C
omp
u
te
r
S
c
i
e
n
c
e
an
d
I
n
for
mati
on
T
e
c
h
n
ol
ogi
e
s
V
ol
.
7
,
N
o
.
1
,
M
a
r
c
h
20
25
,
p
p.
46
~
5
5
IS
S
N
:
2722
-
3221
,
D
O
I
:
10.
1
1591
/
c
s
i
t
.
v
7
i
1
.
pp
46
-
5
5
46
Jou
r
n
al
h
o
m
e
pa
ge
:
ht
t
p:
/
/
i
ae
s
pr
i
m
e
.
c
om
/
i
nd
e
x
.
php
/
c
s
i
t
C
l
o
u
d
-
b
a
s
e
d
p
r
e
d
i
c
t
i
v
e
a
n
a
l
y
t
i
c
s
f
o
r
p
e
n
s
i
o
n
f
u
n
d
p
e
r
f
o
r
ma
n
c
e
o
p
t
i
m
i
z
a
t
i
o
n
Be
au
ty
G
a
r
ab
a
,
M
ai
n
f
or
d
M
u
tan
d
av
ar
i
,
Je
r
i
t
a
C
h
i
b
h
ab
h
a
S
c
h
o
o
l
o
f
In
fo
rm
a
t
i
o
n
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
H
a
ra
re
In
s
t
i
t
u
t
e
o
f
T
e
c
h
n
o
l
o
g
y
(H
IT
),
H
a
ra
re
,
Z
i
m
b
a
b
w
e
A
r
ti
c
l
e
I
n
fo
A
BS
TR
A
C
T
Ar
t
i
c
l
e
h
i
s
t
or
y
:
Re
c
e
i
v
e
d
J
un
9
,
2025
Re
vi
s
e
d
J
u
l
9
,
2025
A
c
c
e
pt
e
d
J
u
l
17
,
2025
T
hi
s
s
t
ud
y
i
n
t
r
o
duc
e
s
a
nov
e
l
,
c
l
o
ud
-
ba
s
e
d
pr
e
d
i
c
t
i
v
e
a
na
l
y
t
i
c
s
f
r
a
m
e
w
or
k
t
a
i
l
o
r
e
d
f
o
r
pe
ns
i
on
f
un
d
pe
r
f
o
r
m
a
n
c
e
m
a
na
g
e
m
e
n
t
i
n
Z
i
m
b
a
bw
e
.
A
ddr
e
s
s
i
ng
l
i
m
i
t
a
t
i
on
s
i
n
t
r
a
di
t
i
ona
l
a
c
t
ua
r
i
a
l
m
o
de
l
s
,
t
he
p
r
op
os
e
d
s
ys
t
e
m
l
e
v
e
r
a
ge
s
r
e
a
l
-
t
i
m
e
d
a
t
a
pi
p
e
l
i
n
e
s
a
n
d
e
xp
l
a
i
na
bl
e
a
r
t
i
f
i
c
i
a
l
i
nt
e
l
l
i
ge
n
c
e
(
X
A
I
)
t
e
c
h
ni
qu
e
s
t
o
e
nha
n
c
e
f
o
r
e
c
a
s
t
i
n
g
a
c
c
u
r
a
c
y
a
n
d
t
r
a
n
s
p
a
r
e
nc
y
.
U
s
i
ng
r
e
gr
e
s
s
i
on
,
c
l
a
s
s
i
f
i
c
a
t
i
on
,
a
nd
de
e
p
l
e
a
r
n
i
ng
m
ode
l
s
,
i
t
f
o
r
e
c
a
s
t
s
m
e
m
be
r
c
on
t
r
i
b
ut
i
on
s
,
i
de
n
t
i
f
i
e
s
r
i
s
ks
o
f
c
on
t
r
i
b
ut
i
on
d
r
op
s
,
a
nd
pr
e
di
c
t
s
m
e
m
be
r
c
h
u
r
n
.
T
he
s
ys
t
e
m
’
s
c
l
oud
de
pl
oy
m
e
n
t
e
n
s
ur
e
s
s
c
a
l
a
b
i
l
i
t
y
a
nd
i
nt
e
r
a
c
t
i
v
e
i
n
t
e
g
r
a
t
i
on
w
i
t
h
t
oo
l
s
l
i
k
e
P
ow
e
r
B
I
f
o
r
de
c
i
s
i
o
n
s
upp
or
t
.
T
hi
s
s
ol
ut
i
on
s
i
g
ni
f
i
c
a
n
t
l
y
a
dva
n
c
e
s
s
us
t
a
i
na
b
l
e
p
e
ns
i
on
f
und
m
a
n
a
ge
m
e
nt
f
or
e
m
e
r
g
i
ng
e
c
ono
m
i
e
s
.
Ke
y
w
or
d
s
:
Cl
oud
c
o
m
pu
t
i
n
g
D
a
t
a
-
dri
v
e
n
de
c
i
s
i
on
m
a
ki
ng
D
e
m
ogr
a
ph
i
c
a
na
l
ys
i
s
N
e
ura
l
n
e
t
w
o
rks
P
e
ns
i
on
fund
m
a
n
a
ge
m
e
nt
P
re
di
c
t
i
ve
a
n
a
l
ys
i
s
Re
gr
e
s
s
i
on
a
n
a
l
ys
i
s
T
hi
s
i
s
an
op
e
n
ac
c
e
s
s
ar
t
i
c
l
e
u
nde
r
t
he
C
C
B
Y
-
SA
l
i
c
e
n
s
e
.
Cor
r
e
s
pon
di
n
g
Au
t
h
or
:
Be
a
u
t
y
G
a
r
a
ba
S
c
hool
of
Infor
m
a
t
i
on
S
c
i
e
nc
e
a
nd
T
e
c
hnol
ogy
,
H
a
ra
r
e
I
ns
t
i
t
ut
e
of
T
e
c
hn
ol
og
y
H
a
ra
r
e
,
Z
i
m
ba
bw
e
e
m
a
i
l
:
h23078
7a
@
hi
t
.
a
c
.
z
w
1.
I
N
TR
O
D
U
C
TI
O
N
P
e
ns
i
on
funds
a
re
a
vi
t
a
l
c
o
m
pon
e
n
t
i
n
gu
a
ra
n
t
e
e
i
ng
r
e
t
i
r
e
e
s
’
l
ong
-
t
e
r
m
f
i
n
a
nc
i
a
l
s
t
a
bi
l
i
t
y.
T
h
e
e
ffe
c
t
i
ve
a
d
m
i
n
i
s
t
r
a
t
i
on
a
nd
p
e
rfor
m
a
n
c
e
opt
i
m
i
z
a
t
i
on
o
f
t
h
e
s
e
fun
ds
a
r
e
now
m
or
e
c
r
uc
i
a
l
t
h
a
n
e
ve
r
due
t
o
t
he
a
gi
n
g
of
t
h
e
pop
ul
a
t
i
on
a
n
d
c
h
a
ngi
n
g
s
oc
i
oe
c
ono
m
i
c
c
i
r
c
um
s
t
a
n
c
e
s
.
K
l
e
rk
hi
g
hl
i
ght
s
t
hi
s
i
n
de
c
i
s
i
o
n
m
a
k
i
ng
und
e
r
unc
e
r
t
a
i
nt
y
a
nd
ri
s
k
[1]
.
F
u
nd
m
a
na
g
e
rs
a
r
e
i
nc
re
a
s
i
ng
l
y
t
a
s
ke
d
w
i
t
h
for
e
c
a
s
t
i
n
g
c
on
t
ri
bu
t
i
o
n
t
re
nds
,
c
ont
ro
l
l
i
ng
c
a
s
h
f
l
ow
,
a
nd
m
a
i
n
t
a
i
ni
n
g
fund
s
us
t
a
i
n
a
bi
l
i
t
y
i
n
t
he
fa
c
e
o
f
s
hi
f
t
i
ng
de
m
ogr
a
ph
i
c
s
a
nd
vol
a
t
i
l
e
m
a
rk
e
t
s
[
2]
.
T
ra
d
i
t
i
ona
l
fore
c
a
s
t
i
ng
a
n
d
p
e
rfor
m
a
nc
e
a
s
s
e
s
s
m
e
nt
m
e
t
ho
ds
oft
e
n
r
e
l
y
h
e
a
v
i
l
y
on
hi
s
t
ori
c
a
l
d
a
t
a
a
nd
s
t
a
t
i
c
fi
n
a
nc
i
a
l
m
od
e
l
s
.
A
l
t
h
ough
t
h
e
s
e
t
e
c
hni
que
s
p
rovi
d
e
foun
da
t
i
on
a
l
know
l
e
dg
e
,
t
h
e
y
o
ft
e
n
l
a
c
k
t
h
e
c
om
p
l
e
xi
t
y
n
e
c
e
s
s
a
ry
t
o
i
d
e
n
t
i
fy
t
h
e
dyn
a
m
i
c
,
n
on
-
l
i
ne
a
r
p
a
t
t
e
rns
found
i
n
a
c
t
ua
l
f
i
na
nc
i
a
l
e
c
os
ys
t
e
m
s
[3]
.
M
a
na
g
e
rs
of
pe
ns
i
on
fu
nds
t
hus
fi
nd
i
t
c
h
a
l
l
e
n
gi
ng
t
o
m
a
ke
p
rom
pt
,
d
a
t
a
-
dr
i
ve
n
de
c
i
s
i
ons
t
ha
t
s
t
r
i
ke
a
b
a
l
a
nc
e
be
t
w
e
e
n
ri
s
k
a
nd
r
e
t
urn
[4],
[5]
.
T
he
a
d
ve
n
t
of
m
a
c
hi
ne
l
e
a
rn
i
ng
(M
L
)
a
nd
pre
di
c
t
i
ve
a
na
l
y
t
i
c
s
[6]
pr
e
s
e
n
t
s
a
t
ra
ns
f
orm
a
t
i
ve
opport
u
ni
t
y
t
o
i
m
pro
ve
pe
ns
i
on
fu
nd
p
e
rfor
m
a
nc
e
m
a
n
a
ge
m
e
n
t
,
a
s
h
i
gh
l
i
gh
t
e
d
by
t
h
e
O
E
CD
i
n
p
e
ns
i
ons
m
a
rk
e
t
i
n
f
oc
us
202
0
[
7]
.
T
hrough
s
oph
i
s
t
i
c
a
t
e
d
m
od
e
l
i
ng
of
hi
s
t
or
i
c
a
l
,
d
e
m
ogr
a
phi
c
,
a
n
d
e
c
ono
m
i
c
da
t
a
,
t
he
s
e
t
e
c
hnol
ogi
e
s
p
rovi
d
e
d
e
e
p
e
r
i
ns
i
ght
s
by
e
xpos
i
n
g
pa
t
t
e
rns
a
n
d
t
re
nds
t
ha
t
w
o
ul
d
o
t
he
rw
i
s
e
go
un
de
t
e
c
t
e
d
[8]
.
By
a
dopt
i
ng
s
uc
h
t
e
c
hni
que
s
,
p
e
ns
i
on
f
und
a
d
m
i
ni
s
t
r
a
t
o
rs
c
a
n
e
nh
a
n
c
e
for
e
c
a
s
t
i
n
g
a
c
c
ur
a
c
y
,
pro
a
c
t
i
v
e
l
y
m
a
n
a
ge
r
i
s
ks
,
a
nd
a
c
h
i
e
v
e
m
or
e
f
a
vor
a
b
l
e
i
nv
e
s
t
m
e
n
t
out
c
o
m
e
s
[4],
[9]
.
