
inanc
ially
by
Unive
r
s
i
ty
of
indones
ia
with
a
DR
P
M
P
UT
I
Q2
2020
gr
a
nt
s
c
he
me.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1
433
-
14
38
1438
RE
F
E
RE
NC
E
S
[1
]
W
an
g
L
.
,
et
al
.
,
“
N
o
rt
h
w
es
t
er
n
U
n
i
v
er
s
i
t
y
Sc
h
i
z
o
p
h
re
n
i
a
D
a
t
a
an
d
So
f
t
w
are
T
o
o
l
(N
U
SD
A
ST
)
,”
F
r
o
n
t
i
er
s
i
n
Neu
r
o
i
n
f
o
r
m
a
t
i
cs
,
v
o
l
.
7
,
p
p
.
2
5
,
2
0
1
3
.
[2
]
Y
as
am
i
M
.
T
.
,
et
al
.
,
“
L
i
v
i
n
g
A
H
ea
l
t
h
y
L
i
fe
w
i
t
h
Sch
i
z
o
p
h
ren
i
a:
Pav
i
n
g
t
h
e
Ro
ad
t
o
Reco
v
ery
,
”
W
o
r
l
d
F
ed
e
r
a
t
i
o
n
f
o
r
M
en
t
a
l
H
e
a
t
l
h
:
W
o
r
l
d
H
ea
l
t
h
O
r
g
a
n
i
z
a
t
i
o
n
,
2
0
1
4
.
[3
]
Sch
i
z
o
p
h
ren
i
a.
co
m.
S
c
h
i
z
o
p
h
r
en
i
a
s
y
m
p
t
o
m
s
a
n
d
d
i
a
g
n
o
s
i
s
o
f
S
ch
i
z
o
p
h
r
e
n
i
a
ret
r
i
ev
e
d
f
ro
m
h
t
t
p
:
/
/
s
c
h
i
z
o
p
h
ren
i
a.
co
m/
s
zfact
s
.
h
t
m
o
n
2
2
O
c
t
o
b
er
2
0
1
9
.
[4
]
A
meri
ca
n
Ps
y
c
h
i
a
t
ri
c
A
s
s
o
c
i
at
i
o
n
.
“
D
i
a
g
n
o
s
t
i
c
a
n
d
s
t
a
t
i
s
t
i
ca
l
m
a
n
u
a
l
o
f
m
e
n
t
a
l
d
i
s
o
r
d
e
r
s
,
”
f
i
f
t
h
ed
i
t
i
o
n
(A
rl
i
n
g
t
o
n
,
V
A
:
A
mer
i
can
P
s
y
c
h
i
a
t
ri
c
A
s
s
o
ci
a
t
i
o
n
)
h
t
t
p
s
:
/
/
d
o
i
.
o
rg
/
1
0
.
1
1
7
6
/
a
p
p
i
.
b
o
o
k
s
.
9
7
8
0
8
9
0
4
2
5
5
9
6
.
2
0
1
3
.
[5
]
Ru
s
t
am
Z
.
,
Ramp
i
s
el
a
T
.
V
.
“
Su
p
p
o
r
t
v
ect
o
r
mach
i
n
es
an
d
t
w
i
n
s
u
p
p
o
rt
v
ec
t
o
r
mach
i
n
e
s
fo
r
cl
as
s
i
fi
ca
t
i
o
n
o
f
s
ch
i
zo
p
h
re
n
i
a
d
a
t
a,
”
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
E
n
g
i
n
eer
i
n
g
&
Tech
n
o
l
o
g
y,
v
o
l
.
7
,
n
o
.
4
,
p
p
.
6
3
7
8
-
6
8
7
7
,
2
0
1
8
.
[6
]
A
h
n
M
.
,
H
o
n
g
J
.
H
.
,
an
d
J
u
n
S
.
C.
,
“
Feas
i
b
i
l
i
t
y
o
f
ap
p
r
o
ach
e
s
co
mb
i
n
i
n
g
s
en
s
o
r
an
d
s
o
u
rce
feat
u
re
s
i
n
b
rai
n
-
co
m
p
u
t
er
i
n
t
erface.
