I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
E
lec
t
r
ical
an
d
Com
p
u
t
e
r
E
n
gin
e
e
r
in
g
(
I
JE
CE
)
Vol.
1
4
,
No.
5
,
Oc
tober
20
2
4
,
pp
.
5534
~
554
2
I
S
S
N:
2088
-
8708
,
DO
I
:
10
.
11591/
ij
e
c
e
.
v
1
4
i
5
.
pp
5
534
-
554
2
5534
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
e
c
e
.
iaes
c
or
e
.
c
om
For
e
c
ast
in
g
c
r
e
d
itw
o
r
t
h
in
e
ss i
n
c
r
e
d
it s
c
or
i
n
g u
si
n
g
m
ac
h
in
e
le
ar
n
i
n
g m
e
t
h
o
d
s
Ayagoz
M
u
k
h
an
ova
1
,
M
ad
iyar
B
ait
e
m
irov
1
,
Az
am
at
Am
irov
2
,
B
olat
T
as
s
u
ov
3
,
Vale
n
t
in
a
M
ak
h
at
ova
4
,
As
s
e
m
gu
l
Kai
p
ova
5
,
Ulz
h
an
M
ak
h
az
h
an
ova
1
,
T
leugai
s
h
a
Os
p
an
ova
1
1
D
e
pa
r
tm
e
nt
of
I
nf
or
ma
ti
on S
ys
te
ms
, L
. N
. G
umi
ly
ov E
ur
a
s
ia
n
N
a
ti
ona
l
U
ni
ve
r
s
it
y, A
s
ta
na
,
R
e
publ
ic
of
K
a
z
a
kh
s
ta
n
2
D
ig
it
a
li
z
a
ti
on D
e
pa
r
tm
e
nt
,
A
byl
ka
s
S
a
gi
nov K
a
r
a
ga
nda
T
e
c
hn
ic
a
l
U
ni
ve
r
s
it
y, K
a
r
a
ga
nda
,
R
e
publ
ic
of
K
a
z
a
kh
s
ta
n
3
F
a
c
ul
t
y of
N
a
tu
r
a
l
S
c
ie
nc
e
s
, N
o
n
-
pr
of
it
L
im
it
e
d
L
ia
bi
li
ty
C
o
mp
a
ny,
M
.
H
.
D
ul
a
ty
T
a
r
a
z
S
ta
t
e
U
ni
v
e
r
s
i
ty
,
T
a
r
a
z
,
R
e
p
ubl
i
c
of
K
a
z
a
kh
s
t
a
n
4
D
e
pa
r
tm
e
nt
of
S
of
twa
r
e
E
ngi
ne
e
r
in
g, A
ty
r
a
u S
ta
te
U
ni
ve
r
s
it
y K
h. D
os
mukha
me
dova
, A
ty
r
a
u, R
e
publ
ic
of
K
a
z
a
kh
s
ta
n
5
D
e
pa
r
tm
e
nt
of
B
io
s
ta
ti
s
ti
c
s
, B
io
in
f
or
ma
ti
c
s
a
nd I
nf
or
ma
ti
on T
e
c
hnol
ogi
e
s
, A
s
t
a
na
M
e
di
c
a
l
U
ni
ve
r
s
it
y, A
s
t
a
na
, R
e
publ
ic
of
K
a
z
a
khs
ta
n
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
F
e
b
27,
2024
R
e
vis
e
d
J
un
20,
2024
Ac
c
e
pted
J
ul
1,
2024
T
h
i
s
ar
t
i
c
l
e
p
r
o
v
i
d
e
s
an
o
v
er
v
i
e
w
o
f
m
o
d
er
n
mac
h
i
n
e
l
earn
i
n
g
met
h
o
d
s
i
n
t
h
e
co
n
t
ex
t
o
f
t
h
ei
r
act
i
v
e
u
s
e
i
n
cred
i
t
s
co
r
i
n
g
,
w
i
t
h
p
art
i
c
u
l
ar
at
t
e
n
t
i
o
n
t
o
t
h
e
fo
l
l
o
w
i
n
g
al
g
o
r
i
t
h
ms
:
l
i
g
h
t
g
ra
d
i
e
n
t
b
o
o
s
t
i
n
g
mach
i
n
e
(
L
G
BM
)
c
l
as
s
i
f
i
er,
l
o
g
i
s
t
i
c
re
g
re
s
s
i
o
n
(L
R)
,
l
i
n
ear
d
i
s
cr
i
mi
n
a
n
t
an
al
y
s
i
s
(L
D
A
)
,
d
eci
s
i
o
n
t
ree
(
D
T
)
c
l
as
s
i
f
i
er,
g
rad
i
en
t
b
o
o
s
t
i
n
g
cl
a
s
s
i
f
i
er
an
d
ex
t
reme
g
rad
i
en
t
b
o
o
s
t
i
n
g
(
X
G
B
)
c
l
a
s
s
i
fi
er.
E
ach
o
f
t
h
e
me
t
h
o
d
s
men
t
i
o
n
e
d
i
s
s
u
b
j
ec
t
t
o
carefu
l
an
a
l
y
s
i
s
t
o
ev
a
l
u
a
t
e
t
h
ei
r
a
p
p
l
i
cab
i
l
i
t
y
an
d
effec
t
i
v
en
e
s
s
i
n
p
red
i
ct
i
n
g
cred
i
t
ri
s
k
.
T
h
e
ar
t
i
c
l
e
ex
ami
n
es
t
h
e
ad
v
an
t
ag
es
an
d
l
i
mi
t
at
i
o
n
s
o
f
each
met
h
o
d
,
i
d
e
n
t
i
fy
i
n
g
t
h
ei
r
i
mp
ac
t
o
n
t
h
e
acc
u
rac
y
an
d
re
l
i
a
b
i
l
i
t
y
o
f
b
o
rr
o
w
er
cred
i
t
w
o
r
t
h
i
n
e
s
s
as
s
es
s
men
t
s
.
Cu
rren
t
t
ren
d
s
i
n
mach
i
n
e
l
ear
n
i
n
g
an
d
cred
i
t
s
co
r
i
n
g
are
al
s
o
c
o
v
ere
d
,
w
ar
n
i
n
g
o
f
c
h
al
l
e
n
g
e
s
an
d
d
i
s
cu
s
s
i
n
g
p
ro
s
p
ec
t
s
.
T
h
e
an
a
l
y
s
i
s
h
i
g
h
l
i
g
h
t
s
t
h
e
s
i
g
n
i
f
i
can
t
co
n
t
r
i
b
u
t
i
o
n
s
o
f
met
h
o
d
s
s
u
c
h
as
L
G
BM
cl
as
s
i
f
i
er,
LR
,
L
D
A
,
DT
cl
as
s
i
f
i
er
,
g
rad
i
en
t
b
o
o
s
t
i
n
g
cl
as
s
i
f
i
er
an
d
X
G
B
c
l
as
s
i
fi
er
t
o
t
h
e
d
e
v
el
o
p
me
n
t
o
f
m
o
d
er
n
cred
i
t
s
c
o
ri
n
g
p
ract
i
ces
,
h
i
g
h
l
i
g
h
t
i
n
g
t
h
e
i
r
p
o
t
e
n
t
i
a
l
fo
r
i
mp
r
o
v
i
n
g
t
h
e
ac
cu
rac
y
an
d
rel
i
a
b
i
l
i
t
y
o
f
b
o
rr
o
w
er
cred
i
t
w
o
r
t
h
i
n
e
s
s
f
o
recas
t
s
i
n
t
h
e
fi
n
a
n
ci
a
l
s
er
v
i
ce
s
i
n
d
u
s
t
ry
.
A
d
d
i
t
i
o
n
al
l
y
,
t
h
e
art
i
cl
e
d
i
s
c
u
s
s
es
t
h
e
i
m
p
o
r
t
an
ce
o
f
carefu
l
s
el
ec
t
i
o
n
o
f
mach
i
n
e
l
earn
i
n
g
mo
d
el
s
an
d
t
h
e
n
eed
t
o
co
n
t
i
n
u
al
l
y
u
p
d
at
e
met
h
o
d
o
l
o
g
y
i
n
l
i
g
h
t
o
f
t
h
e
rap
i
d
l
y
ch
a
n
g
i
n
g
n
at
u
re
o
f
t
h
e
fi
n
a
n
ci
a
l
mark
e
t
.
K
e
y
w
o
r
d
s
:
C
r
e
dit
wor
thi
ne
s
s
De
c
is
ion
tr
e
e
c
las
s
if
ier
Gr
a
dient
boos
ti
ng
c
las
s
if
ier
L
inea
r
dis
c
r
im
inant
a
na
lys
is
L
ogis
ti
c
r
e
gr
e
s
s
ion
M
a
c
hine
lea
r
ning
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
:
M
a
diyar
B
a
it
e
mi
r
ov
De
pa
r
tm
e
nt
of
I
nf
or
mat
ion
S
ys
tems
,
L
.
N.
Gumil
y
ov
E
ur
a
s
ian
Na
ti
ona
l
Unive
r
s
it
y
010000
As
tana
,
R
e
publi
c
of
Ka
z
a
khs
tan
E
mail:
madiya
r
.
ba
ytemir
ov@inbox
.
r
u
1.
I
NT
RODU
C
T
I
ON
I
n
the
moder
n
wor
ld
o
f
f
inanc
e
,
whe
r
e
c
ompeti
ti
on
in
the
lending
mar
ke
t
[
1]
–
[
3]
is
c
ons
tantly
gr
owing,
the
r
e
leva
nc
e
of
de
ve
lopi
ng
e
f
f
e
c
ti
ve
methods
f
or
p
r
e
dicting
c
r
e
dit
wo
r
thi
ne
s
s
is
high.
T
he
a
c
c
ur
a
c
y
a
nd
r
e
li
a
bil
it
y
of
s
uc
h
methods
a
r
e
ke
y
f
a
c
tor
s
f
o
r
f
inanc
ial
ins
ti
tut
ions
s
e
e
king
to
mi
nim
ize
r
is
k
a
n
d
e
ns
ur
e
the
s
us
taina
bil
it
y
o
f
their
loan
por
tf
o
li
os
.
L
ight
g
r
a
dient
boos
ti
ng
mac
hine
(
L
GB
M
)
c
las
s
if
ier
[
4
]
–
[
6]
,
logi
s
ti
c
r
e
gr
e
s
s
ion
[
7]
–
[
9
]
,
li
ne
a
r
dis
c
r
im
inant
a
na
lys
is
[
10
]
–
[
13]
,
de
c
is
ion
tr
e
e
c
las
s
if
ier
[
13]
,
[
14]
,
gr
a
dient
boos
ti
ng
c
las
s
if
ier
[
15]
–
[
17
]
a
nd
e
xt
r
e
me
gr
a
dient
boos
ti
ng
(
XG
B
)
c
las
s
if
ier
[
18]
,
[
19]
a
r
e
a
va
r
iety
o
f
mac
hine
tr
a
ini
ng,
e
a
c
h
with
it
s
own
unique
c
ha
r
a
c
ter
is
ti
c
s
a
nd
a
ppli
c
a
ti
ons
.
T
he
i
r
us
e
in
c
r
e
dit
s
c
or
ing
[
20]
–
[
22]
ope
ns
up
ne
w
oppor
tuni
ti
e
s
f
or
im
p
r
oving
the
a
c
c
ur
a
c
y
of
f
or
e
c
a
s
ts
,
e
s
pe
c
ially
whe
n
wor
king
w
it
h
lar
ge
volum
e
s
of
da
ta
a
nd
c
ompl
e
x
c
r
e
dit
models
.