T
hi
s
s
t
udy
pr
e
s
e
n
t
s
a
c
l
o
ud
-
b
a
s
e
d
pre
di
c
t
i
ve
a
n
a
l
y
t
i
c
s
[
1
0]
m
ode
l
s
p
e
c
i
fi
c
a
l
l
y
d
e
s
i
gn
e
d
for
t
h
e
Z
i
m
ba
bw
e
a
n
p
e
ns
i
o
n
fu
nd
e
nv
i
ron
m
e
n
t
.
T
o
ove
r
c
o
m
e
for
e
c
a
s
t
i
ng
c
ons
t
ra
i
nt
s
i
n
c
onve
nt
i
ona
l
s
ys
t
e
m
s
,
t
h
e
m
ode
l
u
t
i
l
i
s
e
s
c
o
nt
e
m
po
ra
ry
ML
a
l
gor
i
t
h
m
s
,
e
n
a
bl
i
ng
fu
nd
m
a
n
a
ge
rs
t
o
m
a
ke
pro
a
c
t
i
v
e
,
s
t
r
a
t
e
gi
c
,
a
n
d
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
Cl
oud
-
b
as
e
d
pr
e
d
i
c
t
i
v
e
ana
l
y
t
i
c
s
f
or
p
e
ns
i
on
f
u
nd
pe
r
f
or
m
an
c
e
opt
i
m
i
z
a
t
i
on
(
B
e
au
t
y
G
ar
aba
)
47
i
nfor
m
e
d
d
e
c
i
s
i
ons
[4]
.
In
doi
n
g
s
o
,
t
h
i
s
s
t
u
dy
a
i
m
s
t
o
i
m
p
rov
e
pe
ns
i
on
fund
m
a
n
a
g
e
m
e
nt
t
e
c
hni
q
ue
s
a
nd
provi
d
e
a
s
c
a
l
a
b
l
e
fr
a
m
e
w
ork
for
l
on
g
-
t
e
rm
f
i
na
n
c
i
a
l
pl
a
nn
i
n
g
i
n
c
om
p
a
r
a
bl
e
e
m
e
rg
i
ng
e
c
ono
m
i
e
s
.
A
s
hi
ghl
i
gh
t
e
d
i
n
t
h
e
s
urv
e
y
of
b
i
g
d
a
t
a
b
y
Ch
e
n
e
t
al
.
[11]
r
e
ga
rdi
ng
pr
e
d
i
c
t
i
ons
on
t
he
e
vo
l
vi
ng
r
ol
e
of
bi
g
d
a
t
a
i
n
s
h
a
pi
n
g
t
he
fut
u
re
e
c
ono
m
i
c
l
a
nds
c
a
p
e
,
a
s
w
e
l
l
a
s
‘A
pe
rf
orm
a
n
c
e
e
v
a
l
u
a
t
i
on
of
c
l
a
s
s
i
fi
c
a
t
i
on
a
l
gor
i
t
h
m
s
for
b
i
g
da
t
a
’
b
y
H
a
i
e
t
al
.
[12]
.
E
xi
s
t
i
ng
p
e
ns
i
on
f
ore
c
a
s
t
i
ng
s
ys
t
e
m
s
i
n
Z
i
m
ba
bw
e
a
n
d
c
o
m
pa
ra
b
l
e
e
m
e
rgi
ng
e
c
on
om
i
e
s
a
re
l
i
m
i
t
e
d
b
y
s
t
a
t
i
c
m
od
e
l
s
t
h
a
t
c
a
nnot
c
a
pt
ur
e
t
he
c
o
m
p
l
e
x
i
t
y
of
m
od
e
rn
de
m
ogr
a
ph
i
c
a
nd
fi
na
n
c
i
a
l
d
yna
m
i
c
s
.
T
hi
s
s
t
udy
fi
l
l
s
t
ha
t
ga
p
by
i
nt
rod
uc
i
n
g
a
c
l
oud
-
b
a
s
e
d
,
ML
-
dri
v
e
n
s
o
l
ut
i
on
t
ha
t
e
n
a
bl
e
s
re
a
l
-
t
i
m
e
,
a
c
c
ura
t
e
,
a
nd
i
n
t
e
r
pre
t
a
b
l
e
pe
ns
i
on
fund
m
a
n
a
ge
m
e
nt
[12]
,
[13]
.
P
ro
bl
e
m
s
t
a
t
e
m
e
nt
:
p
e
ns
i
on
fun
ds
i
n
Z
i
m
b
a
bw
e
fa
c
e
i
nc
r
e
a
s
i
ng
c
ha
l
l
e
nge
s
d
ue
t
o
de
m
ogr
a
phi
c
s
hi
f
t
s
,
e
c
on
om
i
c
i
ns
t
a
bi
l
i
t
y,
a
nd
l
i
m
i
t
e
d
fore
c
a
s
t
i
ng
t
oo
l
s
.
T
ra
d
i
t
i
on
a
l
a
c
t
u
a
ri
a
l
m
od
e
l
s
oft
e
n
l
a
c
k
a
da
p
t
a
b
i
l
i
t
y
a
n
d
f
a
i
l
t
o
pro
vi
d
e
re
a
l
-
t
i
m
e
,
i
n
t
e
rpr
e
t
a
b
l
e
i
ns
i
gh
t
s
n
e
e
d
e
d
for
pro
a
c
t
i
v
e
d
e
c
i
s
i
on
-
m
a
ki
n
g.
Re
s
e
a
rc
h
ob
j
e
c
t
i
ve
s
:
i
)
t
o
de
v
e
l
op
ML
m
od
e
l
s
fo
r
for
e
c
a
s
t
i
n
g
c
on
t
ri
but
i
ons
,
drop
-
ri
s
k
,
a
nd
m
e
m
b
e
r
c
hurn
i
n
p
e
ns
i
o
n
fun
ds
,
i
i
)
t
o
i
nt
e
gr
a
t
e
a
c
t
u
a
ri
a
l
,
d
e
m
ogra
phi
c
,
a
nd
t
r
a
ns
a
c
t
i
on
a
l
da
t
a
i
nt
o
a
re
a
l
-
t
i
m
e
p
re
d
i
c
t
i
v
e
pi
pe
l
i
n
e
,
a
nd
i
i
i
)
t
o
d
e
pl
oy
m
od
e
l
s
on
a
c
l
oud
p
l
a
t
form
t
o
e
n
a
bl
e
i
n
t
e
r
a
c
t
i
v
e
de
c
i
s
i
on
da
s
h
boa
rds
fo
r
pe
ns
i
on
a
dm
i
ni
s
t
ra
t
ors
.
K
e
y
c
on
t
ri
bu
t
i
o
ns
:
i
)
i
nt
rod
uc
e
d
a
m
odul
a
r,
c
l
o
ud
-
ba
s
e
d
pr
e
d
i
c
t
i
v
e
a
na
l
yt
i
c
s
fr
a
m
e
w
ork
for
pe
ns
i
on
fund
pe
rfor
m
a
n
c
e
for
e
c
a
s
t
i
n
g
i
n
Z
i
m
ba
bw
e
,
i
i
)
e
ngi
n
e
e
r
e
d
d
om
a
i
n
-
s
p
e
c
i
fi
c
M
L
m
od
e
l
s
fo
r
for
e
c
a
s
t
i
ng
c
ont
r
i
bu
t
i
ons
,
drop
-
r
i
s
k,
a
nd
c
h
urn
us
i
ng
a
c
t
ua
r
i
a
l
a
nd
t
ra
ns
a
c
t
i
on
a
l
d
a
t
a
,
i
i
i
)
a
c
h
i
e
v
e
d
hi
g
h
-
pe
rfor
m
a
n
c
e
m
e
t
ri
c
s
(e
.
g
.
,
99.
86%
c
hurn
a
c
c
ur
a
c
y,
0.
8
5
R²
)
w
hi
l
e
m
a
i
nt
a
i
ni
ng
t
r
a
ns
pa
re
n
c
y
v
i
a
e
xp
l
a
i
na
bl
e
fe
a
t
u
re
s
,
i
v
)
de
p
l
oye
d
a
RE
S
T
ful
A
P
I
vi
a
F
a
s
t
A
P
I,
e
na
b
l
i
ng
r
e
a
l
-
t
i
m
e
P
ow
e
r
BI
d
a
s
hbo
a
rds
f
or
fund
m
a
n
a
ge
rs
,
a
nd
v)
br
i
dg
e
d
a
c
t
u
a
ri
a
l
t
he
ory
w
i
t
h
M
L
[14
]
t
o
s
uppo
rt
i
n
t
e
l
l
i
g
e
n
t
,
e
t
h
i
c
a
l
d
e
c
i
s
i
on
-
m
a
k
i
ng
i
n
a
r
e
s
ourc
e
-
c
ons
t
r
a
i
n
e
d
e
nv
i
ron
m
e
nt
.
T
hi
s
r
e
s
e
a
rc
h
i
s
und
e
rpi
nne
d
by
a
m
u
l
t
i
di
s
c
i
p
l
i
n
a
ry
t
he
or
e
t
i
c
a
l
found
a
t
i
on
t
ha
t
i
n
t
e
g
ra
t
e
s
pr
i
nc
i
pl
e
s
from
m
o
de
rn
por
t
fo
l
i
o
t
h
e
ory
(M
P
T
)
,
e
ff
i
c
i
e
n
t
m
a
rke
t
hypo
t
he
s
i
s
(
E
M
H
)
,
a
rt
i
fi
c
i
a
l
i
n
t
e
l
l
i
ge
n
c
e
(A
I)
de
c
i
s
i
on
t
he
or
y
[1
5]
,
a
nd
a
c
t
u
a
ri
a
l
t
h
e
ory
.
T
h
e
s
e
fra
m
e
w
o
rks
c
ol
l
e
c
t
i
ve
l
y
s
ha
pe
t
h
e
c
on
c
e
p
t
ua
l
de
s
i
gn
a
nd
a
n
a
l
yt
i
c
a
l
di
re
c
t
i
on
o
f
t
h
e
p
re
d
i
c
t
i
v
e
m
ode
l
d
e
v
e
l
op
e
d
f
or
e
nha
nc
e
d
pe
ns
i
on
fund
pe
r
form
a
n
c
e
m
a
na
g
e
m
e
n
t
.
H
a
rry
M
a
rkow
i
t
z
'
s
19
52
i
n
t
rodu
c
t
i
on
of
M
P
T
[16]
,
a
found
a
t
i
o
na
l
i
d
e
a
i
n
f
i
n
a
nc
e
,
c
om
p
l
e
t
e
l
y
c
h
a
ng
e
d
how
i
nve
s
t
m
e
nt
por
t
fol
i
os
a
re
pu
t
t
og
e
t
h
e
r.
T
o
r
e
du
c
e
i
n
ve
s
t
m
e
nt
ri
s
k
for
a
g
i
ve
n
l
e
v
e
l
of
re
t
urn
,
t
he
M
P
T
pl
a
c
e
s
a
s
t
rong
e
m
ph
a
s
i
s
on
di
v
e
rs
i
f
i
c
a
t
i
on.
T
h
e
M
P
T
e
m
pha
s
i
z
e
s
d
i
ve
rs
i
fi
c
a
t
i
on
t
o
m
i
n
i
m
i
z
e
i
n
ve
s
t
m
e
nt
r
i
s
k
for
a
gi
ve
n
l
e
ve
l
of
re
t
urn
.
In
t
he
c
on
t
e
x
t
of
p
e
ns
i
on
f
unds
,
M
P
T
s
upport
s
op
t
i
m
a
l
a
s
s
e
t
a
l
l
o
c
a
t
i
on
s
t
ra
t
e
g
i
e
s
a
c
ros
s
a
va
ri
e
t
y
o
f
i
ns
t
rum
e
n
t
s
.