”
Jo
u
r
n
a
l
o
f
n
eu
r
o
s
ci
e
n
ce
m
et
h
o
d
s
,
v
o
l
.
2
0
4
,
n
o
.
1
,
p
p
.
1
6
8
–
1
78
,
2
0
1
2
.
[7
]
N
eu
h
au
s
A
.
H
.
,
et
al
.
,
“
Si
n
g
l
e
-
s
u
b
j
ec
t
cl
as
s
i
fi
ca
t
i
o
n
o
f
s
ch
i
z
o
p
h
ren
i
a
u
s
i
n
g
ev
e
n
t
-
re
l
at
e
d
p
o
t
en
t
i
a
l
s
o
b
t
a
i
n
e
d
d
u
r
i
n
g
au
d
i
t
o
ry
an
d
v
i
s
u
al
o
d
d
b
a
l
l
p
arad
i
g
m
s
,
”
E
u
r
o
p
ea
n
A
r
ch
i
ve
s
o
f
P
s
yc
h
i
a
t
r
y
a
n
d
C
l
i
n
i
c
a
l
Neu
r
o
s
ci
e
n
ce
,
v
o
l
.
263
,
no.
3
,
p
p
.
2
4
1
-
2
4
7
,
2
0
1
3
.
[8
]
H
et
t
i
g
e
N
.
C
.
,
et
al
.
,
“
Cl
as
s
i
f
i
cat
i
o
n
o
f
s
u
i
ci
d
e
at
t
emp
t
er
s
i
n
s
c
h
i
z
o
p
h
ren
i
a
u
s
i
n
g
s
o
c
i
o
c
u
l
t
u
ra
l
an
d
cl
i
n
i
cal
feat
u
r
es
:
A
mach
i
n
e
l
ear
n
i
n
g
ap
p
ro
ac
h
,”
G
e
n
H
o
s
p
P
s
yc
h
i
a
t
r
y
,
v
o
l
.
4
7
,
p
p
.
20
-
2
8
,
2
0
1
7
.
[9
]
Ru
s
t
am
Z
.
,
an
d
Sarag
i
h
G
.
,
“
Pr
ed
i
ct
b
an
k
fi
n
an
c
i
al
fai
l
u
res
u
s
i
n
g
ran
d
o
m
fo
re
s
t
,
”
In
s
t
i
t
u
t
e
o
f
E
l
ec
t
r
i
ca
l
a
n
d
E
l
ec
t
r
o
n
i
cs
E
n
g
i
n
ee
r
s
.
d
o
i
:
1
0
.
1
1
0
9
/
IW
BIS.
2
0
1
8
.
8
4
7
1
7
1
8
.
p
p
.
8
1
-
86
,
2
0
1
8
.
[1
0
]
Z
o
u
Q
.
,
et
al
.
,
“
Pred
i
ct
i
n
g
D
i
a
b
et
e
s
Mel
l
i
t
u
s
w
i
t
h
Mach
i
n
e
L
earn
i
n
g
T
ech
n
i
q
u
e
s
,
”
F
r
o
n
t
i
er
s
i
n
G
e
n
et
i
cs
,
v
o
l
.
9
,
n
o
.
5
1
5
,
2
0
1
8
.
[1
1
]
Pat
i
l
P
.
R
.
,
an
d
K
i
n
ar
i
w
a
l
a
S
.
A
.
,
“
A
u
t
o
mat
ed
D
i
a
g
n
o
s
i
s
o
g
H
eart
D
i
s
ea
s
e
u
s
i
n
g
Ran
d
o
m
F
o
res
t
A
l
g
o
ri
t
h
m
,
”
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
A
d
va
n
ce
R
e
s
ea
r
ch
,
Id
e
a
s
a
n
d
I
n
n
o
va
t
i
o
n
s
i
n
Tec
h
n
o
l
o
g
y.
v
o
l
.
3
,
n
o
.
2
,
p
p
.