I
n
thi
s
s
tudy,
we
will
a
ls
o
a
ddr
e
s
s
is
s
ue
s
of
model
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
F
or
e
c
as
ti
ng
c
r
e
dit
w
or
thi
ne
s
s
in
c
r
e
dit
s
c
or
ing
us
i
ng
mac
hine
lear
nin
g
me
thods
(
A
y
agoz
M
uk
hanov
a
)
5535
int
e
r
pr
e
tabili
ty,
c
omput
a
ti
ona
l
c
ompl
e
xit
y
,
a
nd
po
tential
li
mi
tations
.
T
he
s
e
a
s
pe
c
ts
play
a
n
im
por
tan
t
r
ole
in
the
im
pleme
ntation
o
f
r
e
s
e
a
r
c
h
r
e
s
ult
s
in
r
e
a
l
f
inanc
ial
pr
a
c
ti
c
e
s
.
R
e
s
e
a
r
c
h
int
o
c
r
e
dit
wor
thi
ne
s
s
us
ing
mac
hine
lea
r
ning
methods
is
not
only
of
a
c
a
de
mi
c
int
e
r
e
s
t,
but
a
ls
o
ha
s
d
ir
e
c
t
a
ppli
c
a
ti
on
va
lue
f
or
f
inanc
ial
ins
ti
tut
ions
,
ins
ur
a
nc
e
c
ompanie
s
a
nd
other
mar
ke
t
playe
r
s
.
I
t
is
e
xpe
c
ted
that
the
r
e
s
ult
s
of
thi
s
s
tudy
c
a
n
s
e
r
ve
a
s
a
ba
s
is
f
or
opti
mi
z
ing
de
c
is
ion
-
making
pr
oc
e
s
s
e
s
in
the
lending
indus
tr
y
a
nd
p
r
ovide
mor
e
e
f
f
e
c
ti
ve
r
is
k
mana
ge
ment.
Koc
e
t
al
.
[
23]
e
xplo
r
e
the
r
ole
o
f
c
r
e
dit
r
a
ti
ngs
i
n
a
s
s
e
s
s
ing
f
inanc
ial
s
tabili
ty
a
nd
the
c
r
it
e
r
ia
f
o
r
is
s
uing
a
loan.
T
he
y
r
e
view
e
ight
mac
hine
lea
r
n
ing
methods
,
including
s
uppor
t
ve
c
tor
mac
hines
(
S
VM
)
,
Ga
us
s
ian
na
ive
B
a
ye
s
,
de
c
is
ion
tr
e
e
s
(
D
T
)
,
r
a
nd
om
f
o
r
e
s
t
(
R
F
)
,
XG
B
,
k
-
ne
a
r
e
s
t
ne
ighbor
s
(
KN
N
)
,
mul
ti
-
laye
r
pe
r
c
e
ptr
on
(
M
L
P
)
,
a
nd
logi
s
ti
c
r
e
g
r
e
s
s
ion
(
L
R
)
.
T
he
main
objec
ti
ve
o
f
the
s
tudy
is
to
de
mon
s
tr
a
te
the
be
ne
f
icia
l
a
ppli
c
a
ti
on
of
thes
e
methods
f
or
pr
e
di
c
ti
ng
loan
de
f
a
ult
r
is
k
a
nd
identif
ying
inf
luenc
ing
f
a
c
tor
s
.
T
he
pa
pe
r
pr
ovides
a
n
e
xtens
ive
c
ompar
is
on
e
v
a
l
ua
ti
ng
whic
h
mac
hine
lea
r
ning
models
pe
r
f
or
m
be
tt
e
r
with
a
nd
without
their
own
f
e
a
tur
e
s
e
lec
ti
on
method.
J
i
a
ng
e
t
al.
[
24
]
e
xplor
e
the
p
r
oblem
of
c
r
e
dit
s
c
or
in
g
with
a
f
oc
us
on
identif
ying
a
nomalies
a
nd
maintaining
or
de
r
in
f
inanc
ial
tr
a
ns
a
c
ti
ons
.
T
he
y
highl
ight
the
c
las
s
im
ba
lanc
e
pr
oblem
that
a
r
is
e
s
f
r
om
the
li
m
it
e
d
n
umber
of
de
f
a
ult
r
e
c
or
ds
in
f
inanc
ial
da
ta.
T
o
a
d
dr
e
s
s
thi
s
pr
oblem,
the
a
uthor
s
a
na
lyze
va
r
ious
c
las
s
ica
l
a
ppr
oa
c
he
s
to
lea
r
ning
f
r
om
im
ba
lanc
e
d
da
ta,
i
nc
ludi
ng
r
e
s
a
mpl
ing
methods
,
c
os
t
-
of
-
e
r
r
or
s
tr
a
tegie
s
,
a
nd
t
he
us
e
of
ge
ne
r
a
ti
ve
a
dve
r
s
a
r
ial
ne
twor
ks
(
GA
Ns
)
a
s
a
tool
f
or
lea
r
ning
f
r
om
im
ba
lanc
e
d
da
ta.
Abdol
i
e
t
al
.
[
2
5]
e
xa
m
ine
the
im
po
r
ta
nc
e
o
f
a
uto
mate
d
c
r
e
d
it
s
c
or
ing
a
s
a
r
is
k
mana
ge
ment
to
ol
f
o
r
ba
nks
a
nd
f
inan
c
ial
ins
ti
tut
ions
,
not
ing
it
s
a
tt
r
a
c
t
iv
e
ne
s
s
in
r
e
c
e
n
t
de
c
a
de
s
.
T
he
y
h
ighl
igh
t
tha
t
the
u
n
ba
lanc
e
d
na
tu
r
e
o
f
c
r
e
dit
s
c
o
r
in
g
d
a
tas
e
ts
,
a
s
we
l
l
a
s
f
e
a
t
ur
e
he
te
r
oge
ne
it
y,
p
os
e
c
ha
ll
e
n
ge
s
to
de
ve
l
opi
ng
e
f
f
ic
ient
models
that
c
a
n
ge
ne
r
a
l
ize
t
o
p
r
e
vio
us
ly
u
ns
e
e
n
da
ta.
T
he
pa
pe
r
p
r
opos
e
s
t
he
ba
g
ging
s
upe
r
vis
e
d
a
ut
oe
nc
ode
r
c
las
s
if
ie
r
(
B
S
AC
)
,
a
mo
de
l
tha
t
c
o
mbi
ne
s
th
e
be
n
e
f
it
s
of
s
upe
r
vis
e
d
le
a
r
n
ing
f
or
a
ut
oe
nc
ode
r
s
a
nd
a
ba
gg
ing
mec
ha
nis
m
t
o
h
a
ndle
he
ter
oge
ne
it
ies
in
f
e
a
tu
r
e
s
pa
c
e
,
a
nd
the
r
e
s
ult
s
o
f
e
xtens
i
ve
e
xpe
r
im
e
n
ts
c
o
nf
i
r
m
the
s
upe
r
i
or
it
y
a
nd
r
ob
us
tnes
s
of
t
he
pr
o
pos
e
d
me
thod
in
pr
e
dict
ing
t
he
outco
me
o
f
loan
a
pp
li
c
a
ti
o
ns
.
T
he
r
e
leva
nc
e
of
the
topi
c
o
f
p
r
e
dicting
the
c
r
e
dit
wor
thi
ne
s
s
of
c
r
e
dit
s
c
or
ing
us
ing
mac
hine
lea
r
ning
methods
c
a
nnot
be
ove
r
e
s
ti
mate
d
in
the
li
ght
of
moder
n
c
ha
ll
e
nge
s
in
the
f
inanc
ial
indus
tr
y.
W
it
h
the
incr
e
a
s
ing
volum
e
of
da
ta
a
nd
the
va
r
iety
of
f
a
c
to
r
s
a
f
f
e
c
ti
ng
the
f
inanc
ial
s
it
ua
ti
on
o
f
b
or
r
owe
r
s
,
s
tanda
r
d
methods
of
a
s
s
e
s
s
ing
c
r
e
dit
wor
thi
ne
s
s
a
r
e
not
e
f
f
e
c
ti
ve
e
nough.
T
he
us
e
of
a
dva
nc
e
d
mac
hine
lea
r
ning
methods
pr
ovides
the
oppo
r
tuni
ty
no
t
on
ly
f
or
mor
e
a
c
c
ur
a
te
f
or
e
c
a
s
ti
ng,
but
a
ls
o
f
o
r
de
e
pe
r
da
ta
a
na
lys
is
,
whic
h
in
tur
n
he
lps
to
identif
y
e
a
r
ly
s
ig
ns
of
f
inanc
ial
r
is
ks
.
S
olut
ions
ba
s
e
d
on
L
GB
M
c
las
s
if
ier
,
LR
,
L
DA
,
DT
c
las
s
if
ier
,
gr
a
dient
boos
ti
ng
c
las
s
if
ier
a
nd
XG
B
c
las
s
if
ier
pr
omi
s
e
im
pr
ove
d
c
r
e
dit
s
c
or
ing
r
e
s
ult
s
,
whic
h
a
r
e
c
r
it
ica
l
to
e
ns
ur
ing
the
s
us
taina
bil
it
y
of
f
inanc
ial
ins
ti
tut
ions
a
nd
r
e
duc
ing
the
l
ikelihood
of
f
inanc
ial
c
r
is
e
s
.
2.
M
E
T
HO
D
T
h
e
pu
r
p
os
e
o
f
th
is
s
tu
dy
is
a
c
om
pa
r
a
t
ive
a
n
a
ly
s
is
o
f
mac
hi
ne
le
a
r
n
in
g
m
e
t
ho
ds
f
or
p
r
e
di
c
t
in
g
c
r
e
d
it
wo
r
th
ines
s
in
c
r
e
di
t
s
c
o
r
in
g
.
W
e
s
e
t
ou
r
s
e
lv
e
s
t
he
tas
k
o
f
de
te
r
mi
ni
ng
the
o
pt
i
ma
l
me
th
od
a
mo
ng
L
GB
M
c
las
s
i
f
i
e
r
,
LR
,
L
D
A
,
DT
c
l
a
s
s
i
f
ie
r
,
g
r
a
di
e
n
t
b
oos
ti
n
g
c
las
s
if
ie
r
a
nd
X
GB
c
l
a
s
s
i
f
ie
r
,
a
s
we
l
l
a
s
e
va
l
ua
ti
n
g
th
e
i
r
e
f
f
e
c
t
iv
e
ne
s
s
ba
s
e
d
on
s
tan
da
r
d
c
l
a
s
s
i
f
ica
t
io
n
m
e
t
r
ics
.
As
a
b
a
s
is
f
or
ou
r
r
e
s
e
a
r
c
h
,
we
us
e
d
a
l
a
r
ge
a
nd
d
ive
r
s
e
d
a
t
a
s
e
t
tha
t
in
c
l
ud
e
d
in
f
o
r
m
a
t
io
n
a
bo
ut
b
or
r
owe
r
s
'
f
i
na
n
c
i
a
l
s
i
tu
a
t
io
n
,
c
r
e
d
it
his
t
or
y
,
s
oc
ia
l
f
a
c
to
r
s
a
n
d
ot
he
r
r
e
le
va
n
t
va
r
ia
ble
s
.