H
ow
e
ve
r
,
i
t
s
s
t
a
t
i
c
a
s
s
um
p
t
i
o
ns
—
s
uc
h
a
s
s
t
a
b
l
e
c
orr
e
l
a
t
i
ons
a
nd
r
a
t
i
ona
l
i
nve
s
t
or
be
ha
vi
or
—
l
i
m
i
t
i
t
s
e
ff
e
c
t
i
v
e
ne
s
s
i
n
t
od
a
y’s
vo
l
a
t
i
l
e
fi
na
n
c
i
a
l
e
nv
i
ron
m
e
n
t
.
T
hi
s
s
t
udy
e
nha
n
c
e
s
M
P
T
by
i
nt
e
gra
t
i
n
g
a
da
p
t
i
v
e
,
A
I
-
dri
v
e
n
a
ppr
oa
c
he
s
t
h
a
t
re
s
p
ond
t
o
re
a
l
-
t
i
m
e
m
a
rk
e
t
dyn
a
m
i
c
s
[15]
,
[
17]
.
T
he
e
c
ono
m
i
s
t
E
uge
ne
F
a
m
a
de
ve
l
ope
d
t
he
E
M
H
i
n
t
h
e
197
0s
[18]
,
w
h
i
c
h
c
ont
e
nds
t
ha
t
a
s
s
e
t
pr
i
c
e
s
i
n
fi
na
n
c
i
a
l
m
a
rke
t
s
a
c
c
ura
t
e
l
y
r
e
pr
e
s
e
nt
a
l
l
a
v
a
i
l
a
b
l
e
i
nfor
m
a
t
i
on
a
t
a
ny
gi
v
e
n
m
om
e
n
t
.
P
r
e
d
i
c
t
i
v
e
m
ode
l
i
n
g
i
s
t
he
or
e
t
i
c
a
l
l
y
c
h
a
l
l
e
ng
e
d
by
t
he
E
M
H
,
w
hi
c
h
c
o
nt
e
nds
t
h
a
t
a
s
s
e
t
pr
i
c
e
s
a
l
r
e
a
d
y
t
a
k
e
i
n
t
o
a
c
c
o
unt
a
l
l
a
va
i
l
a
bl
e
i
nfor
m
a
t
i
on
.
H
ow
e
ve
r
,
e
m
pi
ri
c
a
l
da
t
a
d
e
m
o
ns
t
ra
t
e
t
ha
t
b
e
h
a
vi
or
a
l
b
i
a
s
e
s
,
g
e
opo
l
i
t
i
c
a
l
e
v
e
nt
s
,
a
nd
d
e
l
a
y
e
d
re
a
c
t
i
ons
t
o
n
e
w
i
nfor
m
a
t
i
on
fre
que
n
t
l
y
r
e
s
ul
t
i
n
i
n
e
ff
i
c
i
e
nt
m
a
rk
e
t
s
.
E
s
pe
c
i
a
l
l
y
ov
e
r
l
ong
e
r
i
nv
e
s
t
m
e
nt
hori
z
ons
l
i
k
e
t
hos
e
i
n
pe
ns
i
on
f
und
s
t
ra
t
e
g
i
e
s
,
t
hi
s
c
r
e
a
t
e
s
r
oom
for
ML
m
od
e
l
s
t
o
f
i
nd
hi
d
de
n
pa
t
t
e
rns
a
nd
produc
e
us
e
fu
l
i
ns
i
gh
t
s
.
A
fra
m
e
w
o
rk
f
or
a
ut
om
a
t
e
d,
d
a
t
a
-
dr
i
ve
n
d
e
c
i
s
i
on
-
m
a
ki
ng
i
n
t
he
f
a
c
e
of
unc
e
rt
a
i
n
t
y
i
s
prov
i
de
d
by
A
I
de
c
i
s
i
on
t
h
e
ory
.
I
t
m
a
k
e
s
i
t
p
os
s
i
bl
e
t
o
c
re
a
t
e
s
ys
t
e
m
s
t
h
a
t
,
i
n
a
dd
i
t
i
on
t
o
pre
d
i
c
t
i
n
g
r
e
s
ul
t
s
,
s
ugg
e
s
t
t
h
e
b
e
s
t
c
ours
e
of
a
c
t
i
on
b
a
s
e
d
on
r
e
i
nf
orc
e
m
e
nt
l
e
a
rni
ng
a
nd
pr
oba
b
i
l
i
s
t
i
c
r
e
a
s
on
i
ng
.
T
h
i
s
i
s
i
n
l
i
ne
w
i
t
h
t
h
e
i
n
t
ri
c
a
t
e
ne
e
ds
of
m
a
na
g
i
ng
p
e
ns
i
o
n
funds
,
w
he
r
e
c
hoi
c
e
s
m
us
t
t
a
k
e
c
h
a
ng
i
ng
d
e
m
o
gra
p
hi
c
pa
t
t
e
rns
,
u
ns
t
a
b
l
e
e
c
on
om
i
e
s
,
a
nd
re
gu
l
a
t
ory
c
ha
ng
e
s
i
nt
o
c
ons
i
d
e
ra
t
i
o
n.
A
c
t
u
a
ri
a
l
t
h
e
ory
,
l
ong
e
s
t
a
bl
i
s
he
d
i
n
p
e
ns
i
on
for
e
c
a
s
t
i
ng,
pr
ovi
de
s
s
t
a
t
i
s
t
i
c
a
l
m
ode
l
s
for
e
s
t
i
m
a
t
i
ng
l
i
a
bi
l
i
t
i
e
s
a
nd
f
ut
ur
e
fu
nd
ob
l
i
g
a
t
i
ons
.
D
e
s
p
i
t
e
t
h
e
i
r
robus
t
n
e
s
s
,
t
r
a
d
i
t
i
ona
l
a
c
t
ua
r
i
a
l
m
e
t
hods
a
re
not
a
l
w
a
ys
re
s
pons
i
v
e
t
o
a
brup
t
c
h
a
ng
e
s
a
nd
fr
e
qu
e
nt
l
y
r
e
l
y
on
f
i
xe
d
a
s
s
um
pt
i
ons
.
T
o
i
nc
re
a
s
e
fo
re
c
a
s
t
i
ng
g
ra
nu
l
a
ri
t
y,
a
da
p
t
i
v
i
t
y
,
a
nd
r
e
a
l
-
t
i
m
e
r
e
s
pons
i
ve
n
e
s
s
,
t
h
i
s
s
t
ud
y
s
ug
ge
s
t
s
a
hy
bri
d
a
ppro
a
c
h
t
h
a
t
c
om
b
i
n
e
s
ML
t
e
c
hni
q
ue
s
w
i
t
h
a
c
t
u
a
ri
a
l
m
o
de
l
s
.
T
he
us
e
fu
l
ne
s
s
of
ML
a
nd
pre
d
i
c
t
i
v
e
a
na
l
yt
i
c
s
i
n
p
e
ns
i
on
fun
ds
a
nd
f
i
na
nc
i
a
l
pe
rfor
m
a
n
c
e
fore
c
a
s
t
i
ng
i
s
e
m
pi
ri
c
a
l
l
y
s
uppor
t
e
d
by
a
n
e
x
pa
nd
i
ng
body
of
re
s
e
a
r
c
h.
R
e
s
e
a
rc
h
c
ont
i
nuous
l
y
de
m
ons
t
r
a
t
e
s
t
ha
t
M
L
m
ode
l
s
—
s
uc
h
a
s
gr
a
di
e
nt
boos
t
i
ng
m
a
c
h
i
ne
s
,
l
ong
s
hort
-
t
e
r
m
m
e
m
ory
(L
S
T
M
)
n
e
t
w
or
ks
,
a
nd
re
i
nfor
c
e
m
e
nt
l
e
a
rni
n
g
—
p
e
rfor
m
m
o
re
a
c
c
ura
t
e
l
y
t
ha
n
c
onv
e
nt
i
on
a
l
s
t
a
t
i
s
t
i
c
a
l
t
e
c
hn
i
que
s
,
pa
rt
i
c
u
l
a
r
l
y
w
he
n
de
a
l
i
ng
w
i
t
h
c
om
p
l
e
x
d
a
t
a
a
nd
un
c
e
rt
a
i
n
m
a
rk
e
t
s
.
T
o
i
m
prov
e
m
o
de
l
p
e
rf
orm
a
nc
e
,
r
e
s
e
a
r
c
h
a
l
s
o
e
m
pha
s
i
z
e
s
how
c
ruc
i
a
l
i
t
i
s
t
o
i
n
c
orpor
a
t
e
d
e
m
ogra
phi
c
,
m
a
c
ro
e
c
ono
m
i
c
,
a
nd
a
l
t
e
rna
t
i
ve
d
a
t
a
(
l
i
k
e
s
e
n
t
i
m
e
nt
a
na
l
ys
i
s
a
nd
E
S
G
m
e
t
ri
c
s
).
N
one
t
h
e
l
e
s
s
,
t
he
r
e
a
re
s
t
i
l
l
s
e
v
e
ra
l
g
a
ps
i
n
l
i
t
e
ra
t
ure
.
T
h
e
s
e
i
nc
l
ude
i
s
s
ue
s
w
i
t
h
c
o
m
pu
t
a
t
i
o
na
l
de
m
a
nds
,
m
ode
l
i
nt
e
rpr
e
t
a
bi
l
i
t
y,
a
n
d
e
t
h
i
c
a
l
a
nd
l
e
ga
l
c
om
p
l
i
a
nc
e
.
F
or
e
xa
m
p
l
e
,
e
ve
n
t
h
ough
de
e
p
l
e
a
rn
i
ng
m
ode
l
s
a
r
e
ve
ry
a
c
c
u
ra
t
e
[19
],
[2
0]
,
t
he
i
r
"
bl
a
c
k
box"
n
a
t
ur
e
c
a
n
m
a
k
e
t
h
e
m
l
e
s
s
t
ra
ns
p
a
re
nt
,
w
hi
c
h
i
s
c
r
uc
i
a
l
for
re
g
ul
a
t
ors
a
nd
s
t
a
ke
hol
d
e
rs
i
n
pe
ns
i
on
fu
nds
.
S
i
m
i
l
a
r
t
o
t
hi
s
,
a
dva
nc
e
d
m
ode
l
s
'
hi
gh
r
e
s
our
c
e
r
e
qu
i
re
m
e
nt
s
pre
ve
nt
t
h
e
m
fro
m
be
i
ng
us
e
d
i
n
s
e
t
t
i
ngs
w
i
t
h
fe
w
e
r
re
s
our
c
e
s
or
i
n
s
m
a
l
l
e
r
pe
ns
i
on
s
c
he
m
e
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
46
-
5
5
48
T
hi
s
r
e
s
e
a
r
c
h
a
ddr
e
s
s
e
s
t
he
s
e
ga
ps
b
y
propos
i
ng
a
c
l
o
ud
-
ba
s
e
d
p
re
d
i
c
t
i
v
e
a
n
a
l
y
t
i
c
s
fra
m
e
w
o
rk
t
a
i
l
or
e
d
s
pe
c
i
f
i
c
a
l
l
y
t
o
pe
ns
i
on
fun
d
m
a
na
g
e
m
e
n
t
.