5
7
9
–
5
8
9
,
2
0
1
7
.
[1
2
]
Hul
j
a
n
ah
M
.
,
et
al
.
,
“
F
ea
t
u
r
e
S
el
e
ct
i
o
n
u
s
i
n
g
R
a
n
d
o
m
F
o
r
es
t
Cl
a
s
s
i
f
i
er
f
o
r
P
r
ed
i
ct
i
n
g
P
r
o
s
t
a
t
e
Ca
n
ce
r
,”
IO
P
Co
n
fere
n
ce
Seri
es
:
Mat
er
i
al
s
Sci
en
ce
an
d
E
n
g
i
n
eeri
n
g
,
v
o
l
.
5
4
6
,
n
o
.
5
,
2
0
1
9
.
[1
3
]
Ru
s
t
am
Z
.
,
Su
d
ars
o
n
o
E
.
,
an
d
Sar
w
i
n
d
a
D
.
,
“
R
a
n
d
o
m
-
F
o
r
e
s
t
(R
F
)
a
n
d
S
u
p
p
o
r
t
V
ec
t
o
r
M
a
ch
i
n
e
(S
V
M
)
Im
p
l
em
e
n
t
a
t
i
o
n
f
o
r
A
n
a
l
ys
i
s
o
f
G
en
e
E
xp
r
es
s
i
o
n
D
a
t
a
i
n
Ch
r
o
n
i
c
Ki
d
n
ey
D
i
s
ea
s
e
(CKD
),
”
IO
P
C
o
n
fere
n
ce
Ser
i
es
:
Ma
t
eri
a
l
s
Sci
e
n
ce
an
d
E
n
g
i
n
eeri
n
g
,
v
o
l
.
5
4
6
,
n
o
.
5
,
2
0
1
9
.
[1
4
]
A
p
r
i
l
l
i
a
n
i
U
.
,
an
d
Ru
s
t
am
Z
.
,
“
O
s
t
e
o
a
rt
h
ri
t
i
s
D
i
s
ea
s
e
P
r
ed
i
c
t
i
o
n
Ba
s
ed
o
n
Ran
d
o
m
F
o
res
t
,
”
In
s
t
i
t
u
t
e
o
f
E
l
ect
r
i
ca
l
a
n
d
E
l
ec
t
r
o
n
i
cs
E
n
g
i
n
eer
s
.
D
O
I
:
1
0
.
1
1
0
9
/
ICA
CSIS.
2
0
1
8
.
8
6
1
8
1
6
6
.
p
p
.
2
3
7
-
2
4
0
,
2
0
1
8
.
[1
5
]
A
n
d
reas
e
n
N
.
C.
,
“
S
ca
l
e
f
o
r
t
h
e
A
s
s
es
s
m
e
n
t
o
f
N
eg
a
t
i
ve
S
y
m
p
t
o
m
s
(SA
N
S).
”
co
n
cep
t
u
a
l
a
n
d
t
h
eo
r
et
i
ca
l
f
o
u
n
d
a
t
i
o
n
s
.
Th
e
B
r
i
t
i
s
h
j
o
u
r
n
a
l
o
f
p
s
yc
h
i
a
t
r
y
,
p
p
.
4
9
-
5
2
,
1
9
8
9
.
[1
6
]
A
n
d
reas
e
n
N
.
C.
,
“
S
ca
l
e
f
o
r
t
h
e
A
s
s
e
s
s
m
en
t
o
f
P
o
s
i
t
i
ve
S
ym
p
t
o
m
s
”
(SA
p
S.
Io
w
a
Ci
t
y
:
U
n
i
v
ers
i
t
y
o
f
Io
w
a).
1
9
8
4
.
[1
7
]
E
fro
n
B
.
,
an
d
T
i
b
s
h
i
ra
n
i
R.
,
“
A
n
I
n
t
r
o
d
u
ct
i
o
n
t
o
t
h
e
Bo
o
t
s
t
r
a
,”
N
ew
Y
o
rk
:
Ch
a
p
man
&
H
a
l
l
.