T
his
da
tas
e
t
p
r
ov
id
e
s
us
wi
th
t
he
o
pp
or
tu
n
it
y
to
m
o
r
e
c
o
m
pr
e
he
ns
ive
l
y
a
na
ly
z
e
a
n
d
e
va
lu
a
te
t
he
pr
op
os
e
d
me
t
ho
ds
.
B
e
f
o
r
e
a
pp
ly
in
g
m
a
c
h
i
ne
lea
r
ni
ng
met
ho
ds
,
c
a
r
e
f
ul
da
ta
p
r
e
p
r
oc
e
s
s
i
ng
w
a
s
c
a
r
r
i
e
d
o
u
t
,
i
nc
l
ud
in
g
ha
nd
li
ng
mi
s
s
i
ng
v
a
l
ue
s
,
e
nc
od
in
g
c
a
te
go
r
ic
a
l
f
e
a
t
ur
e
s
,
n
or
ma
l
izi
ng
nu
me
r
ica
l
da
ta
,
a
nd
p
r
o
c
e
s
s
ing
o
u
tl
ie
r
s
.
T
h
is
s
ta
ge
a
ll
ows
yo
u
t
o
e
ns
u
r
e
t
he
c
o
r
r
e
c
tn
e
s
s
a
nd
s
ta
bi
li
t
y
o
f
the
mo
de
ls
.
L
GB
M
is
a
gr
a
dient
a
mpl
if
ica
ti
on
method
opti
mi
z
e
d
f
or
e
f
f
icie
nt
wo
r
k
with
lar
ge
volum
e
s
of
da
ta.
T
his
method
dr
a
ws
a
tt
e
nti
on
to
taking
int
o
a
c
c
ount
unba
lanc
e
d
c
las
s
e
s
,
whic
h
is
a
n
im
por
tant
a
s
pe
c
t
i
n
c
r
e
dit
s
c
or
ing
pr
oblems
.
L
ogis
ti
c
r
e
g
r
e
s
s
ion
is
a
c
las
s
ic
binar
y
c
las
s
if
ica
ti
on
method
ba
s
e
d
on
a
logi
s
ti
c
f
unc
ti
on.
W
e
us
e
it
in
the
c
ontext
o
f
c
r
e
dit
s
c
or
ing
to
a
s
s
e
s
s
the
li
ke
li
hood
of
a
bor
r
owe
r
's
c
r
e
dit
wor
th
ines
s
a
nd
make
a
de
c
is
ion
ba
s
e
d
on
that
li
ke
li
hood.
L
DA
is
a
li
n
e
a
r
dis
c
r
im
inant
a
na
lys
is
method
de
s
igned
to
m
a
xim
ize
dif
f
e
r
e
nc
e
s
be
twe
e
n
c
las
s
e
s
.
I
n
c
r
e
dit
s
c
or
ing,
thi
s
method
c
a
n
be
e
f
f
e
c
ti
ve
f
or
highl
ight
ing
ke
y
f
e
a
tur
e
s
that
a
f
f
e
c
t
c
r
e
dit
wor
thi
ne
s
s
.
De
c
is
ion
tr
e
e
c
las
s
if
ier
s
pr
ovide
a
vis
ua
l
r
e
pr
e
s
e
ntation
of
de
c
is
ion
making
a
nd
a
r
e
c
a
pa
ble
of
c
a
ptur
ing
c
ompl
e
x
r
e
lations
hips
in
da
t
a
.
W
e
us
e
thi
s
method
to
identi
f
y
the
s
tr
uc
tur
e
o
f
c
r
it
e
r
ia
that
inf
luenc
e
the
f
or
e
c
a
s
ti
ng
of
c
r
e
dit
wor
thi
ne
s
s
.
T
he
gr
a
dient
boos
ti
ng
c
las
s
if
ier
a
ll
ows
the
c
ons
tr
uc
ti
on
of
e
ns
e
mbl
e
s
of
de
c
is
ion
tr
e
e
s
,
whic
h
c
a
n
im
p
r
ove
t
he
pr
e
dictive
powe
r
o
f
the
model.
W
e
'll
e
xplor
e
i
ts
us
e
in
c
r
e
dit
s
c
or
ing
a
nd
e
va
luate
how
it
ha
ndles
c
ompl
e
x
da
ta
s
tr
uc
tur
e
s
.
XG
B
is
a
gr
a
dient
boos
ti
ng
im
pleme
ntation
that
p
r
ovides
a
ddit
ional
op
ti
mi
z
a
t
ions
a
nd
r
e
gular
iza
ti
ons
.
W
e
wi
ll
look
a
t
it
s
i
mpac
t
on
the
a
c
c
ur
a
c
y
of
c
r
e
dit
f
or
e
c
a
s
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
1
4
,
No.
5
,
Oc
tober
2
02
4
:
553
-
554
2
5536
T
o
ob
jec
t
iv
e
l
y
c
o
mp
a
r
e
m
a
c
h
in
e
l
e
a
r
n
in
g
m
e
t
ho
ds
,
w
e
us
e
d
s
ta
nd
a
r
d
met
r
ics
s
uc
h
a
s
a
c
c
u
r
a
c
y
,
r
e
c
a
l
l,
p
r
e
c
is
i
on
a
nd
F
1
-
me
a
s
u
r
e
.
T
he
s
e
me
t
r
ics
e
va
l
ua
te
b
o
th
the
ov
e
r
a
l
l
p
e
r
f
o
r
m
a
nc
e
o
f
t
he
mo
de
ls
a
n
d
t
he
i
r
a
b
il
i
ty
t
o
c
o
r
r
e
c
tl
y
i
de
nt
i
f
y
bo
r
r
o
we
r
s
w
it
h
p
r
ob
le
ma
ti
c
c
r
e
d
i
t
h
is
t
o
r
i
e
s
.
W
e
c
on
du
c
t
e
d
a
s
e
r
ies
o
f
e
x
pe
r
im
e
n
ts
,
t
r
a
i
ni
ng
e
a
c
h
m
ode
l
on
t
he
tr
a
i
ni
ng
s
e
t
a
nd
tes
ti
ng
o
n
t
he
tes
t
s
e
t
.
T
he
r
e
s
ul
ts
a
r
e
a
na
lyz
e
d
us
in
g
e
v
a
l
ua
ti
on
m
e
tr
ics
to
i
de
n
ti
f
y
t
he
b
e
s
t
me
th
ods
t
ha
t
c
a
n
e
f
f
e
c
ti
ve
ly
s
ol
ve
th
e
p
r
ob
le
m
o
f
c
r
e
d
it
wo
r
th
ine
s
s
f
o
r
e
c
a
s
t
in
g
.
3.
RE
S
UL
T
S
AN
D
DI
S
CU
S
S
I
ON
T
o
buil
d
the
models
,
we
us
e
d
the
home
c
r
e
dit
da
tas
e
t
f
r
om
k
aggle.
c
om
,
c
ontaining
65
c
olum
ns
.
T
he
da
ta
s
e
t
include
s
a
c
olum
n
c
a
ll
e
d
T
AR
GE
T
,
whic
h
r
e
pr
e
s
e
nts
the
tar
ge
t
va
r
iable
(
1
-
c
us
tom
e
r
with
pa
yment
dif
f
iculti
e
s
:
he
wa
s
late
in
pa
yment,
0
-
a
ll
other
c
a
s
e
s
)
.
C
olum
n
na
mes
a
nd
de
s
c
r
ipt
ions
a
r
e
given
i
n
T
a
ble
1
(
s
e
e
in
Appe
ndix
)
.
I
n
e
xpe
r
im
e
nts
with
c
r
e
dit
p
r
e
diction
in
a
bin
a
r
y
c
las
s
if
ica
ti
on
tas
k,
we
a
ppli
e
d
a
c
omm
on
thr
e
s
holdi
ng
method
to
c
onve
r
t
p
r
e
dicte
d
p
r
oba
bil
i
ti
e
s
(
)
int
o
binar
y
c
las
s
labe
ls
.
T
his
p
r
oc
e
s
s
is
c
a
r
r
i
e
d
out
us
ing
a
th
r
e
s
hold
s
e
t
a
t
0.
5
:
pr
oba
bil
it
ies
e
qua
l
to
or
gr
e
a
ter
than
0
.
5
a
r
e
r
ounde
d
to
1
(
pos
it
iv
e
c
las
s
)
,
while
pr
oba
bil
it
ies
be
low
0.
5
a
r
e
r
ounde
d
to
0
(
ne
ga
ti
ve
c
las
s
)
.
T
his
a
ppr
oa
c
h
a
ll
ows
us
to
obta
in
c
lea
r
c
a
tegor
ies
of
objec
t
membe
r
s
hip
in
c
las
s
e
s
,
whic
h
s
im
pli
f
ies
the
int
e
r
pr
e
tation
of
c
las
s
if
ica
ti
on
r
e
s
ult
s
a
nd
pr
e
ve
nts
a
mbi
guit
ies
a
s
s
oc
iate
d
with
th
r
e
s
hold
v
a
lues
.
I
n
the
c
ons
ider
e
d
methods
(
L
GB
M
c
las
s
if
ier
,
L
R
,
L
DA
,
DT
c
las
s
if
ier
,
g
r
a
dient
boos
ti
ng
c
las
s
if
ier
a
nd
XG
B
c
las
s
if
ier
)
,
the
ke
y
metr
ics
f
o
r
e
a
c
h
of
th
e
m
we
r
e
e
va
luate
d
a
f
ter
a
pplyi
ng
a
s
e
t
thr
e
s
hold,
a
s
s
ho
wn
in
F
igur
e
1
(
a
)
to
(
f
)
,
whe
r
e
the
r
e
s
ult
s
a
r
e
pr
e
s
e
n
t
:
R
OC
_AU
C
,
a
c
c
ur
a
c
y,
pr
e
c
is
ion,
r
e
c
a
ll
,
s
pe
c
if
icit
y
a
nd
F
1
-
s
c
or
e
.
T
he
s
e
r
e
s
ult
s
r
e
f
lec
t
the
pe
r
f
or
manc
e
of
e
a
c
h
method
a
f
ter
a
pplyi
ng
a
s
tanda
r
d
thr
e
s
hol
d
of
0
.
5.
T
his
pr
oba
bil
i
ty
r
ounding
tec
hnique
plays
a
n
im
por
tant
r
o
le
in
the
c
ons
tr
uc
ti
on
of
binar
y
c
las
s
if
ica
ti
on
models
,
e
ns
ur
ing
their
int
e
r
pr
e
tabi
li
ty
a
nd
a
ppli
c
a
bil
it
y
in
va
r
ious
s
ubjec
t
a
r
e
a
s
,
including
c
r
e
dit
s
c
or
ing
.
(
a
)
(
b)
(
c
)
(
d)
(
e
)
(f)
F
igur
e
1.
M
e
tr
ic
r
e
s
ult
s
a
f
ter
a
djus
tm
e
nt
by
metho
ds
:
(
a
)
R
OC
_AU
C
,
(
b)
a
c
c
ur
a
c
y,
(
c
)
p
r
e
c
is
ion,
(
d
)
r
e
c
a
ll
,
(
e
)
s
pe
c
if
icity,
a
nd
(
f
)
F
1
-
s
c
or
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
F
or
e
c
as
ti
ng
c
r
e
dit
w
or
thi
ne
s
s
in
c
r
e
dit
s
c
or
ing
us
i
ng
mac
hine
lear
nin
g
me
thods
(
A
y
agoz
M
uk
hanov
a
)
5537
B
a
s
e
d
on
the
a
na
lys
i
s
of
the
gr
a
phs
in
F
igu
r
e
1,
we
c
a
n
c
onc
lude
that
ne
ga
ti
ve
c
las
s
e
s
we
r
e
c
or
r
e
c
tl
y
pr
e
dicte
d
in
mos
t
c
a
s
e
s
,
whic
h
is
c
onf
ir
med
by
high
s
pe
c
if
icity
va
lues
.