T
he
m
od
e
l
i
n
t
e
g
ra
t
e
s
re
a
l
-
t
i
m
e
d
a
t
a
pi
p
e
l
i
ne
s
,
e
xp
l
a
i
na
bl
e
a
rt
i
fi
c
i
a
l
i
n
t
e
l
l
i
ge
nc
e
(X
A
I)
t
e
c
h
ni
qu
e
s
[1
5]
,
[17]
,
a
nd
s
c
a
l
a
b
l
e
i
nfra
s
t
ruc
t
ur
e
t
o
e
ns
ure
pra
c
t
i
c
a
l
a
ppl
i
c
a
bi
l
i
t
y
a
c
ros
s
di
ve
rs
e
i
ns
t
i
t
u
t
i
on
a
l
s
e
t
t
i
ngs
.
By
bri
dgi
ng
t
he
ore
t
i
c
a
l
robus
t
n
e
s
s
w
i
t
h
e
m
pi
ri
c
a
l
i
nnov
a
t
i
on
,
t
h
i
s
s
t
udy
a
i
m
s
t
o
d
e
l
i
ve
r
a
pr
e
di
c
t
i
ve
s
ol
u
t
i
on
t
ha
t
i
s
not
on
l
y
a
c
c
ur
a
t
e
a
nd
a
d
a
pt
i
ve
bu
t
a
l
s
o
t
ra
ns
pa
r
e
nt
,
e
t
hi
c
a
l
,
a
nd
re
gul
a
t
ory
-
c
om
p
l
i
a
nt
.
2.
M
ET
H
O
D
T
hi
s
s
t
udy
pr
e
s
e
nt
s
t
h
e
c
l
oud
-
b
a
s
e
d
pr
e
di
c
t
i
ve
a
na
l
yt
i
c
s
f
or
pe
ns
i
on
fun
d
p
e
rfor
m
a
nc
e
o
pt
i
m
i
z
a
t
i
on;
a
s
t
ruc
t
ur
e
d
ML
fr
a
m
e
w
ork
d
e
s
i
gn
e
d
for
p
e
ns
i
on
f
und
p
e
rfor
m
a
n
c
e
fo
re
c
a
s
t
i
ng
.
It
a
dopt
s
a
d
a
t
a
-
dr
i
ve
n
a
ppro
a
c
h
t
o
d
e
ve
l
op
i
ng
a
nd
e
va
l
ua
t
i
ng
M
L
m
od
e
l
s
f
or
pe
ns
i
on
fund
pe
rfor
m
a
n
c
e
op
t
i
m
i
z
a
t
i
on
.
T
h
e
m
e
t
hodo
l
ogy
i
nvol
ve
s
a
s
t
ruc
t
ur
e
d
pi
pe
l
i
n
e
t
h
a
t
i
nc
l
ude
s
d
a
t
a
i
n
t
e
gr
a
t
i
on
,
fe
a
t
ur
e
e
ngi
n
e
e
ri
ng
,
m
ode
l
t
r
a
i
n
i
ng
a
nd
e
v
a
l
u
a
t
i
on
,
a
nd
m
ode
l
de
p
l
oy
m
e
nt
.
T
hr
e
e
pr
e
di
c
t
i
ve
m
ode
l
s
w
e
r
e
de
v
e
l
ope
d
:
i
)
a
r
e
gr
e
s
s
i
on
m
o
de
l
f
or
t
ot
a
l
c
ont
ri
bu
t
i
o
n
for
e
c
a
s
t
i
ng,
ii
)
a
bi
n
a
ry
c
l
a
s
s
i
fi
c
a
t
i
on
m
ode
l
fo
r
c
o
nt
r
i
but
i
on
d
rop
r
i
s
k
d
e
t
e
c
t
i
on
,
a
nd
iii
)
a
b
i
na
ry
c
l
a
s
s
i
f
i
c
a
t
i
on
m
od
e
l
fo
r
c
h
urn
pr
e
di
c
t
i
on
.
T
he
o
ve
ra
l
l
a
rc
hi
t
e
c
t
ur
e
i
s
i
l
l
us
t
ra
t
e
d
i
n
F
i
gur
e
1
,
w
hi
l
e
t
he
m
e
t
h
odol
o
gi
c
a
l
w
or
kfl
ow
i
s
de
p
i
c
t
e
d
i
n
F
i
gure
2
.
F
i
gure
1
.
S
ys
t
e
m
a
rc
h
i
t
e
c
t
ure
F
i
gure
2
.
M
e
t
ho
dol
og
i
c
a
l
w
orkf
l
ow
2.
1
.
P
r
op
os
e
d
s
ol
u
ti
on
ov
e
r
vi
e
w
T
he
prop
os
e
d
s
ol
u
t
i
on
i
s
a
m
odu
l
a
r
M
L
-
ba
s
e
d
d
e
c
i
s
i
on
s
upp
ort
s
ys
t
e
m
d
e
s
i
gn
e
d
t
o
e
nha
nc
e
s
t
ra
t
e
g
i
c
pe
ns
i
on
fund
m
a
na
g
e
m
e
n
t
t
hr
ough
p
re
d
i
c
t
i
v
e
a
na
l
yt
i
c
s
.
It
i
nt
e
gr
a
t
e
s
a
c
t
u
a
ri
a
l
a
nd
t
r
a
ns
a
c
t
i
ona
l
d
a
t
a
i
n
t
o
a
n
i
nt
e
l
l
i
ge
nt
pi
p
e
l
i
ne
c
a
p
a
bl
e
o
f
e
s
t
i
m
a
t
i
ng
fu
t
ure
c
ont
r
i
bu
t
i
ons
,
i
de
n
t
i
f
yi
ng
hi
gh
-
ri
s
k
c
ont
r
i
but
ors
,
a
nd
d
e
t
e
c
t
i
ng
pot
e
n
t
i
a
l
m
e
m
be
r
c
hurn
.
T
he
s
e
pr
e
d
i
c
t
i
ons
a
r
e
de
pl
oy
e
d
t
h
rough
a
R
E
S
T
fu
l
A
P
I
us
i
ng
F
a
s
t
A
P
I
,
e
n
a
bl
i
ng
s
e
a
m
l
e
s
s
i
n
t
e
gra
t
i
on
w
i
t
h
r
e
a
l
-
t
i
m
e
d
a
s
hboa
rds
s
u
c
h
a
s
M
i
c
ros
o
ft
P
ow
e
r
BI
for
i
nt
e
ra
c
t
i
ve
v
i
s
ua
l
i
z
a
t
i
on
a
nd
de
c
i
s
i
on
-
m
a
ki
ng
s
uppo
rt
.
2.
2
.
D
ata
c
o
l
l
e
c
t
i
on
an
d
i
n
t
e
gr
a
ti
on
D
a
t
a
w
a
s
obt
a
i
ne
d
f
rom
t
hre
e
E
xc
e
l
s
he
e
t
s
:
a
c
t
ua
ri
a
l
da
t
a
,
c
ont
ri
but
i
on
s
bre
a
kdo
w
n,
a
nd
e
xi
t
s
.
E
a
c
h
da
t
a
s
e
t
c
ont
a
i
ne
d
de
t
a
i
l
e
d
m
e
m
be
r
-
l
e
ve
l
pe
n
s
i
on
re
c
ords
,
i
nc
l
udi
ng
de
m
ogra
phi
c
a
t
t
ri
but
e
s
,
s
a
l
a
ry
hi
s
t
ory,
f
und
e
nt
ry
a
n
d
e
xi
t
da
t
e
s
,
a
nd
c
ont
ri
but
i
on
t
ra
ns
a
c
t
i
on
s
.
T
he
da
t
a
s
e
t
s
w
e
re
m
e
rge
d
u
s
i
n
g
uni
que
i
de
nt
i
fi
e
r
s
(e
.
g.
,
s
y
s
t
e
m
no
),
a
n
d
da
t
a
p
re
proc
e
s
s
i
ng
s
t
e
ps
e
ns
u
re
d
c
on
s
i
s
t
e
nc
y
i
n
da
t
e
f
orm
a
t
s
,
c
ol
um
n
na
m
i
ng,
a
nd
va
l
ue
re
p
re
s
e
nt
a
t
i
o
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
Cl
oud
-
b
as
e
d
pr
e
d
i
c
t
i
v
e
ana
l
y
t
i
c
s
f
or
p
e
ns
i
on
f
u
nd
pe
r
f
or
m
an
c
e
opt
i
m
i
z
a
t
i
on
(
B
e
au
t
y
G
ar
aba
)
49
2.
3
.
F
e
atu
r
e
e
n
gi
n
e
e
r
i
n
g
F
e
a
t
u
re
e
ng
i
ne
e
ri
n
g
w
a
s
c
ondu
c
t
e
d
t
o
e
nha
n
c
e
t
h
e
pr
e
di
c
t
i
v
e
c
a
pa
c
i
t
y
of
t
h
e
m
ode
l
s
:
a
ge
w
a
s
c
a
l
c
u
l
a
t
e
d
fro
m
da
t
e
of
bi
r
t
h
,
a
nd
fund
t
e
nur
e
w
a
s
de
r
i
ve
d
from
fund
e
nt
ry
d
a
t
e
s
.
T
o
t
a
l
c
on
t
ri
b
ut
i
ons
w
e
re
c
om
p
ut
e
d
a
s
t
he
s
u
m
of
e
m
p
l
oy
e
e
(
E
E
)
a
nd
e
m
pl
oy
e
r
(
E
R)
c
on
t
ri
b
ut
i
ons
.
D
rop
ri
s
k
w
a
s
e
ngi
ne
e
re
d
by
c
om
p
a
r
i
ng
t
he
a
ve
ra
g
e
m
ont
hl
y
c
on
t
ri
bu
t
i
o
ns
be
t
w
e
e
n
e
a
r
l
y
a
nd
l
a
t
e
r
m
on
t
hs
,
i
d
e
nt
i
fyi
ng
a
≥30%
de
c
l
i
ne
.
Churn
w
a
s
d
e
fi
n
e
d
a
s
a
pe
r
i
od
of
s
i
x
or
m
or
e
c
ons
e
c
ut
i
ve
m
on
t
hs
w
i
t
h
z
e
r
o
c
ont
ri
bu
t
i
o
ns
.
Ca
t
e
go
ri
c
a
l
v
a
ri
a
bl
e
s
(e
.
g
.
,
g
e
nd
e
r,
p
a
y
p
oi
n
t
)
w
e
re
e
nc
od
e
d
us
i
ng
on
e
-
ho
t
e
nc
odi
ng
,
a
nd
m
i
s
s
i
ng
va
l
ue
s
w
e
r
e
i
m
p
ut
e
d
us
i
ng
s
t
a
t
i
s
t
i
c
a
l
m
e
t
hods
ba
s
e
d
on
f
e
a
t
ur
e
t
yp
e
a
nd
di
s
t
r
i
bu
t
i
on
.
2.
3
.
1.
M
od
e
l
1:
tot
al
c
on
t
r
i
b
u
ti
on
p
r
e
d
i
c
ti
on
(r
e
gr
e
s
s
i
on
)
A
re
gr
e
s
s
i
on
prob
l
e
m
w
a
s
fo
rm
u
l
a
t
e
d
t
o
pr
e
di
c
t
t
ot
a
l
fut
ur
e
c
ont
ri
b
ut
i
ons
.
T
w
o
m
ode
l
s
w
e
re
t
r
a
i
n
e
d:
i
)
l
i
n
e
a
r
re
g
re
s
s
i
on
,
a
nd
i
i
)
r
a
ndo
m
fore
s
t
r
e
gre
s
s
or
.
T
h
e
da
t
a
s
e
t
w
a
s
p
a
rt
i
t
i
one
d
i
n
t
o
t
r
a
i
n
i
ng
a
nd
t
e
s
t
s
e
t
s
.
M
od
e
l
s
w
e
re
e
v
a
l
u
a
t
e
d
us
i
ng
m
e
a
n
a
bs
o
l
ut
e
e
rror
(M
A
E
),
r
oot
m
e
a
n
s
qu
a
r
e
d
e
r
ror
(RM
S
E
),
a
nd
t
h
e
R²
s
c
ore
.