1
9
9
3
.
[1
8
]
H
as
t
i
e
T
.
,
T
i
b
s
h
i
ra
n
i
R
.
,
an
d
Fri
e
d
man
J
.
,
“
T
h
e
el
e
men
t
o
f
s
t
at
i
s
t
i
ca
l
l
earn
i
n
g
d
a
t
a
mi
n
i
n
g
,
i
n
fere
n
ce,
an
d
p
red
i
ct
i
o
n
,
”
Cal
i
f
o
rn
i
a:
Sp
r
i
n
g
er.
2
0
0
9
.
[1
9
]
J
ames
G
.
,
et
al
.
,
“
Fi
rs
t
E
d
i
t
i
o
n
;
A
n
In
t
ro
d
u
c
t
i
o
n
t
o
St
at
i
s
t
i
cal
L
earn
i
n
g
w
i
t
h
A
p
p
l
i
cat
i
o
n
i
n
R
,
”
S
p
r
i
n
g
e
r
.
N
ew
Y
o
r
k
:
S
p
r
i
n
g
er
-
V
er
l
ag
N
ew
Y
o
rk
,
v
o
l
.
1
1
2
,
p
p
.
3
-
7
,
2
0
1
3
.
[2
0
]
Brei
man
L
.
,
“
Ran
d
o
m
fo
re
s
t
s
,
”
M
a
c
h
.
Lea
r
n
,
v
o
l
.
4
5
,
n
o
.
1
,
p
p
.
5
–
3
2
,
2
0
0
1
.
[2
1
]
Mi
n
d
erman
n
S.
,
“
Ran
d
o
m
fo
re
s
t
s
,”
T
h
es
i
s
.
A
m
s
t
er
d
am
:
K
o
rt
e
w
eg
-
d
e
V
r
i
es
I
n
s
t
i
t
u
u
t
v
o
o
r
W
i
s
k
u
n
d
e,
U
n
i
v
er
s
i
t
ei
t
v
an
A
m
s
t
er
d
am,
2
0
1
6
.
[2
2
]
St
ro
b
l
C.
,
“
St
at
i
s
t
i
ca
l
Is
s
u
e
s
i
n
Mac
h
i
n
e
L
earn
i
n
g
-
T
o
w
ard
s
Rel
i
a
b
l
e
S
p
l
i
t
Se
l
ect
i
o
n
an
d
V
ari
a
b
l
e
Im
p
o
r
t
a
n
ce
Meaas
u
re
s
,
”
Mu
n
c
h
en
,
2
0
0
8
.
[2
3
]
Brei
man
L
.
,
et
al
.
,
“
Cl
as
s
i
fi
ca
t
i
o
n
an
d
reg
re
s
s
i
o
n
t
ree
s
”
W
a
d
s
wo
r
t
h
In
t
,
1
9
8
4
.
[2
4
]
G
rab
czew
s
k
i
K
.
,
“
Met
a
-
L
earn
i
n
g
i
n
d
eci
s
i
o
n
t
ree
i
n
d
u
c
t
i
o
n
,”
To
r
u
n
:
S
p
r
i
n
g
er
,
v
o
l
.
1
,
2
0
1
4
.
[2
5
]
K
o
h
av
i
R
.
an
d
Pro
v
o
s
t
F.
,
“
O
n
ap
p
l
i
e
d
res
earch
i
n
mach
i
n
e
l
ear
n
i
n
g
,
”
E
d
i
t
o
r
i
a
l
f
o
r
t
h
e
S
p
e
ci
a
l
Is
s
u
e
o
n
A
p
p
l
i
ca
t
i
o
n
M
a
c
h
i
n
e
Lea
r
n
i
n
g
a
n
d
t
h
e
K
n
o
w
l
ed
g
e
D
i
s
co
ve
r
y
P
r
o
ce
s
s
,
Co
l
u
m
b
i
a
U
n
i
ver
s
i
t
y,
v
o
l
.
30
,
p
p
.
1
2
7
-
1
3
2
,
1
9
9
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.