How
e
ve
r
,
the
ne
e
d
to
a
d
jus
t
the
de
c
is
ion
thr
e
s
hold
is
a
n
int
e
gr
a
l
s
tep
in
opti
mi
z
ing
models
,
e
s
pe
c
ially
whe
n
ba
lanc
ing
be
twe
e
n
f
a
ls
e
pos
it
ives
a
nd
f
a
ls
e
ne
ga
ti
ve
s
.
I
n
thi
s
c
ontext,
it
wa
s
de
c
ided
to
c
onduc
t
e
xpe
r
im
e
nts
with
dif
f
e
r
e
nt
th
r
e
s
hold
va
lues
a
nd
e
va
luate
their
im
pa
c
t
on
ke
y
met
r
ics
s
uc
h
a
s
pr
e
c
is
ion
(
P
r
e
c
is
ion)
,
r
e
c
a
ll
(
R
e
c
a
ll
)
a
nd
F
1
-
mea
s
ur
e
.
Us
ing
dif
f
e
r
e
nt
th
r
e
s
holds
f
or
c
las
s
if
ica
ti
on
a
ll
ows
you
to
tune
the
s
e
ns
it
ivi
ty
of
the
model
to
s
pe
c
if
ic
c
las
s
e
s
in
a
c
c
or
da
nc
e
with
the
r
e
quir
e
ments
of
the
a
ppli
c
a
ti
on
domain.
T
his
is
e
s
pe
c
ially
im
por
tant
in
the
c
o
ntext
of
c
r
e
dit
s
c
or
ing,
whe
r
e
the
we
ight
of
va
r
ious
types
of
e
r
r
o
r
s
c
a
n
be
c
r
it
ica
l.
B
y
f
indi
ng
the
opti
mal
th
r
e
s
hold,
a
ba
lanc
e
c
a
n
be
a
c
hieve
d
be
twe
e
n
mi
nim
izing
f
a
ls
e
pos
it
ives
a
nd
f
a
ls
e
ne
ga
ti
ve
s
,
whic
h
in
tu
r
n
wi
ll
im
pr
ove
the
qua
li
ty
o
f
the
model's
pr
e
dictions
,
a
s
s
hown
in
F
igur
e
2
(
a
)
to
(
f
)
,
whic
h
pr
e
s
e
nts
metr
ic
r
e
s
ult
s
a
f
ter
a
djus
tm
e
nt
by
methods
:
R
OC
_AU
C
,
a
c
c
ur
a
c
y,
pr
e
c
is
ion,
r
e
c
a
ll
,
s
pe
c
if
icity,
a
nd
F
1
-
s
c
or
e
.
(
a
)
(b
)
(
c
)
(d
)
(
e
)
(f)
F
igur
e
2.
M
e
tr
ic
r
e
s
ult
s
a
f
ter
a
djus
tm
e
nt
by
metho
ds
:
(
a
)
R
OC
_
AUC
,
(
b)
a
c
c
ur
a
c
y,
(
c
)
p
r
e
c
is
ion,
(
d
)
r
e
c
a
ll
,
(
e
)
s
pe
c
if
icity,
a
nd
(
f
)
F1
-
s
c
or
e
E
xpe
r
im
e
nts
with
dif
f
e
r
e
nt
de
c
is
ion
thr
e
s
holds
pr
ovide
a
ddit
ional
e
xplor
a
tor
y
e
videnc
e
,
e
xpa
nding
our
unde
r
s
tanding
of
the
model's
s
e
ns
it
ivi
ty
to
dif
f
e
r
e
nt
leve
ls
of
de
c
is
ion
c
onf
idenc
e
.
Ana
lys
is
of
the
c
onf
us
ion
matr
ix
in
F
igur
e
3,
or
e
r
r
or
mat
r
ix,
ba
s
e
d
on
dif
f
e
r
e
nt
thr
e
s
holds
a
ll
ows
you
to
look
in
mo
r
e
de
tail
a
t
the
im
pa
c
t
of
c
ha
nging
the
thr
e
s
hold
on
the
qua
li
ty
of
c
las
s
if
ica
ti
on.
T
his
a
ppr
oa
c
h
is
im
por
tant
f
or
r
e
f
ini
ng
the
model
s
e
tt
ings
in
a
c
c
or
da
nc
e
with
the
s
pe
c
if
ic
r
e
quir
e
ments
a
nd
pr
e
f
e
r
e
nc
e
s
of
the
bus
in
e
s
s
.
T
he
r
e
s
ult
s
obtaine
d
c
a
n
s
e
r
ve
a
s
the
ba
s
is
f
or
mor
e
a
c
c
ur
a
te
a
nd
f
lexible
a
djus
tm
e
nt
of
the
model
w
it
hin
the
f
r
a
mew
or
k
o
f
c
r
e
dit
s
c
or
ing
r
e
qui
r
e
ments
.
I
n
the
f
igur
e
s
,
the
metr
ics
of
the
de
c
is
ion
tr
e
e
c
l
a
s
s
if
ier
model
f
or
the
tr
a
ini
ng
a
nd
tes
t
da
ta
s
e
ts
r
e
maine
d
unc
ha
nge
d.
T
his
is
be
c
a
us
e
the
tr
e
e
r
e
tur
ns
pr
e
dictions
not
a
s
pr
oba
bil
it
ies
,
but
a
s
int
e
ge
r
s
.
T
he
a
c
c
ur
a
c
y
of
the
models
is
high
in
the
or
igi
na
l
tabl
e
s
with
a
th
r
e
s
hold
of
0
.
5,
s
ince
in
the
da
ta
unde
r
s
tudy
the
number
of
one
c
las
s
s
igni
f
ica
ntl
y
e
xc
e
e
ds
the
number
of
a
nother
.
M
ode
ls
a
r
e
good
a
t
pr
e
dicting
ba
d
c
us
tom
e
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
1
4
,
No.
5
,
Oc
tober
2
02
4
:
553
-
554
2
5538
da
ta,
but
ba
d
a
t
pr
e
dicting
good
one
s
.
T
he
r
e
f
or
e
,
it
is
a
dvis
a
ble
to
r
e
ly
on
other
indi
c
a
tor
s
.
Af
ter
a
djus
ti
ng
the
thr
e
s
hold,
the
main
metr
ics
incr
e
a
s
e
d
s
igni
f
ica
ntl
y
f
or
a
ll
models
e
xc
e
pt
the
de
c
is
ion
tr
e
e
,
indi
c
a
ti
ng
the
pos
it
ive
im
pa
c
t
of
c
hoos
ing
the
r
ight
thr
e
s
holds
.
F
igur
e
3.
C
onf
us
ion
matr
ix
of
XG
B
oos
t
met
hod
4.
CONC
L
USI
ON
I
n
thi
s
s
tudy,
a
na
lyzing
models
in
the
c
ontext
of
c
r
e
dit
s
c
or
ing,
s
ix
di
f
f
e
r
e
nt
c
las
s
if
ica
ti
on
a
lgor
it
hms
we
r
e
e
xa
mi
ne
d.
E
a
c
h
o
f
thes
e
models
ha
s
de
mons
tr
a
ted
it
s
unique
c
ha
r
a
c
ter
is
ti
c
s
a
nd
pe
r
f
or
manc
e
,
e
nr
iching
our
unde
r
s
tanding
of
their
a
ppli
c
a
bil
it
y
in
c
r
e
dit
f
or
e
c
a
s
ti
ng
tas
ks
.
T
he
r
e
s
ult
s
highl
ight
the
out
s
tanding
pe
r
f
or
manc
e
of
XG
B
c
las
s
if
ier
,
whic
h
s
tands
out
a
mong
other
models
in
a
ll
metr
ics
r
e
view
e
d,
i
nc
ludi
ng
R
OC
_AU
C
,
a
c
c
ur
a
c
y,
F
1
-
s
c
or
e
,
a
nd
s
pe
c
if
icity.
T
his
de
mons
tr
a
tes
the
high
pe
r
f
or
manc
e
o
f
XG
B
c
las
s
if
ier
in
the
c
ontext
o
f
c
r
e
dit
s
c
or
ing
a
nd
it
s
a
bil
it
y
to
pr
e
dict
c
r
e
dit
wor
thi
ne
s
s
with
a
h
igh
de
g
r
e
e
of
a
c
c
ur
a
c
y.
Additi
ona
l
r
e
s
e
a
r
c
h
int
o
the
va
r
iation
of
de
c
is
ion
thr
e
s
holds
whe
n
r
ounding
c
las
s
membe
r
s
hip
pr
o
ba
bil
it
ies
r
e
ve
a
led
a
s
igni
f
ica
nt
im
p
r
ove
ment
in
the
pr
e
di
c
ti
ve
a
bil
it
y
of
the
models
.
Ana
lyzing
met
r
ics
s
uc
h
a
s
R
OC
_AU
C
,
pr
e
c
is
ion,
r
e
c
a
ll
,
s
pe
c
if
icity,
a
nd
F
1
-
s
c
or
e
a
t
dif
f
e
r
e
nt
thr
e
s
holds
highl
ight
s
the
i
mpor
tanc
e
of
f
indi
ng
a
tr
a
de
-
of
f
be
twe
e
n
f
a
ls
e
pos
it
ives
a
nd
f
a
ls
e
ne
ga
ti
ve
s
.
Ove
r
a
ll
,
the
r
e
s
ult
s
o
f
ou
r
s
tudy
not
only
e
nr
ich
the
unde
r
s
tanding
of
the
pe
r
f
or
manc
e
of
va
r
ious
models
in
c
r
e
dit
s
c
or
ing,
but
a
ls
o
high
li
ght
the
im
por
tanc
e
of
c
a
r
e
f
ul
c
a
li
br
a
ti
on
a
nd
s
e
lec
ti
on
of
opti
mal
th
r
e
s
holds
to
im
pr
ove
f
or
e
c
a
s
ti
ng
pe
r
f
or
manc
e
.
T
he
s
e
f
indi
ngs
pr
ovide
va
luable
guidanc
e
f
or
de
c
i
s
ion
making
in
the
f
ield
of
c
r
e
dit
s
c
or
ing
a
nd
in
th
e
c
ontext
of
f
inanc
ial
r
is
k
mana
ge
ment.
AP
P
E
ND
I
X
T
a
ble
1.