T
h
e
ra
nd
om
for
e
s
t
m
ode
l
de
m
ons
t
r
a
t
e
d
s
upe
ri
or
p
e
r
form
a
nc
e
w
i
t
h
a
n
R²
of
0
.
85
,
i
nd
i
c
a
t
i
ng
s
t
ron
g
pre
di
c
t
i
ve
a
bi
l
i
t
y.
2.
3
.
2.
M
od
e
l
2:
c
on
tr
i
b
u
ti
on
d
r
op
r
i
s
k
c
l
as
s
i
fi
c
ati
on
T
hi
s
m
ode
l
i
de
n
t
i
f
i
e
s
m
e
m
be
rs
l
i
k
e
l
y
t
o
r
e
du
c
e
t
he
i
r
c
ont
r
i
b
ut
i
ons
.
D
rop
r
i
s
k
w
a
s
l
a
b
e
l
e
d
a
s
a
≥30
%
drop
i
n
a
v
e
ra
ge
c
ont
r
i
but
i
ons
.
a
ra
n
dom
for
e
s
t
c
l
a
s
s
i
f
i
e
r
w
a
s
e
m
p
l
oy
e
d,
us
i
ng
fe
a
t
ur
e
s
s
u
c
h
a
s
s
a
l
a
ry
,
ope
ni
ng
a
nd
c
l
os
i
ng
b
a
l
a
n
c
e
s
,
pa
y
po
i
nt
,
a
nd
t
e
nure
.
G
i
ve
n
t
h
e
c
l
a
s
s
i
m
ba
l
a
n
c
e
,
s
yn
t
he
t
i
c
m
i
n
ori
t
y
ov
e
rs
a
m
pl
i
ng
t
e
c
hni
q
ue
(S
M
O
T
E
)
w
a
s
us
e
d
t
o
b
a
l
a
nc
e
t
h
e
da
t
a
s
e
t
.
P
os
t
-
S
M
O
T
E
t
r
a
i
n
i
ng
yi
e
l
d
e
d
p
e
rfe
c
t
c
l
a
s
s
i
fi
c
a
t
i
on
m
e
t
ri
c
s
,
a
c
hi
e
vi
n
g
a
n
A
U
C
s
c
ore
of
1.
00,
w
i
t
h
ve
r
i
fi
c
a
t
i
on
t
h
rough
t
he
c
onf
us
i
on
m
a
t
ri
x
a
nd
RO
C
a
n
a
l
ys
i
s
.
2.
3
.
3.
M
od
e
l
3:
c
h
u
r
n
p
r
e
d
i
c
ti
on
Churn
w
a
s
de
f
i
ne
d
b
a
s
e
d
on
a
s
i
x
-
m
o
nt
h
c
o
ns
e
c
ut
i
ve
z
e
ro
c
o
nt
ri
but
i
on
rul
e
.
T
he
m
ode
l
pi
p
e
l
i
ne
i
nc
l
ude
d
a
c
o
l
u
m
n
t
r
a
ns
for
m
e
r
f
or
s
c
a
l
i
n
g
num
e
ri
c
a
l
fe
a
t
u
re
s
a
n
d
one
-
hot
e
nc
o
di
ng
c
a
t
e
go
ri
c
a
l
v
a
r
i
a
b
l
e
s
.
A
ra
ndo
m
f
ore
s
t
c
l
a
s
s
i
fi
e
r
w
a
s
us
e
d
w
i
t
h
i
n
t
h
i
s
p
i
pe
l
i
n
e
.
M
od
e
l
pe
rf
orm
a
n
c
e
w
a
s
e
v
a
l
u
a
t
e
d
us
i
ng
RO
C
-
A
U
C
,
a
c
c
ura
c
y
,
pre
c
i
s
i
on
,
a
nd
re
c
a
l
l
.
T
h
e
m
ode
l
’s
pr
e
di
c
t
i
ve
p
e
rf
orm
a
nc
e
,
e
va
l
u
a
t
e
d
us
i
ng
RO
C
-
A
U
C
,
a
c
c
ur
a
c
y,
pre
c
i
s
i
on
,
a
nd
r
e
c
a
l
l
,
i
s
pre
s
e
nt
e
d
i
n
T
a
bl
e
1
.
T
he
s
e
re
s
u
l
t
s
h
i
ghl
i
gh
t
t
h
e
m
ode
l
’s
e
xc
e
pt
i
ona
l
a
bi
l
i
t
y
t
o
d
i
s
t
i
ng
ui
s
h
b
e
t
w
e
e
n
c
hur
ne
d
a
nd
a
c
t
i
v
e
m
e
m
be
rs
.
R
a
ndo
m
fore
s
t
w
a
s
s
e
l
e
c
t
e
d
for
i
t
s
robus
t
n
e
s
s
,
i
n
t
e
rpr
e
t
a
bi
l
i
t
y,
a
n
d
h
i
gh
p
e
rfo
rm
a
nc
e
w
i
t
h
m
i
xe
d
da
t
a
t
yp
e
s
.
C
om
p
a
re
d
t
o
b
l
a
c
k
-
box
de
e
p
l
e
a
r
n
i
ng
m
od
e
l
s
[19]
,
[20
]
,
i
t
off
e
rs
a
b
e
t
t
e
r
t
ra
d
e
-
off
be
t
w
e
e
n
a
c
c
ur
a
c
y
a
nd
e
xp
l
a
i
n
a
b
i
l
i
t
y
,
c
ri
t
i
c
a
l
i
n
r
e
gu
l
a
t
e
d
p
e
ns
i
on
e
nvi
ro
nm
e
nt
s
.
2.
4
.
M
od
e
l
e
val
u
at
i
on
an
d
d
e
p
l
o
yme
n
t
A
l
l
m
od
e
l
s
w
e
r
e
s
e
ri
a
l
i
z
e
d
us
i
ng
j
ob
l
i
b
for
p
e
rs
i
s
t
e
nc
e
a
nd
l
a
t
e
r
us
e
.
T
he
de
pl
oy
m
e
nt
a
r
c
hi
t
e
c
t
u
re
ut
i
l
i
z
e
s
F
a
s
t
A
P
I
to
s
e
rv
e
pr
e
di
c
t
i
ons
t
h
rough
RE
S
T
ful
e
ndpo
i
nt
s
.
T
he
s
e
A
P
Is
can
be
i
nt
e
gr
a
t
e
d
w
i
t
h
P
ow
e
r
B
I
t
o
prov
i
d
e
a
re
a
l
-
t
i
m
e
da
s
hbo
a
rd
for
p
e
ns
i
o
n
a
d
m
i
ni
s
t
r
a
t
ors
,
e
nh
a
nc
i
ng
t
he
i
r
a
bi
l
i
t
y
t
o
m
a
k
e
pro
a
c
t
i
v
e
,
da
t
a
-
dri
ve
n
de
c
i
s
i
ons
.
V
i
s
ua
l
re
pre
s
e
nt
a
t
i
on:
t
h
e
s
ys
t
e
m
a
rc
h
i
t
e
c
t
ure
d
i
a
gr
a
m
i
n
F
i
gur
e
1
ou
t
l
i
ne
s
t
he
ful
l
e
nd
-
to
-
e
nd
w
orkf
l
ow
,
fro
m
ra
w
d
a
t
a
i
n
put
t
o
m
ode
l
de
pl
oy
m
e
n
t
a
nd
da
s
hbo
a
rd
i
nt
e
gr
a
t
i
on.
T
h
e
m
e
t
ho
dol
o
gy
w
orkfl
ow
d
i
a
gr
a
m
i
n
F
i
g
ure
2
p
rovi
de
s
a
d
e
t
a
i
l
e
d
f
l
ow
of
t
he
a
n
a
l
y
t
i
c
a
l
pro
c
e
s
s
e
s
i
nvol
v
e
d
i
n
t
r
a
i
n
i
ng
a
nd
e
va
l
ua
t
i
ng
e
a
c
h
m
ode
l
.
T
o
pr
e
s
e
rv
e
da
t
a
pr
i
v
a
c
y
,
a
l
l
m
e
m
be
r
d
a
t
a
w
a
s
a
nony
m
i
z
e
d
b
e
for
e
us
e
.
D
uri
ng
de
pl
o
ym
e
nt
,
s
e
c
ur
e
H
T
T
P
S
e
ndpo
i
nt
s
a
nd
e
n
c
rypt
e
d
s
t
ora
ge
w
e
r
e
us
e
d
t
o
pr
ot
e
c
t
s
e
ns
i
t
i
v
e
i
nfor
m
a
t
i
on
,
e
ns
uri
n
g
c
om
pl
i
a
n
c
e
w
i
t
h
d
a
t
a
prot
e
c
t
i
o
n
s
t
a
nd
a
rds
.
T
a
b
l
e
1
.
M
od
e
l
m
e
t
ri
c
s
RO
C
-
AUC
A
c
c
u
ra
c
y
P
re
c
i
s
i
o
n
a
n
d
r
e
c
a
l
l
0
.
9
9
9
9
7
9
9
.
8
6
%
1
.
0
0
3.
R
ES
U
LTS
A
N
D
D
I
S
C
U
S
S
I
O
N
3
.
1
.
M
od
e
l
1:
tota
l
c
on
t
r
i
b
u
ti
on
p
r
e
d
i
c
t
i
on
(r
e
gr
e
s
s
i
on
)
T
hi
s
m
o
de
l
us
e
s
r
e
gre
s
s
i
on
a
n
a
l
ys
i
s
vi
a
ra
n
dom
for
e
s
t
re
gre
s
s
or
t
o
for
e
c
a
s
t
fut
ur
e
c
on
t
ri
but
i
on
a
m
ou
nt
s
pe
r
m
e
m
b
e
r
.
It
i
s
a
c
or
e
pa
rt
of
p
e
rfor
m
a
nc
e
fore
c
a
s
t
i
ng
a
n
d
a
l
i
gns
di
re
c
t
l
y
t
o
i
m
pr
ove
a
c
c
ura
c
y
us
i
ng
ML
[14]
.
M
L
t
e
c
hni
q
ue
:
re
gr
e
s
s
i
on
(r
a
ndo
m
for
e
s
t
,
opt
i
o
na
l
l
y
e
x
t
e
n
d
t
o
n
e
ur
a
l
n
e
t
w
orks
)
.
Con
t
ri
but
i
on:
di
re
c
t
l
y
s
uppor
t
s
pe
rfo
rm
a
nc
e
for
e
c
a
s
t
i
ng
by
pr
e
di
c
t
i
ng
c
a
s
h
i
nfl
ow
s
fro
m
m
e
m
be
rs
.
F
i
gur
e
3
i
l
l
us
t
ra
t
e
s
t
h
e
t
ra
i
ni
ng
p
e
rfo
rm
a
nc
e
m
e
t
r
i
c
s
of
t
he
r
e
gre
s
s
i
on
m
o
de
l
s
,
c
om
pa
ri
ng
l
i
n
e
a
r
r
e
gr
e
s
s
i
on
a
nd
ra
n
dom
fo
re
s
t
a
ppro
a
c
h
e
s
.
F
i
gure
4
d
e
p
i
c
t
s
t
he
r
e
l
a
t
i
ons
hi
p
b
e
t
w
e
e
n
a
c
t
u
a
l
a
nd
pre
di
c
t
e
d
c
on
t
ri
b
ut
i
on
va
l
u
e
s
g
e
n
e
ra
t
e
d
b
y
t
he
ra
n
dom
for
e
s
t
re
gre
s
s
or
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
46
-
5
5
50
F
i
gure
3
.
T
ra
i
ni
ng
m
e
t
ri
c
s
:
l
i
ne
a
r
re
gr
e
s
s
i
on
a
n
d
r
a
ndo
m
fore
s
t
F
i
gure
4
.
R
a
ndo
m
for
e
s
t
a
c
t
ua
l
vs
pr
e
di
c
t
e
d
3
.
2
.
M
od
e
l
2:
c
on
t
r
i
b
u
t
i
on
d
r
op
r
i
s
k
d
e
te
c
ti
on
(c
l
as
s
i
f
i
c
at
i
on
)
T
hi
s
m
ode
l
i
de
nt
i
fi
e
s
m
e
m
be
r
s
l
i
ke
l
y
t
o
re
duc
e
c
o
nt
ri
but
i
o
ns
,
w
hi
c
h
i
s
e
s
s
e
nt
i
a
l
for
unde
r
s
t
a
ndi
ng
c
ont
ri
but
i
o
n
t
re
nd
s
.
It
w
or
ks
by
c
om
pa
ri
ng
e
a
c
h
m
e
m
be
r
'
s
c
ont
ri
but
i
o
n
a
ga
i
n
s
t
a
t
hre
s
hol
d
(e
.
g.
,
70%
of
t
h
e
a
ve
ra
ge
).
M
L
t
e
c
hni
que
:
c
l
a
s
s
i
fi
c
a
t
i
on
(ra
n
dom
fo
re
s
t
c
l
a
s
s
i
fi
e
r).
C
ont
ri
but
i
on
:
h
e
l
ps
fore
c
a
s
t
do
w
nw
a
rd
t
re
nds
i
n
c
ont
ri
b
ut
i
on
s
a
nd
c
a
s
h
fl
ow
ri
s
k
s
.
O
pt
i
m
i
z
a
t
i
on:
t
he
fra
m
e
w
ork
c
a
n
be
e
xpa
nde
d
t
o
i
nc
l
ude
e
xt
e
rna
l
fe
a
t
ure
s
(e
.
g.
,
m
a
c
roe
c
on
om
i
c
i
ndi
c
a
t
or
s
or
m
e
m
be
r
s
a
l
a
r
y
ba
nds
)
t
o
e
nri
c
h
a
c
c
ura
c
y.
F
i
gure
5
p
re
s
e
nt
s
t
he
e
va
l
ua
t
i
on
m
e
t
ri
c
s
o
f
t
he
ra
n
dom
fo
re
s
t
c
l
a
s
s
i
fi
e
r
us
e
d
t
o
i
de
nt
i
fy
m
e
m
be
r
s
a
t
ri
s
k
of
re
duc
i
ng
t
he
i
r
c
ont
ri
but
i
o
ns
.
F
i
g
ure
6
s
ho
w
s
t
he
c
onfu
s
i
on
m
a
t
ri
x
of
t
he
ra
ndom
fore
s
t
c
l
a
s
s
i
fi
e
r,
i
l
l
us
t
ra
t
i
ng
i
t
s
c
l
a
s
s
i
fi
c
a
t
i
o
n
pe
rf
orm
a
nc
e
i
n
di
s
t
i
ngui
s
hi
n
g
hi
gh
-
ri
s
k
a
nd
l
ow
-
ri
s
k
m
e
m
be
r
s
.
F
i
g
ure
7
di
s
pl
a
y
s
t
he
RO
C
c
urve
for
drop
-
ri
s
k
pre
di
c
t
i
on,
hi
ghl
i
g
ht
i
ng
t
he
c
l
a
s
s
i
fi
e
r’
s
a
bi
l
i
t
y
t
o
di
s
c
ri
m
i
na
t
e
be
t
w
e
e
n
m
e
m
be
rs
w
i
t
h
de
c
l
i
ni
ng
a
nd
s
t
a
bl
e
c
ont
ri
b
ut
i
on
pa
t
t
e
rn
s
.
F
i
gure
5
.
E
va
l
ua
t
i
on
m
e
t
ri
c
s
r
a
ndo
m
c
l
a
s
s
i
f
i
e
r
F
i
gure
6
.
R
a
ndo
m
c
l
a
s
s
i
f
i
e
r
c
o
nfus
i
on
m
a
t
ri
x
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
Cl
oud
-
b
as
e
d
pr
e
d
i
c
t
i
v
e
ana
l
y
t
i
c
s
f
or
p
e
ns
i
on
f
u
nd
pe
r
f
or
m
an
c
e
opt
i
m
i
z
a
t
i
on
(
B
e
au
t
y
G
ar
aba
)
51
F
i
gure
7
.
RO
C
c
urv
e
fo
r
dr
op
ri
s
k
3
.
3
.
M
od
e
l
3:
m
e
mb
e
r
c
h
u
r
n
p
r
e
d
i
c
t
i
on
(
c
l
as
s
i
fi
c
ati
on
)
T
hi
s
m
od
e
l
pr
e
di
c
t
s
w
he
t
he
r
a
m
e
m
b
e
r
i
s
l
i
k
e
l
y
t
o
e
xi
t
t
he
f
und,
w
hi
c
h
a
ff
e
c
t
s
l
ong
-
t
e
r
m
c
a
s
h
fl
ow
a
nd
l
i
a
bi
l
i
t
i
e
s
.
Ch
urn
p
re
d
i
c
t
i
on
i
s
a
r
i
s
k
a
s
s
e
s
s
m
e
n
t
t
ool
,
e
s
pe
c
i
a
l
l
y
i
f
l
i
nk
e
d
t
o
d
e
m
ogr
a
phi
c
or
e
c
ono
m
i
c
i
ndi
c
a
t
ors
.
M
L
t
e
c
hni
que
[6]
,
[14]
,
[21]
,
[22]
:
c
l
a
s
s
i
fi
c
a
t
i
on
(
ra
ndo
m
fore
s
t
c
l
a
s
s
i
fi
e
r)
.
Cont
r
i
bu
t
i
on
:
s
u
ppor
t
s
ri
s
k
m
od
e
l
i
ng
[23]
a
n
d
f
i
na
nc
i
a
l
s
t
a
bi
l
i
t
y
by
for
e
c
a
s
t
i
n
g
p
ot
e
nt
i
a
l
fu
nd
e
x
i
t
s
.
E
xt
e
ns
i
on
:
c
a
n
i
nc
l
ud
e
f
e
a
t
ur
e
s
l
i
k
e
a
ge
,
t
e
nure
,
or
e
c
ono
m
i
c
s
t
r
e
s
s
i
ndi
c
a
t
ors
t
o
s
i
m
ul
a
t
e
t
he
i
m
pa
c
t
of
m
a
rk
e
t
/
i
nf
l
a
t
i
on
s
ho
c
ks
.
Com
p
a
re
d
to
ba
s
e
l
i
n
e
l
i
ne
a
r
m
od
e
l
s
,
t
h
e
r
a
ndo
m
for
e
s
t
a
nd
ML
m
od
e
l
s
a
c
hi
e
ve
d
s
i
gni
f
i
c
a
nt
l
y
hi
g
he
r
pre
di
c
t
i
ve
a
c
c
ura
c
y
.
F
or
i
ns
t
a
n
c
e
,
t
he
c
h
urn
m
od
e
l
re
a
c
h
e
d
99.
86
%
a
c
c
ura
c
y
(vs
.
78%
i
n
ba
s
e
l
i
ne
)
,
a
nd
t
h
e
c
ont
r
i
bu
t
i
on
pre
di
c
t
i
on
m
ode
l
a
c
h
i
e
v
e
d
an
R²
of
0.
8
5
(vs
.
0.
6
1
in
b
a
s
e
l
i
n
e
).
F
i
gur
e
8
s
u
m
m
a
r
i
s
e
s
t
he
e
v
a
l
u
a
t
i
on
m
e
t
ri
c
s
for
t
h
e
m
e
m
be
r
c
h
urn
p
re
d
i
c
t
i
on
m
ode
l
,
de
m
ons
t
ra
t
i
ng
i
t
s
s
t
r
ong
p
re
d
i
c
t
i
v
e
p
e
rfor
m
a
nc
e
r
e
l
a
t
i
ve
t
o
ba
s
e
l
i
n
e
a
ppro
a
c
h
e
s
.
T
he
h
i
gh
A
U
C
va
l
ue
s
(
a
ppro
a
c
h
i
ng
1
.
0)
i
nd
i
c
a
t
e
s
t
rong
m
o
de
l
d
i
s
c
r
i
m
i
na
t
i
o
n
a
b
i
l
i
t
y
.
T
h
i
s
i
m
pl
i
e
s
t
ha
t
t
h
e
c
l
a
s
s
i
f
i
e
rs
c
a
n
r
e
l
i
a
b
l
y
d
i
s
t
i
ng
ui
s
h
b
e
t
w
e
e
n
r
i
s
ky
a
nd
non
-
ri
s
ky
m
e
m
b
e
rs
,
e
na
bl
i
ng
m
o
re
pr
e
c
i
s
e
fu
nd
m
a
n
a
ge
m
e
nt
d
e
c
i
s
i
ons
.
F
i
gur
e
9
p
re
s
e
n
t
s
t
h
e
c
onfus
i
o
n
m
a
t
ri
x
f
or
t
h
e
dro
p
-
ri
s
k
c
l
a
s
s
i
fi
c
a
t
i
on
m
od
e
l
,
i
l
l
us
t
ra
t
i
ng
t
h
e
m
o
de
l
’s
e
ff
e
c
t
i
v
e
ne
s
s
i
n
c
or
re
c
t
l
y
i
d
e
nt
i
fyi
ng
m
e
m
be
rs
a
t
r
i
s
k
of
r
e
du
c
e
d
c
on
t
ri
bu
t
i
ons
.
T
he
s
e
re
s
ul
t
s
ha
v
e
d
i
r
e
c
t
i
m
pl
i
c
a
t
i
ons
f
or
p
e
ns
i
on
fund
po
l
i
c
y
a
nd
s
t
r
a
t
e
gy.
A
c
c
ur
a
t
e
c
hurn
pre
d
i
c
t
i
on
e
n
a
b
l
e
s
pro
a
c
t
i
ve
re
t
e
n
t
i
o
n
s
t
ra
t
e
g
i
e
s
.
D
ro
p
-
ri
s
k
d
e
t
e
c
t
i
on
c
a
n
i
nf
or
m
c
on
t
ri
bu
t
i
o
n
e
nfor
c
e
m
e
nt
or
i
n
c
e
n
t
i
v
e
s
.
R
e
a
l
-
t
i
m
e
for
e
c
a
s
t
i
ng
a
l
i
gns
w
i
t
h
s
t
ra
t
e
gi
c
fi
n
a
nc
i
a
l
pl
a
nn
i
ng,
e
ns
ur
i
ng
s
us
t
a
i
n
a
bi
l
i
t
y
und
e
r
un
c
e
rt
a
i
n
m
a
c
roe
c
ono
m
i
c
c
on
di
t
i
ons
.
F
i
gure
8
.
M
e
t
r
i
c
s
s
u
m
m
a
ry
for
m
e
m
be
r
c
hurn
p
r
e
di
c
t
i
on
3.
4
.
M
od
e
l
e
xp
l
ai
n
ab
i
l
i
ty
an
d
e
th
i
c
a
l
c
on
s
i
d
e
r
at
i
on
s
G
i
ve
n
t
h
e
re
gu
l
a
t
ory
n
a
t
ur
e
of
pe
ns
i
on
f
und
m
a
na
ge
m
e
n
t
,
e
x
pl
a
i
na
b
i
l
i
t
y
w
a
s
a
ke
y
r
e
qu
i
re
m
e
nt
[24]
.