Da
ta
s
e
t
with
de
s
c
r
ipt
ions
C
ol
u
m
n
na
m
e
D
e
s
c
r
i
b
e
E
X
T
_SO
U
R
C
E
_3
N
or
m
a
l
i
z
e
d
s
c
or
e
f
r
o
m
e
x
t
e
r
n
a
l
da
t
a
s
our
c
e
E
X
T
_SO
U
R
C
E
_1
N
or
m
a
l
i
z
e
d
s
c
or
e
f
r
o
m
e
x
t
e
r
n
a
l
da
t
a
s
our
c
e
E
X
T
_SO
U
R
C
E
_2
N
or
m
a
l
i
z
e
d
s
c
or
e
f
r
o
m
e
x
t
e
r
n
a
l
da
t
a
s
our
c
e
D
A
Y
S_B
I
R
T
H
C
l
i
e
nt
'
s
a
g
e
i
n
da
ys
a
t
t
he
t
i
m
e
o
f
a
pp
l
i
c
a
t
i
o
n
A
M
T
_C
R
E
D
I
T
C
r
e
d
i
t
a
m
o
unt
o
f
t
h
e
l
oa
n
A
M
T
_A
N
N
U
I
T
Y
L
oa
n
a
nn
ui
t
y
A
M
T
_G
O
O
D
S_
P
R
I
C
E
F
or
c
on
s
u
m
e
r
l
o
a
ns
i
t
i
s
t
he
p
r
i
c
e
of
t
he
go
ods
f
or
w
h
i
c
h
t
he
l
o
a
n
i
s
gi
ve
n
O
W
N
_C
A
R
_A
G
E
A
ge
o
f
c
l
i
e
nt
'
s
c
a
r
D
A
Y
S_E
M
P
L
O
Y
E
D
H
ow
m
a
ny
d
a
ys
b
e
f
o
r
e
t
he
a
pp
l
i
c
a
t
i
on
t
h
e
pe
r
s
on
s
t
a
r
t
e
d
c
u
r
r
e
n
t
e
m
p
l
oy
m
e
nt
D
A
Y
S_R
E
G
I
ST
R
A
T
I
O
N
H
ow
m
a
ny
d
a
ys
b
e
f
o
r
e
t
he
a
pp
l
i
c
a
t
i
on
di
d
c
l
i
e
n
t
c
h
a
nge
h
i
s
r
e
g
i
s
t
r
a
t
i
o
n
R
E
G
I
O
N
_P
O
P
U
L
A
T
I
O
N
_R
E
L
A
T
I
V
E
N
or
m
a
l
i
z
e
d
pop
ul
a
t
i
on
o
f
r
e
g
i
on
w
h
e
r
e
c
l
i
e
n
t
l
i
ve
s
(
h
i
ghe
r
nu
m
be
r
m
e
a
ns
t
he
c
l
i
e
nt
l
i
v
e
s
i
n
m
o
r
e
pop
ul
a
t
e
d
r
e
g
i
o
n)
D
A
Y
S_I
D
_P
U
B
L
I
SH
H
ow
m
a
ny
d
a
ys
b
e
f
o
r
e
t
he
a
pp
l
i
c
a
t
i
on
di
d
c
l
i
e
n
t
c
h
a
nge
t
he
i
de
n
t
i
t
y
do
c
u
m
e
n
t
w
i
t
h
w
hi
c
h
he
a
ppl
i
e
d
f
o
r
t
he
l
oa
n
A
M
T
_
I
N
C
O
M
E
_T
O
T
A
L
I
nc
om
e
o
f
t
h
e
c
l
i
e
nt
D
A
Y
S_L
A
ST
_P
H
O
N
E
_C
H
A
N
G
E
H
ow
m
a
ny
d
a
ys
b
e
f
o
r
e
a
p
pl
i
c
a
t
i
on
d
i
d
c
l
i
e
nt
c
h
a
ng
e
phon
e
E
N
T
R
A
N
C
E
S_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
F
or
e
c
as
ti
ng
c
r
e
dit
w
or
thi
ne
s
s
in
c
r
e
dit
s
c
or
ing
us
i
ng
mac
hine
lear
nin
g
me
thods
(
A
y
agoz
M
uk
hanov
a
)
5539
T
a
ble
1.
Da
ta
s
e
t
with
de
s
c
r
ipt
ions
(
c
onti
nue
)
C
ol
u
m
n
na
m
e
D
e
s
c
r
i
b
e
A
M
T
_R
E
Q
_C
R
E
D
I
T
_B
U
R
E
A
U
_Y
E
A
R
N
um
b
e
r
o
f
e
nqu
i
r
i
e
s
t
o
C
r
e
di
t
B
ur
e
a
u
a
bou
t
t
he
c
l
i
e
n
t
one
d
a
y
ye
a
r
(
e
xc
l
u
di
n
g
l
a
s
t
3
m
on
t
hs
be
f
o
r
e
a
p
pl
i
c
a
t
i
on
)
Y
E
A
R
S_B
U
I
L
D
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
A
M
T
_R
E
Q
_C
R
E
D
I
T
_B
U
R
E
A
U
_
M
O
N
N
um
b
e
r
o
f
e
nqu
i
r
i
e
s
t
o
C
r
e
di
t
B
ur
e
a
u
a
bou
t
t
he
c
l
i
e
n
t
one
m
on
t
h
b
e
f
o
r
e
a
pp
l
i
c
a
t
i
on
(
e
xc
l
u
di
n
g
one
w
e
e
k
b
e
f
or
e
a
pp
l
i
c
a
t
i
o
n)
L
I
V
I
N
G
A
R
E
A
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
C
O
M
M
O
N
A
R
E
A
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
L
A
N
D
A
R
E
A
_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
L
I
V
I
N
G
A
P
A
R
T
M
E
N
T
S_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
O
B
S_30_C
N
T
_SO
C
I
A
L
_C
I
R
C
L
E
H
ow
m
a
ny
o
bs
e
r
va
t
i
ons
o
f
c
l
i
e
nt
'
s
s
oc
i
a
l
s
u
r
r
ound
i
ng
s
w
i
t
h
obs
e
r
v
a
bl
e
30
D
P
D
(
d
a
ys
p
a
s
t
due
)
d
e
f
a
ul
t
F
L
O
O
R
SM
A
X
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
L
I
V
I
N
G
A
P
A
R
T
M
E
N
T
S_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
L
A
N
D
A
R
E
A
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
L
A
N
D
A
R
E
A
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
N
O
N
L
I
V
I
N
G
A
R
E
A
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
H
O
U
R
_A
P
P
R
_P
R
O
C
E
SS
_ST
A
R
T
A
ppr
ox
i
m
a
t
e
l
y
a
t
w
h
a
t
ho
ur
d
i
d
t
h
e
c
l
i
e
n
t
a
pp
l
y
f
or
t
he
l
o
a
n
R
E
G
I
O
N
_R
A
T
I
N
G
_C
L
I
E
N
T
_
W
_C
I
T
Y
O
ur
r
a
t
i
ng
of
t
he
r
e
gi
on
w
he
r
e
c
l
i
e
nt
l
i
v
e
s
w
i
t
h
t
a
ki
ng
c
i
t
y
i
nt
o
a
c
c
o
unt
(
1
,
2
,
3
)
L
I
V
I
N
G
A
P
A
R
T
M
E
N
T
S_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
L
I
V
I
N
G
A
R
E
A
_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
C
N
T
_F
A
M
_M
E
M
B
E
R
S
H
ow
m
a
ny
f
a
m
i
l
y
m
e
m
be
r
s
do
e
s
c
l
i
e
n
t
h
a
ve
Y
E
A
R
S_B
E
G
I
N
E
X
P
L
U
A
T
A
T
I
O
N
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
A
M
T
_R
E
Q
_C
R
E
D
I
T
_B
U
R
E
A
U
_Q
R
T
N
um
b
e
r
o
f
e
nqu
i
r
i
e
s
t
o
C
r
e
di
t
B
ur
e
a
u
a
bou
t
t
he
c
l
i
e
n
t
3
m
o
nt
h
b
e
f
o
r
e
a
pp
l
i
c
a
t
i
on
(
e
xc
l
ud
i
ng
one
m
ont
h
be
f
or
e
a
pp
l
i
c
a
t
i
o
n)
F
L
O
O
R
SM
I
N
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
Y
E
A
R
S_B
U
I
L
D
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
O
B
S_60_C
N
T
_SO
C
I
A
L
_C
I
R
C
L
E
H
ow
m
a
ny
o
bs
e
r
va
t
i
ons
o
f
c
l
i
e
nt
'
s
s
oc
i
a
l
s
u
r
r
ound
i
ng
s
w
i
t
h
obs
e
r
v
a
bl
e
60
D
P
D
(
d
a
ys
p
a
s
t
due
)
d
e
f
a
ul
t
B
A
SE
M
E
N
T
A
R
E
A
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
g
e
of
bui
l
d
i
ng
,
nu
m
b
e
r
of
e
l
e
va
t
or
s
,
nu
m
be
r
o
f
e
n
t
r
a
nc
e
s
,
s
t
a
t
e
o
f
t
h
e
b
ui
l
di
n
g,
nu
m
be
r
o
f
f
l
oo
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
1
4
,
No.
5
,
Oc
tober
2
02
4
:
553
-
554
2
5540
T
a
ble
1.