F
e
a
t
u
re
i
m
por
t
a
n
c
e
p
l
ot
s
a
n
d
S
H
a
pl
e
y
A
d
di
t
i
v
e
e
xP
l
a
na
t
i
o
ns
(S
H
A
P
)
v
a
l
u
e
s
w
e
re
us
e
d
t
o
i
n
t
e
rp
re
t
m
od
e
l
out
pu
t
s
,
off
e
ri
ng
t
r
a
ns
p
a
re
n
c
y
t
o
fund
a
d
m
i
n
i
s
t
r
a
t
ors
.
F
i
gu
re
10
i
l
l
us
t
r
a
t
e
s
t
h
e
s
t
ru
c
t
u
re
a
nd
d
e
c
i
s
i
on
l
ogi
c
of
t
he
r
a
ndo
m
f
ore
s
t
c
l
a
s
s
i
f
i
e
r
,
s
upp
ort
i
ng
t
he
i
n
t
e
rpr
e
t
a
b
i
l
i
t
y
of
t
he
m
o
de
l
out
put
s
.
F
i
gu
re
1
1
c
om
p
a
re
s
ba
s
e
l
i
n
e
c
ont
r
i
bu
t
i
on
e
s
t
i
m
a
t
e
s
w
i
t
h
m
od
e
l
-
pre
d
i
c
t
e
d
v
a
l
u
e
s
,
i
l
l
us
t
ra
t
i
ng
t
he
i
m
prov
e
m
e
n
t
a
c
h
i
e
v
e
d
t
hrou
gh
t
he
ML
a
ppr
oa
c
h.
E
t
hi
c
a
l
c
ons
i
de
r
a
t
i
ons
w
e
r
e
a
l
s
o
obs
e
rve
d
i
n
da
t
a
h
a
nd
l
i
ng
,
m
ode
l
f
a
i
rn
e
s
s
,
a
nd
d
e
c
i
s
i
on
a
c
c
ount
a
b
i
l
i
t
y
[25]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
46
-
5
5
52
F
i
gure
9
.
C
onfus
i
o
n
m
a
t
ri
x
for
dro
p
r
i
s
k
F
i
gure
10
.
Ra
nd
om
for
e
s
t
c
l
a
s
s
i
fi
e
r
F
i
gure
1
1
.
Ba
s
e
l
i
n
e
vs
pr
e
di
c
t
e
d
c
on
t
ri
but
i
ons
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
Cl
oud
-
b
as
e
d
pr
e
d
i
c
t
i
v
e
ana
l
y
t
i
c
s
f
or
p
e
ns
i
on
f
u
nd
pe
r
f
or
m
an
c
e
opt
i
m
i
z
a
t
i
on
(
B
e
au
t
y
G
ar
aba
)
53
4.
C
O
N
C
LU
S
I
O
N
M
ore
s
o
phi
s
t
i
c
a
t
e
d
a
nd
f
l
e
x
i
bl
e
fore
c
a
s
t
i
ng
t
ool
s
a
r
e
re
q
ui
re
d
due
t
o
t
he
gr
ow
i
ng
c
om
p
l
e
x
i
t
y
of
m
a
n
a
gi
ng
p
e
ns
i
on
funds
,
w
hi
c
h
i
s
fue
l
e
d
by
m
a
r
ke
t
v
ol
a
t
i
l
i
t
y
,
d
e
m
ogra
phi
c
c
ha
ng
e
s
,
a
nd
e
c
ono
m
i
c
unc
e
r
t
a
i
nt
y.
By
c
re
a
t
i
ng
a
c
l
oud
-
ba
s
e
d
pr
e
di
c
t
i
ve
a
n
a
l
y
t
i
c
s
m
od
e
l
s
p
e
c
i
fi
c
a
l
l
y
s
ui
t
e
d
t
o
t
he
pe
ns
i
on
f
und
e
nvi
r
onm
e
nt
i
n
Z
i
m
b
a
bw
e
,
t
hi
s
s
t
udy
t
a
c
kl
e
d
t
h
e
s
e
i
s
s
ue
s
.
T
he
m
od
e
l
e
n
ha
n
c
e
s
t
he
a
c
c
ura
c
y
of
c
a
s
h
fl
ow
proj
e
c
t
i
ons
,
c
ont
r
i
bu
t
i
on
t
re
nds
,
a
nd
p
e
rfor
m
a
n
c
e
pr
e
di
c
t
i
ons
by
i
n
c
orpo
ra
t
i
ng
ML
t
e
c
hni
q
ue
s
,
s
u
c
h
a
s
re
gre
s
s
i
o
n
a
n
a
l
ys
i
s
,
ne
ur
a
l
n
e
t
w
or
ks
,
a
nd
t
i
m
e
s
e
r
i
e
s
for
e
c
a
s
t
i
ng
.
T
h
rough
s
c
a
l
a
bl
e
c
l
oud
de
p
l
oy
m
e
n
t
a
nd
re
a
l
-
t
i
m
e
da
t
a
i
n
t
e
gr
a
t
i
on
,
t
h
e
s
o
l
ut
i
on
no
t
on
l
y
i
m
pr
ove
s
fund
m
a
n
a
ge
rs
'
c
a
p
a
c
i
t
y
t
o
m
a
k
e
d
a
t
a
-
dr
i
ve
n
i
nve
s
t
m
e
nt
de
c
i
s
i
ons
b
ut
a
l
s
o
pr
om
o
t
e
s
l
on
g
-
t
e
rm
fi
na
n
c
i
a
l
s
us
t
a
i
n
a
b
i
l
i
t
y
.
T
he
m
ode
l
l
a
ys
t
h
e
groun
dw
ork
for
fut
ur
e
s
t
udy
a
nd
a
dv
a
nc
e
m
e
n
t
i
n
pre
d
i
c
t
i
v
e
p
e
ns
i
on
fund
m
a
n
a
ge
m
e
nt
i
n
r
e
l
a
t
e
d
c
o
nt
e
xt
s
a
nd
pro
vi
d
e
s
a
c
re
a
t
i
ve
,
us
e
ful
s
ol
ut
i
on
t
o
t
he
dr
a
w
ba
c
ks
of
c
onv
e
n
t
i
on
a
l
for
e
c
a
s
t
i
n
g
t
e
c
h
ni
qu
e
s
.
U
l
t
i
m
a
t
e
l
y
,
t
hi
s
w
or
k
m
a
k
e
s
a
s
i
gn
i
fi
c
a
n
t
c
ont
r
i
bu
t
i
on
t
o
t
he
dom
a
i
ns
o
f
da
t
a
-
dri
v
e
n
d
e
c
i
s
i
on
-
m
a
ki
n
g,
pe
ns
i
on
fun
d
a
d
m
i
ni
s
t
r
a
t
i
on
,
a
n
d
fi
na
n
c
i
a
l
a
n
a
l
y
t
i
c
s
.
T
hi
s
r
e
s
e
a
r
c
h
o
pe
ns
s
e
ve
r
a
l
a
ve
n
ue
s
for
e
xt
e
ns
i
on:
i
)
t
e
c
hni
c
a
l
e
nh
a
n
c
e
m
e
n
t
s
:
i
nc
o
rpor
a
t
e
de
e
p
-
l
e
a
rn
i
ng
m
od
e
l
s
(
L
S
T
M
s
)
for
s
e
que
nc
e
for
e
c
a
s
t
i
ng
o
f
l
ong
-
t
e
rm
t
r
e
nds
,
i
i
)
po
l
i
c
y
i
nt
e
gra
t
i
o
n:
a
l
i
gn
pre
di
c
t
i
ve
out
put
s
w
i
t
h
c
o
m
pl
i
a
n
c
e
c
h
e
c
ks
a
nd
r
e
gu
l
a
t
ory
r
e
qui
r
e
m
e
n
t
s
,
i
i
i
)
c
r
os
s
-
s
e
c
t
or
a
ppl
i
c
a
t
i
ons
:
a
da
pt
t
he
m
ode
l
fra
m
e
w
or
k
t
o
ot
h
e
r
s
e
c
t
ors
s
uc
h
a
s
h
e
a
l
t
h
i
ns
ura
n
c
e
a
n
d
s
o
c
i
a
l
w
e
l
fa
r
e
,
i
v)
re
s
ou
rc
e
o
pt
i
m
i
z
a
t
i
o
n:
opt
i
m
i
z
e
m
od
e
l
t
ra
i
ni
n
g
a
n
d
i
nf
e
re
nc
e
c
os
t
s
for
l
ow
-
re
s
our
c
e
i
ns
t
i
t
u
t
i
o
ns
us
i
ng
s
e
rve
r
l
e
s
s
c
om
p
ut
i
ng
or
e
dg
e
de
vi
c
e
s
,
a
nd
v)
c
ons
i
de
r
na
m
i
n
g
t
h
e
m
od
e
l
i
ng
a
ppro
a
c
h
a
s
a
fr
a
m
e
w
ork
(
e
.
g.
,
“
c
l
oud
-
a
c
t
ua
r
i
a
l
pr
e
di
c
t
i
ve
opt
i
m
i
z
a
t
i
on
m
ode
l
”
(CA
P
O
M
)
.
A
C
K
N
O
WL
ED
G
M
EN
TS
I
e
xt
e
nd
m
y
he
a
rt
f
e
l
t
gr
a
t
i
t
ud
e
t
o
m
y
a
c
a
d
e
m
i
c
s
u
pe
rv
i
s
ors
,
w
hos
e
i
ns
i
ght
f
ul
g
ui
d
a
nc
e
,
c
ons
t
ruc
t
i
v
e
c
ri
t
i
qu
e
s
,
a
nd
e
n
c
our
a
ge
m
e
n
t
w
e
re
i
ns
t
ru
m
e
nt
a
l
t
hrou
ghout
t
he
de
v
e
l
o
pm
e
nt
of
t
hi
s
r
e
s
e
a
rc
h
.
I
a
m
e
qu
a
l
l
y
t
ha
nk
ful
t
o
t
h
e
fa
c
ul
t
y
m
e
m
be
rs
o
f
t
h
e
c
l
oud
c
om
p
ut
i
ng
de
p
a
rt
m
e
n
t
a
t
H
a
r
a
re
I
ns
t
i
t
ut
e
of
T
e
c
hnol
ogy,
w
h
os
e
a
c
a
de
m
i
c
s
uppor
t
a
nd
r
e
s
ourc
e
s
prov
i
d
e
d
t
h
e
fo
unda
t
i
o
n
for
t
hi
s
w
ork.
T
h
e
i
r
de
d
i
c
a
t
i
on
t
o
fos
t
e
ri
ng
i
nnov
a
t
i
on
a
nd
re
s
e
a
r
c
h
e
xc
e
l
l
e
n
c
e
h
a
s
b
e
e
n
i
nva
l
ua
bl
e
.
S
pe
c
i
a
l
a
p
pre
c
i
a
t
i
on
a
l
s
o
goe
s
t
o
t
h
e
p
e
ns
i
o
n
fund
i
ns
t
i
t
u
t
i
o
ns
a
nd
re
g
ul
a
t
ory
bod
i
e
s
i
n
Z
i
m
b
a
bw
e
t
h
a
t
fa
c
i
l
i
t
a
t
e
d
a
c
c
e
s
s
t
o
v
i
t
a
l
d
a
t
a
a
nd
d
om
a
i
n
k
now
l
e
dge
.
T
he
i
r
c
oope
ra
t
i
on
s
i
g
ni
f
i
c
a
nt
l
y
e
n
ri
c
he
d
t
he
pr
a
c
t
i
c
a
l
re
l
e
va
n
c
e
of
t
hi
s
s
t
udy
.