Da
ta
s
e
t
with
de
s
c
r
ipt
ions
(
c
onti
nue
)
C
ol
u
m
n
na
m
e
D
e
s
c
r
i
b
e
Y
E
A
R
S_B
U
I
L
D
_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
A
P
A
R
T
M
E
N
T
S_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
C
O
M
M
O
N
A
R
E
A
_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
B
A
SE
M
E
N
T
A
R
E
A
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
N
O
N
L
I
V
I
N
G
A
R
E
A
_M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
A
P
A
R
T
M
E
N
T
S_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
N
O
N
L
I
V
I
N
G
A
P
A
R
T
M
E
N
T
S_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
D
E
F
_60_C
N
T
_
SO
C
I
A
L
_C
I
R
C
L
E
H
ow
m
a
ny
o
bs
e
r
va
t
i
ons
o
f
c
l
i
e
nt
'
s
s
oc
i
a
l
s
u
r
r
ound
i
ng
s
de
f
a
u
l
t
e
d
on
60
(
da
y
s
p
a
s
t
due
)
D
P
D
A
M
T
_R
E
Q
_C
R
E
D
I
T
_B
U
R
E
A
U
_
W
E
E
K
N
um
b
e
r
o
f
e
nqu
i
r
i
e
s
t
o
C
r
e
di
t
B
ur
e
a
u
a
bou
t
t
he
c
l
i
e
n
t
one
w
e
e
k
be
f
or
e
a
p
pl
i
c
a
t
i
on
(
e
xc
l
ud
i
ng
one
da
y
b
e
f
o
r
e
a
p
pl
i
c
a
t
i
on
)
T
O
T
A
L
A
R
E
A
_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
D
E
F
_30_C
N
T
_
SO
C
I
A
L
_C
I
R
C
L
E
H
ow
m
a
ny
o
bs
e
r
va
t
i
ons
o
f
c
l
i
e
nt
'
s
s
oc
i
a
l
s
u
r
r
ound
i
ng
s
de
f
a
u
l
t
e
d
on
30
D
P
D
(
da
ys
pa
s
t
du
e
)
Y
E
A
R
S_B
E
G
I
N
E
X
P
L
U
A
T
A
T
I
O
N
_A
V
G
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
R
E
G
_C
I
T
Y
_N
O
T
_L
I
V
E
_C
I
T
Y
F
l
a
g
i
f
c
l
i
e
nt
'
s
pe
r
m
a
n
e
nt
a
ddr
e
s
s
d
oe
s
n
ot
m
a
t
c
h
c
o
nt
a
c
t
a
ddr
e
s
s
(
1
=
di
f
f
e
r
e
n
t
,
0=
s
a
m
e
,
a
t
c
i
t
y
l
e
ve
l
)
F
L
A
G
_D
O
C
U
M
E
N
T
_
18
D
i
d
c
l
i
e
n
t
pr
ov
i
de
do
c
u
m
e
n
t
18
F
L
A
G
_D
O
C
U
M
E
N
T
_
16
D
i
d
c
l
i
e
n
t
pr
ov
i
de
do
c
u
m
e
n
t
16
F
L
A
G
_D
O
C
U
M
E
N
T
_
8
D
i
d
c
l
i
e
n
t
pr
ov
i
de
do
c
u
m
e
n
t
8
F
L
A
G
_W
O
R
K
_P
H
O
N
E
D
i
d
c
l
i
e
n
t
pr
ov
i
de
ho
m
e
p
hone
(
1=
Y
E
S
,
0=
N
O
)
Y
E
A
R
S_B
E
G
I
N
E
X
P
L
U
A
T
A
T
I
O
N
_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
N
O
N
L
I
V
I
N
G
A
P
A
R
T
M
E
N
T
S_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
F
L
A
G
_D
O
C
U
M
E
N
T
_
3
D
i
d
c
l
i
e
n
t
pr
ov
i
de
do
c
u
m
e
n
t
3
F
L
A
G
_D
O
C
U
M
E
N
T
_
6
D
i
d
c
l
i
e
n
t
pr
ov
i
de
do
c
u
m
e
n
t
6
F
L
A
G
_D
O
C
U
M
E
N
T
_
14
D
i
d
c
l
i
e
n
t
pr
ov
i
de
do
c
u
m
e
n
t
14
A
P
A
R
T
M
E
N
T
S_
M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
B
A
SE
M
E
N
T
A
R
E
A
_
M
E
D
I
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
F
L
O
O
R
SM
A
X
_M
O
D
E
N
or
m
a
l
i
z
e
d
i
n
f
o
r
m
a
t
i
o
n
a
bou
t
bu
i
l
d
i
n
g
w
h
e
r
e
t
he
c
l
i
e
n
t
l
i
ve
s
,
w
ha
t
i
s
a
v
e
r
a
ge
(
_
A
V
G
s
uf
f
i
x
)
,
m
od
us
(
_
M
O
D
E
s
u
f
f
i
x
)
,
m
e
di
a
n
(
_
M
E
D
I
s
uf
f
i
x
)
a
pa
r
t
m
e
nt
s
i
z
e
,
c
o
m
m
on
a
r
e
a
,
l
i
v
i
ng
a
r
e
a
,
a
ge
of
bu
i
l
di
ng
,
n
u
m
be
r
of
e
l
e
v
a
t
o
r
s
,
nu
m
b
e
r
of
e
nt
r
a
n
c
e
s
,
s
t
a
t
e
of
t
he
b
ui
l
d
i
ng
,
nu
m
be
r
of
f
l
o
or
F
L
A
G
_E
M
P
_P
H
O
N
E
D
i
d
c
l
i
e
n
t
pr
ov
i
de
w
or
k
phon
e
(
1=
Y
E
S
,
0=
N
O
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
F
or
e
c
as
ti
ng
c
r
e
dit
w
or
thi
ne
s
s
in
c
r
e
dit
s
c
or
ing
us
i
ng
mac
hine
lear
nin
g
me
thods
(
A
y
agoz
M
uk
hanov
a
)
5541
AC
KNOWL
E
DGE
M
E
NT
S
T
his
r
e
s
e
a
r
c
h
ha
s
be
e
n
f
unde
d
by
the
S
c
ienc
e
C
omm
it
tee
of
the
M
ini
s
tr
y
of
S
c
ienc
e
a
nd
Highe
r
E
duc
a
ti
on
of
the
R
e
publi
c
of
Ka
z
a
khs
tan
(
Gr
a
nt
N
o.
AP19677451)
.
RE
F
E
RE
NC
E
S
[
1]
Z
.
H
e
,
J
.
H
ua
ng,
a
nd
J
.
Z
hou,
“
O
pe
n
ba
nki
ng:
c
r
e
di
t
ma
r
ke
t
c
ompe
ti
ti
on
w
he
n
bor
r
ow
e
r
s
ow
n
th
e
da
ta
,”
J
our
nal
of
F
in
an
c
ia
l
E
c
onomic
s
, vol
. 147, no. 2, pp. 449
–
474, F
e
b. 2023, doi:
10.1016/j
.j
f
in
e
c
o.2022.12.003.
[
2]
U
.
T
.
M
a
kha
z
ha
nov
a
,
F
.
A
.
M
ur
z
in
,
A
.
A
.
M
ukha
nova
,
a
nd
E
.
P
.
A
br
a
mov,
“
F
uz
z
y
lo
gi
c
o
f
Z
a
de
h
a
nd
de
c
is
io
n
-
ma
ki
ng
in
th
e
f
ie
ld
of
l
oa
n,”
J
our
nal
of
t
he
or
e
ti
c
al
and appli
e
d I
nf
or
m
at
io
n T
e
c
hnol
ogy
, vol
. 98, no. 06, pp. 1076
–
1086, 2020.
[
3]
U
.
M
a
kha
z
ha
nova
e
t
al
.
,
“
T
he
e
va
lu
a
ti
on
of
c
r
e
di
twor
th
in
e
s
s
of
tr
a
de
a
nd
e
nt
e
r
p
r
is
e
s
of
s
e
r
vi
c
e
us
in
g
th
e
me
th
od
ba
s
e
d
on
f
u
z
z
y
lo
gi
c
,”
A
ppl
ie
d Sc
ie
nc
e
s
, vol
. 12, no. 22, Nov. 20
22, doi:
10.33
90/
a
pp122211515.
[
4]
Y
.
W
a
ng,
Y
.
L
iu
,
J
.
Z
ha
o,
a
nd
Q
.
Z
ha
ng,
“
L
ow
-
c
ompl
e
xi
ty
f
a
s
t
C
U
c
la
s
s
if
ic
a
ti
on
de
c
is
io
n
me
th
od
b
a
s
e
d
on
L
G
B
M
c
la
s
s
if
ie
r
,”
E
le
c
tr
oni
c
s
, vol
. 12, no. 11, M
a
y 2023, doi:
10.3390/ele
c
tr
oni
c
s
12112488.
[
5]
İ
.
F
.
K
il
in
ç
e
r
a
nd
O
.
K
a
ta
r
,
“
A
ne
w
in
tr
us
io
n
de
te
c
ti
on
s
ys
te
m
f
or
s
e
c
ur
e
d
I
oT
/I
I
oT
n
e
twor
ks
ba
s
e
d
on
L
G
B
M
,”
G
a
z
i
U
ni
v
e
r
s
it
y
J
our
nal
of
Sc
ie
nc
e
P
ar
t
C
:
D
e
s
ig
n and T
e
c
hnol
ogy
, 2023.
[
6]
T
.
L
iu
,
X
.
Z
ha
ng,
R
.
C
he
n,
X
.
D
e
ng,
a
nd
B
.
F
u,
“
D
e
ve
lo
pme
n
t,
c
ompa
r
is
on,
a
nd
v
a
li
da
ti
on
of
f
our
in
te
ll
ig
e
nt
,
pr
a
c
ti
c
a
l
ma
c
hi
n
e
le
a
r
ni
ng
mode
ls
f
or
pa
ti
e
nt
s
w
it
h
pr
os
ta
te
-
s
pe
c
if
ic
a
nt
ig
e
n
in
th
e
gr
a
y
z
one
,”
F
r
ont
ie
r
s
in
O
nc
ol
ogy
,
vol
.
13,
J
un.
2023,
doi
:
10.3389/f
onc
.2023.1157384.
[
7]
J
.
T
u
s
s
upov
e
t
al
.
,
“
A
n
a
ly
s
is
of
f
or
ma
l
c
onc
e
pt
s
f
or
ve
r
if
ic
a
ti
on
of
pe
s
ts
a
nd
di
s
e
a
s
e
s
of
c
r
ops
us
in
g
ma
c
hi
ne
le
a
r
ni
ng
me
th
o
ds
,”
I
E
E
E
A
c
c
e
s
s
, vol
. 12, pp. 19902
–
19910, 2024, doi:
10.1109/AC
C
E
S
S
.2024.3361046.
[
8]
W.
-
Y
. L
oh, “
L
ogi
s
ti
c
r
e
gr
e
s
s
io
n t
r
e
e
a
na
ly
s
i
s
,”
i
n
Spr
in
ge
r
H
a
ndbook
of
E
ngi
ne
e
r
in
g St
at
is
ti
c
s
, 2023, pp. 593
–
604.
[
9]
G
. T
r
oi
a
no
e
t
a
l.
, “
D
e
ve
lo
pme
nt
a
nd i
nt
e
r
na
ti
ona
l
va
li
da
ti
on o
f
l
ogi
s
ti
c
r
e
gr
e
s
s
io
n a
nd ma
c
hi
ne
‐
le
a
r
ni
ng mode
ls
f
or
th
e
pr
e
di
c
ti
on
of
10‐
ye
a
r
mol
a
r
l
os
s
,”
J
our
nal
of
C
li
ni
c
al
P
e
r
io
dont
ol
ogy
, vol
.
50, no. 3, pp. 348
–
357, M
a
r
. 2023,
doi
:
10.1111/j
c
pe
.13739.
[
10]
R
.
G
r
a
f
,
M
.
Z
e
ld
ovi
c
h,
a
nd
S
.
F
r
ie
dr
ic
h,
“
C
ompa
r
in
g
li
ne
a
r
di
s
c
r
im
in
a
nt
a
na
ly
s
is
a
nd
s
upe
r
vi
s
e
d
le
a
r
ni
ng
a
lg
or
it
hms
f
or
bi
n
a
r
y
c
la
s
s
if
ic
a
ti
on
—
a
me
th
od c
omp
a
r
is
on s
tu
dy,”
B
io
m
e
tr
ic
al
J
our
n
al
, vol
. 66, no. 1, J
a
n. 2024,
doi
:
10.1002/bi
mj
.202200098.
[
11]
T
.
S
ue
s
s
e
,
A
.
B
r
e
nni
ng,
a
nd
V
.
G
r
upp,
“
S
pa
ti
a
l
li
ne
a
r
di
s
c
r
im
i
na
nt
a
na
ly
s
i
s
a
ppr
oa
c
he
s
f
or
r
e
mot
e
-
s
e
ns
in
g c
la
s
s
if
ic
a
ti
on,”
Sp
at
ia
l
St
at
is
ti
c
s
, vol
. 57, Oc
t.
2023, doi:
10.1016/j
.s
pa
s
ta
.2023.10077
5.
[
12]
G
. S
in
gh,
Y
. P
a
l,
a
nd A
. K
. D
a
hi
ya
, “
C
la
s
s
if
ic
a
ti
on of
pow
e
r
q
ua
li
ty
di
s
tu
r
ba
nc
e
s
us
in
g l
in
e
a
r
di
s
c
r
im
in
a
nt
a
na
ly
s
is
,”
A
ppl
ie
d
Sof
t
C
om
put
in
g
, vol
. 138, M
a
y 2023, doi:
10.1016/j
.a
s
oc
.2023.1101
81.
[
13]
G
.
D
e
vi
s
e
tt
y
a
nd
N
.
S
.
K
uma
r
,
“
P
r
e
di
c
ti
on
of
br
a
dyc
a
r
di
a
us
in
g
de
c
is
io
n
tr
e
e
a
lg
or
it
hm
a
nd
c
ompa
r
in
g
th
e
a
c
c
ur
a
c
y
w
it
h
s
up
por
t
ve
c
to
r
ma
c
hi
ne
,”
E
3S W
e
b of
C
onf
e
r
e
n
c
e
s
, vol
. 399,
J
ul
. 2023, doi:
10.1051/e3s
c
onf
/2
02339909004.
[
14]
H
.
C
he
n,
G
.
Z
ha
ng,
X
.
P
a
n,
a
nd
R
.
J
ia
,
“
U
s
in
g
dua
l
e
vol
ut
io
na
r
y
s
e
a
r
c
h
to
c
on
s
tr
uc
t
de
c
i
s
io
n
tr
e
e
b
a
s
e
d
e
ns
e
mbl
e
c
l
a
s
s
if
i
e
r
,”
C
om
pl
e
x
& I
nt
e
ll
ig
e
nt
Sy
s
te
m
s
, vol
. 9, no. 2, pp. 1327
–
1345, A
pr
. 2023, doi:
10.1007/s
40747
-
022
-
00855
-
x.
[
15]
A
bdul
la
h
-
A
ll
-
T
a
nvi
r
,
I
.
A
.
K
ha
ndoka
r
,
A
.
K
.
M
.
M
.
I
s
la
m,
S
.
I
s
la
m,
a
nd
S
.
S
ha
ta
bda
,
“
A
gr
a
di
e
nt
boos
ti
ng
c
la
s
s
if
ie
r
f
or
pur
c
ha
s
e
i
nt
e
nt
io
n pr
e
di
c
ti
on of
onl
in
e
s
hoppe
r
s
,”
H
e
li
y
on
, vol
. 9, no. 4, Apr
. 2023,
doi
:
10.1016/j
.he
li
yon.2023.e
15163.
[
16]
R
.
S
uhe
ndr
a
e
t
al
.
,
“
E
va
lu
a
ti
on
of
gr
a
di
e
nt
boos
te
d
c
la
s
s
if
ie
r
in
a
to
pi
c
de
r
ma
ti
ti
s
s
e
ve
r
it
y
s
c
or
e
c
la
s
s
if
ic
a
ti
on,”
H
e
c
a
J
our
na
l
of
A
ppl
ie
d Sc
ie
nc
e
s
, vol
. 1, no. 2, pp. 54
–
61, S
e
p. 2023, doi:
10.6
0084/hj
a
s
.v1i
2.85.
[
17]
H
.
N
ha
t
-
D
uc
a
nd
T
.
V
a
n
-
D
uc
,
“
C
ompa
r
is
on
of
hi
s
to
gr
a
m
-
ba
s
e
d
gr
a
di
e
nt
boo
s
ti
ng
c
la
s
s
if
ic
a
ti
on
ma
c
hi
ne
,
r
a
ndom
F
or
e
s
t,
a
nd
de
e
p
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
f
or
pa
ve
me
nt
r
a
ve
li
ng
s
e
ve
r
it
y
c
la
s
s
if
ic
a
ti
on,”
A
ut
om
at
io
n
in
C
ons
tr
uc
ti
on
,
vol
.
148,
A
pr
.
2023, doi:
10.1016/j
.a
ut
c
on.2023.104767.
[
18]
V
.
J
a
in
a
nd
M
.
A
gr
a
w
a
l,
“
H
e
a
r
t
f
a
il
ur
e
pr
e
di
c
ti
on
u
s
in
g
X
G
B
c
la
s
s
if
ie
r
,
lo
gi
s
ti
c
r
e
gr
e
s
s
io
n
a
nd
s
uppor
t
ve
c
to
r
c
la
s
s
if
ie
r
,”
i
n
2023
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
A
dv
anc
e
m
e
nt
in
C
om
put
at
io
n
&
C
om
put
e
r
T
e
c
hnol
ogi
e
s
(
I
nC
A
C
C
T
)
,
M
a
y
2023,
pp.
1
–
5,
doi
:
10.1109/I
nC
A
C
C
T
57535.2023.10141752.
[
19]
K
.
K
ona
r
,
S
.
D
a
s
,
a
nd
S
.
D
a
s
,
“
E
mpl
oye
e
a
tt
r
it
io
n
pr
e
di
c
ti
on
f
or
im
ba
la
nc
e
d
da
ta
us
in
g
ge
n
e
ti
c
a
lg
or
it
hm
-
ba
s
e
d
pa
r
a
me
te
r
opt
im
iz
a
ti
on
of
X
G
B
c
la
s
s
if
ie
r
,”
in
2023
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
C
om
put
e
r
,
E
le
c
tr
ic
al
&
C
om
m
uni
c
at
io
n
E
ngi
ne
e
r
in
g
(
I
C
C
E
C
E
)
, J
a
n. 2023, pp. 1
–
6, doi:
10.1109/I
C
C
E
C
E
51049.20
23.10085402.
[
20]
X
.
D
a
s
ti
le
,
T
.
C
e
li
k,
a
nd
M
.
P
ot
s
a
ne
,
“
S
ta
ti
s
ti
c
a
l
a
nd
ma
c
hi
ne
le
a
r
ni
ng
mode
ls
in
c
r
e
di
t
s
c
or
in
g:
a
s
y
s
te
ma
ti
c
li
te
r
a
tu
r
e
s
ur
v
e
y,”
A
ppl
ie
d Soft
C
om
put
in
g
, vol
. 91, J
un. 2020, doi:
10.1016/j
.a
s
oc
.2020.106263.
[
21]
G
. T
e
l
e
s
, J
. J
. P
.
C
.
R
odr
ig
ue
s
,
K
. S
a
l
e
e
m, S
.
K
oz
lo
v,
a
nd
R
.
A
.
L
. R
a
b
ê
lo
, “
M
a
c
hi
ne
l
e
a
r
n
in
g
a
nd d
e
c
is
i
on
s
upp
or
t
s
ys
te
m
on
c
r
e
di
t
s
c
or
i
ng,
”
N
e
u
r
al
C
o
m
put
in
g
and A
p
pl
ic
at
io
n
s
,
vol
. 32
, no
. 14
, p
p. 98
09
–
9
826,
J
u
l.
2
020,
doi
:
10.
1007/
s
0
0521
-
019
-
0
4537
-
7.
[
22]
E
.
S
.
K
a
m
im
u
r
a
,
A
.
R
.
F
.
P
in
to
,
a
n
d
M
.
S
.
N
a
g
a
n
o,
“
A
r
e
c
e
n
t
r
e
v
i
e
w
on
op
ti
m
i
s
a
ti
on
me
th
od
s
a
ppl
i
e
d
to
c
r
e
d
it
s
c
or
in
g
m
od
e
l
s
,
”
J
o
u
r
n
al
of
E
c
o
no
m
i
c
s
,
F
i
na
n
c
e
a
nd
A
d
m
i
ni
s
t
r
a
ti
v
e
S
c
ie
n
c
e
,
vo
l.
2
8
,
no
.
56
,
pp
.
35
2
–
37
1,
2
02
3,
d
oi
:
1
0.
11
08
/
J
E
F
A
S
-
09
-
20
21
-
01
93
.
[
23]
O
.
K
oc
,
O
.
U
gur
,
a
nd
A
.
S
.
K
e
s
te
l,
“
T
he
im
pa
c
t
of
f
e
a
tu
r
e
s
e
le
c
ti
on
a
nd
tr
a
n
s
f
or
ma
ti
on
on
ma
c
hi
ne
le
a
r
ni
ng
me
th
ods
in
de
te
r
mi
ni
ng t
he
c
r
e
di
t
s
c
or
in
g,”
ar
X
iv
pr
e
p
r
in
t
ar
X
iv
:
2303.05427
, 2023.
[
24]
C
.
J
ia
ng,
W
.
L
u,
Z
.
W
a
ng,
a
nd
Y
.
D
in
g,
“
B
e
nc
hma
r
ki
ng
s
ta
te
-
of
-
th
e
-
a
r
t
im
ba
la
nc
e
d
da
ta
le
a
r
ni
ng
a
ppr
oa
c
he
s
f
or
c
r
e
di
t
s
c
or
i
ng,”
E
x
pe
r
t
Sy
s
te
m
s
w
it
h A
ppl
ic
at
io
ns
, vol
. 213, M
a
r
. 2023, doi:
10.
1016/j
.e
s
w
a
.2022.118878.
[
25]
M
.
A
bdol
i,
M
.
A
kba
r
i,
a
nd
J
.
S
ha
hr
a
bi
,
“
B
a
ggi
ng
s
upe
r
vi
s
e
d
a
ut
oe
nc
ode
r
c
la
s
s
if
ie
r
f
or
c
r
e
di
t
s
c
or
in
g,”
E
x
pe
r
t
Sy
s
te
m
s
w
it
h
A
ppl
ic
at
io
ns
, vol
. 213, M
a
r
. 2023, doi:
10.1016/j
.e
s
w
a
.2022.11
8991.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
A
y
a
g
o
z
M
ukha
no
v
a
recei
v
ed
h
er
Ph
D
i
n
2
0
1
5
i
n
i
n
fo
rmat
i
o
n
s
y
s
t
em
s
fro
m
L
.
N
.
G
u
mi
l
y
o
v
E
u
ra
s
i
a
n
N
a
t
i
o
n
a
l
U
n
i
v
er
s
i
t
y
,
K
aza
k
h
s
t
a
n
.
Cu
rren
t
l
y
,
s
h
e
i
s
an
as
s
o
c
i
at
e
p
ro
fe
s
s
o
r
o
f
t
h
e
D
ep
ar
t
men
t
o
f
I
n
fo
rma
t
i
o
n
Sy
s
t
ems
at
t
h
e
s
ame
u
n
i
v
er
s
i
t
y
.
H
er
r
es
earc
h
i
n
t
ere
s
t
s
i
n
cl
u
d
e
art
i
f
i
ci
a
l
i
n
t
e
l
l
i
g
e
n
ce
an
d
d
ec
i
s
i
o
n
mak
i
n
g
.
Sh
e
can
b
e
co
n
t
ac
t
ed
at
emai
l
:
ay
ag
o
z1
9
8
3
0
2
@
mai
l
.
ru
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
1
4
,
No.
5
,
Oc
tober
2
02
4
:
553
-
554
2
5542
M
a
di
y
a
r
Ba
i
te
m
i
r
o
v
i
n
2
0
2
3
,
h
e
g
rad
u
at
e
d
fro
m
t
h
e
E
u
ra
s
i
a
n
N
at
i
o
n
al
U
n
i
v
er
s
i
t
y
n
amed
aft
er
L
.
N
.
G
u
mi
l
y
o
v
w
i
t
h
a
d
eg
ree
i
n
i
n
fo
rma
t
i
o
n
s
y
s
t
ems
.
Cu
rre
n
t
l
y
,
h
e
i
s
a
mas
t
er's
s
t
u
d
e
n
t
at
t
h
e
E
u
ra
s
i
a
n
N
a
t
i
o
n
a
l
U
n
i
v
ers
i
t
y
.
Sci
en
t
i
f
i
c
i
n
t
ere
s
t
s
-
mach
i
n
e
l
ear
n
i
n
g
,
d
a
t
a
s
c
i
en
ce,
d
at
a
mi
n
i
n
g
,
an
d
art
i
f
i
ci
a
l
i
n
t
e
l
l
i
g
e
n
ce.
H
e
can
b
e
co
n
t
act
e
d
v
i
a
ema
i
l
at
:
mad
i
y
ar.
b
ay
t
emi
r
o
v
@
i
n
b
o
x
.
r
u
.
A
za
m
a
t
A
m
i
r
o
v
in
i
n
f
o
rmat
i
cs
,
co
m
p
u
t
i
n
g
,
an
d
co
n
t
r
o
l
s
y
s
t
ems
fro
m
L
.
N
.
G
u
mi
l
y
o
v
E
u
ra
s
i
a
n
N
a
t
i
o
n
a
l
U
n
i
v
er
s
i
t
y
,
K
az
ak
h
s
t
an
.
Cu
rre
n
t
l
y
,
h
e
i
s
t
h
e
ac
t
i
n
g
a
s
s
o
ci
a
t
e
p
ro
fe
s
s
o
r
an
d
d
i
rec
t
o
r
o
f
t
h
e
D
i
g
i
t
al
i
zat
i
o
n
D
e
p
art
me
n
t
at
A
b
y
l
k
a
s
Sag
i
n
o
v
K
arag
a
n
d
a
T
ec
h
n
i
cal
U
n
i
v
er
s
i
t
y
.
H
i
s
res
earc
h
i
n
t
eres
t
s
i
n
cl
u
d
e
i
n
fo
rma
t
i
o
n
t
ec
h
n
o
l
o
g
y
,
s
mart
ci
t
i
es
,
b
i
g
d
at
a,
an
d
t
h
e
i
n
t
ern
e
t
o
f
t
h
i
n
g
s
.
H
e
can
b
e
co
n
t
act
e
d
at
ema
i
l
:
a.
ami
ro
v
@
k
s
t
u
.
k
z.
Bo
l
a
t
Ta
s
s
u
o
v
i
n
1
9
9
7
h
e
g
rad
u
a
t
ed
fro
m
Z
h
am
b
y
l
U
n
i
v
ers
i
t
y
w
i
t
h
a
d
eg
ree
i
n
mat
h
ema
t
i
c
s
an
d
co
m
p
u
t
er
s
ci
e
n
ce
.
In
2
0
0
6
,
h
e
d
efen
d
ed
h
i
s
d
i
s
s
er
t
at
i
o
n
an
d
recei
v
ed
t
h
e
d
eg
ree
o
f
Can
d
i
d
at
e
o
f
T
ech
n
i
ca
l
Sci
e
n
ces
.
H
e
s
t
art
e
d
h
i
s
care
er
i
n
1
9
9
8
a
s
a
s
o
f
t
w
are
en
g
i
n
eer
at
M.
H
.
D
u
l
at
y
T
araz
St
at
e
U
n
i
v
er
s
i
t
y
.
Cu
rre
n
t
l
y
,
h
e
i
s
t
h
e
d
ean
o
f
t
h
e
Facu
l
t
y
o
f
"
N
at
u
ra
l
Sci
en
ce
s
"
o
f
t
h
e
N
o
n
-
p
r
o
fi
t
l
i
mi
t
ed
L
i
ab
i
l
i
t
y
Co
m
p
an
y
"
M.
H
.
D
u
l
a
t
y
T
a
raz
St
at
e
U
n
i
v
ers
i
t
y
"
.
H
e
i
s
t
h
e
a
u
t
h
o
r
o
f
mo
re
t
h
a
n
6
0
s
c
i
en
t
i
f
i
c
p
ap
e
rs
,
i
n
cl
u
d
i
n
g
1
mo
n
o
g
rap
h
,
3
t
ex
t
b
o
o
k
s
a
n
d
2
ar
t
i
c
l
es
i
n
t
h
e
Sco
p
u
s
d
at
a
b
as
e.
Res
earc
h
i
n
t
eres
t
s
–
i
n
fo
rma
t
i
o
n
s
ecu
r
i
t
y
,
co
mp
u
t
er
mo
d
e
l
i
n
g
.
H
e
can
b
e
c
o
n
t
act
e
d
at
emai
l
:
b
o
l
at
_
k
t
n
@
mai
l
.
ru
.
V
a
l
enti
na
M
a
kh
a
to
v
a
c
an
d
i
d
at
e
o
f
T
ech
n
i
ca
l
Sci
e
n
ces
,
cu
rren
t
l
y
p
r
o
fes
s
o
r
o
f
t
h
e
D
ep
ar
t
men
t
o
f
S
o
ft
w
are
E
n
g
i
n
eeri
n
g
a
t
A
t
y
ra
u
St
a
t
e
U
n
i
v
ers
i
t
y
K
h
.
D
o
s
mu
k
h
am
ed
o
v
a,
A
t
y
ra
u
,
K
azak
h
s
t
an
.
Sh
e
h
a
s
mo
re
t
h
a
n
1
4
0
s
c
i
en
t
i
f
i
c
p
a
p
ers
,
i
n
cl
u
d
i
n
g
8
p
a
p
ers
i
n
W
e
b
o
f
Sc
i
en
ce
an
d
Sco
p
u
s
ra
t
i
n
g
p
u
b
l
i
ca
t
i
o
n
s
,
3
m
o
n
o
g
ra
p
h
s
,
8
t
e
x
t
b
o
o
k
s
an
d
5
c
o
p
y
ri
g
h
t
cert
i
fi
ca
t
es
o
f
i
n
t
el
l
ect
u
a
l
p
ro
p
ert
y
.
H
-
i
n
d
e
x
–
5
.
Sh
e
w
a
s
t
h
e
ex
ec
u
t
o
r
o
f
t
h
e
p
ro
j
ect
o
f
s
earc
h
an
d
i
n
i
t
i
at
i
v
e
re
s
earch
w
o
rk
o
n
t
h
e
t
o
p
i
c
fu
n
d
amen
t
al
p
a
t
t
er
n
s
o
f
rh
eo
l
o
g
i
ca
l
p
ro
p
e
rt
i
e
s
o
f
n
an
o
c
o
mp
o
s
i
t
e
mat
er
i
al
s
”
.
G
ran
t
fu
n
d
i
n
g
fro
m
t
h
e
M
i
n
i
s
t
ry
o
f
E
d
u
ca
t
i
o
n
a
n
d
Sc
i
en
ce
o
f
t
h
e
Rep
u
b
l
i
c
o
f
K
aza
k
h
s
t
a
n
2
0
1
8
-
2
0
2
0
.
Sh
e
can
b
e
co
n
t
ac
t
ed
a
t
emai
l
:
mah
v
e@
mai
l
.
r
u
.
A
s
s
em
g
ul
Ka
i
po
v
a
Mas
t
er
o
f
T
ech
n
i
cal
Sci
e
n
ces
.
Sh
e
i
s
cu
rren
t
l
y
a
s
en
i
o
r
l
ect
u
rer
i
n
t
h
e
D
e
p
art
me
n
t
o
f
Bi
o
s
t
a
t
i
s
t
i
cs
,
Bi
o
i
n
fo
rma
t
i
c
s
an
d
In
fo
rmat
i
o
n
T
ec
h
n
o
l
o
g
i
es
.
Sh
e
i
s
t
h
e
au
t
h
o
r
o
f
mo
re
t
h
a
n
1
0
s
c
i
en
t
i
f
i
c
p
ap
er
s
,
1
art
i
c
l
e
i
n
t
h
e
Sco
p
u
s
d
at
a
b
as
e.
Sh
e
can
b
e
co
n
t
ac
t
ed
b
y
emai
l
:
k
ai
p
o
v
a.
a@
amu
.
k
z
.
U
l
zha
n
M
a
kh
a
zha
n
o
v
a
i
n
2
0
0
8
s
h
e
g
rad
u
at
ed
fro
m
E
u
ras
i
an
N
at
i
o
n
al
U
n
i
v
er
s
i
t
y
w
i
t
h
a
d
eg
ree
i
n
co
m
p
u
t
er
s
ci
e
n
ce
.
In
2
0
1
1
h
e
recei
v
e
d
a
mas
t
er's
d
eg
ree
i
n
co
m
p
u
t
er
s
ci
e
n
c
e.
In
2
0
2
1
,
s
h
e
g
ra
d
u
a
t
ed
fr
o
m
E
u
ras
i
an
N
a
t
i
o
n
a
l
U
n
i
v
e
rs
i
t
y
d
o
c
t
o
ra
l
s
t
u
d
i
e
s
w
i
t
h
a
s
p
eci
a
l
t
y
6
D
0
7
0
3
0
0
–
i
n
f
o
rmat
i
o
n
s
y
s
t
em
s
.
Fro
m
2
0
2
0
t
o
t
h
e
p
res
en
t
,
h
e
i
s
a
s
en
i
o
r
l
ect
u
rer
at
t
h
e
D
ep
ar
t
men
t
o
f
I
n
fo
rma
t
i
o
n
S
y
s
t
ems
,
L
.
N
.
G
u
mi
l
y
o
v
E
u
ras
i
a
n
N
a
t
i
o
n
a
l
U
n
i
v
ers
i
t
y
,
A
s
t
an
a.
S
h
e
i
s
t
h
e
a
u
t
h
o
r
o
f
mo
re
t
h
a
n
2
2
w
o
r
k
s
.
Res
earch
i
n
t
eres
t
s
:
d
at
a
a
n
al
y
s
i
s
,
b
i
g
d
a
t
a,
mach
i
n
e
l
earn
i
n
g
,
fu
zzy
l
o
g
i
c.
Sh
e
can
b
e
c
o
n
t
act
e
d
at
emai
l
:
ma
k
h
az
h
an
.
u
t
@
g
mai
l
.
co
m
.
Tl
eug
a
i
s
h
a
Os
p
a
no
v
a
1
9
8
1
g
ra
d
u
a
t
ed
fr
o
m
t
h
e
K
azak
h
St
a
t
e
U
n
i
v
ers
i
t
y
n
ame
d
aft
er.
S.
M.
K
i
r
o
v
a,
Facu
l
t
y
o
f
Mech
a
n
i
c
s
an
d
A
p
p
l
i
ed
Ma
t
h
ema
t
i
c
s
,
s
p
eci
a
l
t
y
ap
p
l
i
ed
mat
h
ema
t
i
c
s
.
Can
d
i
d
a
t
e
o
f
T
ech
n
i
ca
l
Sci
e
n
ces
,
s
p
ec
i
a
l
t
y
1
3
/
0
5
/
2
0
1
8
-
ma
t
h
ema
t
i
ca
l
mo
d
el
i
n
g
,
n
u
mer
i
cal
met
h
o
d
s
a
n
d
s
o
ft
w
are
p
ac
k
ag
e
s
.
Si
n
ce
2
0
1
1
,
s
h
e
h
a
s
b
ee
n
w
o
r
k
i
n
g
a
t
E
N
U
n
ame
d
aft
er.
L
.
N
.
G
u
mi
l
ev
a,
D
ep
ar
t
men
t
o
f
In
fo
rmat
i
o
n
Sy
s
t
e
ms
.
H
er
res
earch
i
n
t
ere
s
t
’
s
i
n
f
o
rmat
i
o
n
an
d
co
mp
u
t
er
t
ec
h
n
o
l
o
g
i
es
a
n
d
s
y
s
t
ems
,
art
i
fi
c
i
al
i
n
t
e
l
l
i
g
en
ce
.
Sh
e
can
b
e
co
n
t
act
e
d
at
emai
l
:
T
l
e
u
2
0
0
9
@
mai
l
.
ru
.
Evaluation Warning : The document was created with Spire.PDF for Python.