I
a
m
fu
rt
h
e
r
i
nd
e
bt
e
d
t
o
m
y
fa
m
i
l
y
f
or
t
he
i
r
unw
a
ve
r
i
ng
s
uppor
t
,
p
a
t
i
e
n
c
e
,
a
nd
m
ot
i
va
t
i
on
du
ri
ng
t
h
i
s
r
e
s
e
a
rc
h
j
ourn
e
y.
F
U
N
D
I
N
G
I
N
F
O
R
M
A
TI
O
N
U
t
i
l
i
z
i
ng
a
va
i
l
a
bl
e
re
s
our
c
e
s
,
t
h
e
a
ut
ho
rs
a
dv
i
s
e
d
t
ha
t
t
h
e
r
e
s
e
a
rc
h
w
a
s
c
ondu
c
t
e
d
w
i
t
hout
t
h
e
ne
e
d
for
e
xt
e
rna
l
fund
i
ng,
d
e
m
o
ns
t
ra
t
i
n
g
t
he
c
a
p
a
c
i
t
y
t
o
a
c
h
i
e
v
e
r
e
s
ul
t
s
c
os
t
-
e
ff
e
c
t
i
ve
l
y.
A
U
TH
O
R
C
O
N
TR
I
BU
TI
O
N
S
S
TA
T
EM
EN
T
T
hi
s
j
our
na
l
us
e
s
t
h
e
Cont
ri
bu
t
or
Rol
e
s
T
a
x
ono
m
y
(
C
Re
di
T
)
t
o
re
c
ogn
i
z
e
i
nd
i
vi
du
a
l
a
ut
hor
c
ont
r
i
bu
t
i
ons
,
r
e
du
c
e
a
ut
hors
h
i
p
di
s
pu
t
e
s
,
a
nd
fa
c
i
l
i
t
a
t
e
c
o
l
l
a
bora
t
i
on
.
N
ame
of
A
u
th
o
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Be
a
u
t
y
G
a
r
a
ba
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
M
a
i
nfor
d
M
u
t
a
n
da
v
a
ri
✓
✓
✓
✓
✓
✓
✓
✓
✓
J
e
r
i
t
a
Chi
b
ha
bh
a
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
ft
w
a
re
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
rm
a
l
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
s
o
u
rc
e
s
D
:
D
a
t
a
Cu
ra
t
i
o
n
O
:
W
ri
t
i
n
g
-
O
ri
g
i
n
a
l
D
ra
ft
E
:
W
ri
t
i
n
g
-
Re
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
s
u
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
rv
i
s
i
o
n
P
:
P
ro
j
e
c
t
a
d
m
i
n
i
s
t
ra
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
s
i
t
i
o
n
C
O
N
F
LI
C
T
O
F
I
N
T
ER
ES
T
S
TA
T
EM
EN
T
T
he
a
u
t
hors
de
c
l
a
r
e
t
ha
t
t
he
r
e
a
r
e
n
o
c
onf
l
i
c
t
s
of
i
nt
e
re
s
t
r
e
l
a
t
e
d
t
o
t
h
i
s
r
e
s
e
a
rc
h.
I
N
F
O
R
M
ED
C
O
N
S
EN
T
T
he
pro
t
e
c
t
i
on
of
pri
v
a
c
y
i
s
a
l
e
g
a
l
ri
g
ht
t
h
a
t
m
us
t
n
ot
be
bre
a
c
he
d
w
i
t
hou
t
i
ndi
vi
du
a
l
i
n
form
e
d
c
ons
e
n
t
.
In
c
a
s
e
s
w
he
r
e
t
he
i
de
n
t
i
f
i
c
a
t
i
on
of
p
e
rs
on
a
l
i
nf
or
m
a
t
i
on
i
s
ne
c
e
s
s
a
ry
f
or
s
c
i
e
nt
i
fi
c
r
e
a
s
ons
,
a
u
t
hors
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2722
-
3221
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
,
V
o
l
.
7
,
N
o
.
1
,
M
a
rc
h
20
26
:
46
-
5
5
54
s
houl
d
ob
t
a
i
n
ful
l
do
c
u
m
e
n
t
a
t
i
on
of
i
nf
orm
e
d
c
o
ns
e
n
t
,
i
nc
l
u
di
ng
w
ri
t
t
e
n
p
e
rm
i
s
s
i
on
fr
om
t
he
p
a
t
i
e
n
t
pr
i
or
t
o
i
nc
l
us
i
on
i
n
t
he
s
t
udy
.
I
nc
or
pora
t
e
t
h
e
fo
l
l
ow
i
ng
(o
r
a
s
i
m
i
l
a
r
)
s
t
a
t
e
m
e
n
t
:
W
e
h
a
ve
obt
a
i
n
e
d
i
nfor
m
e
d
c
ons
e
n
t
from
a
l
l
i
nd
i
vi
dua
l
s
i
nc
l
ude
d
i
n
t
hi
s
s
t
udy
.
ETH
I
C
A
L
A
P
P
R
O
V
A
L
T
hi
s
re
s
e
a
r
c
h
u
t
i
l
i
z
e
d
a
nony
m
i
z
e
d
p
e
ns
i
o
n
da
t
a
a
nd
di
d
not
r
e
qui
r
e
e
t
hi
c
a
l
a
p
prova
l
fro
m
t
h
e
da
t
a
s
ubj
e
c
t
s
.
T
h
e
us
e
of
a
n
ony
m
i
z
e
d
d
a
t
a
e
ns
ure
d
t
h
a
t
no
i
d
e
nt
i
f
i
a
b
l
e
i
nfor
m
a
t
i
on
w
a
s
c
ol
l
e
c
t
e
d
.
T
hi
s
a
dh
e
re
s
t
o
e
t
h
i
c
a
l
gui
d
e
l
i
n
e
s
for
r
e
s
e
a
r
c
h
i
n
t
he
p
e
ns
i
ons
a
nd
i
ns
u
ra
n
c
e
s
e
c
t
o
r.
D
A
TA
A
V
A
I
LA
BI
LI
TY
D
ue
t
o
pri
v
a
c
y
a
nd
c
o
nfi
d
e
nt
i
a
l
i
t
y
c
on
c
e
rns
,
t
h
e
da
t
a
us
e
d
t
o
s
uppor
t
t
h
e
r
e
s
e
a
rc
h'
s
c
on
c
l
us
i
ons
ha
s
be
e
n
a
nony
m
i
z
e
d
a
nd
i
s
n
ot
a
va
i
l
a
bl
e
f
or
pub
l
i
c
s
h
a
ri
ng.
H
ow
e
ve
r
,
t
h
e
a
n
onym
i
z
e
d
d
a
t
a
c
a
n
be
m
a
d
e
a
v
a
i
l
a
b
l
e
upon
r
e
a
s
o
na
b
l
e
re
que
s
t
fr
om
t
h
e
a
u
t
hor
,
p
rovi
d
e
d
t
h
a
t
t
h
e
r
e
qu
e
s
t
a
dh
e
re
s
t
o
e
t
h
i
c
a
l
s
t
a
nda
rds
a
nd
da
t
a
prot
e
c
t
i
on
l
a
w
s
.
R
EF
ER
EN
C
ES
[1
]
A
.
M
.
d
e
K
l
e
rk
,
“
D
e
c
i
s
i
o
n
m
a
k
i
n
g
u
n
d
e
r
u
n
c
e
rt
a
i
n
t
y
a
n
d
ri
s
k
:
a
re
t
i
re
e
’s
p
e
rs
p
e
c
t
i
v
e
o
n
c
h
o
o
s
i
n
g
b
e
t
w
e
e
n
p
e
n
s
i
o
n
o
p
t
i
o
n
s
,
”
2
0
1
7
P
o
r
t
l
a
n
d
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
M
a
n
a
g
e
m
e
n
t
o
f
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
(P
ICM
E
T
)
,
p
p
.
1
–
1
2
,
2
0
1
7
,
d
o
i
:
1
0
.
2
3
9
1
9
/
P
ICM
E
T
.
2
0
1
7
.
8
1
2
5
3
0
9
.
[2
]
A
.
S
h
a
rm
a
a
n
d
S
.
K
a
b
a
d
e
,
“
O
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
ri
t
h
m
s
fo
r
p
e
n
s
i
o
n
a
s
s
e
t
a
l
l
o
c
a
t
i
o
n
u
n
d
e
r
m
a
rk
e
t
v
o
l
a
t
i
l
i
t
y
,
”
S
S
R
N
E
l
e
c
t
r
o
n
i
c
J
o
u
r
n
a
l
,
2
0
2
5
,
d
o
i
:
1
0
.
2
1
3
9
/
s
s
rn
.
5
3
7
3
8
8
2
.
[3
]
In
t
e
rn
a
t
i
o
n
a
l
M
o
n
e
t
a
ry
F
u
n
d
(IM
F
),
M
a
n
a
g
i
n
g
d
i
v
e
r
g
e
n
t
r
e
c
o
v
e
r
i
e
s
:
g
l
o
b
a
l
e
c
o
n
o
m
y
o
n
f
i
r
m
e
r
g
r
o
u
n
d
,
b
u
t
wi
t
h
d
i
v
e
r
g
e
n
t
r
e
c
o
v
e
r
i
e
s
a
m
i
d
h
i
g
h
u
n
c
e
r
t
a
i
n
t
y
l
i
s
t
e
n
wi
t
h
s
p
e
e
c
h
i
f
y
.
W
a
s
h
i
n
g
t
o
n
,
D
C,
U
n
i
t
e
d
S
t
a
t
e
s
:
In
t
e
rn
a
t
i
o
n
a
l
M
o
n
e
t
a
ry
F
u
n
d
,
2
0
2
1
.
[4
]
O
.
E
.
A
ro
,
“
P
re
d
i
c
t
i
v
e
a
n
a
l
y
t
i
c
s
i
n
fi
n
a
n
c
i
a
l
m
a
n
a
g
e
m
e
n
t
:
e
n
h
a
n
c
i
n
g
d
e
c
i
s
i
o
n
-
m
a
k
i
n
g
a
n
d
ri
s
k
m
a
n
a
g
e
m
e
n
t
,
”
In
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
R
e
s
e
a
r
c
h
P
u
b
l
i
c
a
t
i
o
n
a
n
d
R
e
v
i
e
ws
,
v
o
l
.
5
,
n
o
.
1
0
,
p
p
.
2
1
8
1
–
2
1
9
4
,
O
c
t
.
2
0
2
4
,
d
o
i
:
1
0
.
5
5
2
4
8
/
g
e
n
g
p
i
.
5
.
1
0
2
4
.
2
8
1
9
.
[5
]
A
.
A
.
Be
l
o
l
i
p
e
t
s
k
i
i
a
n
d
M
.
A
.
L
e
p
s
k
a
y
a
,
“
A
m
a
t
h
e
m
a
t
i
c
a
l
m
o
d
e
l
o
f
p
e
n
s
i
o
n
fu
n
d
o
p
e
ra
t
i
o
n
a
n
d
m
e
t
h
o
d
s
o
f
fu
n
d
s
t
a
b
i
l
i
t
y
a
n
a
l
y
s
i
s
,
”
Co
m
p
u
t
a
t
i
o
n
a
l
M
a
t
h
e
m
a
t
i
c
s
a
n
d
M
o
d
e
l
i
n
g
,
v
o
l
.
2
9
,
n
o
.
2
,
p
p
.
2
3
3
–
2
4
3
,
A
p
r
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
5
9
8
-
0
1
8
-
9
4
0
4
-
7.
[6
]
Y
.
Y
a
o
,
Z
.
X
i
a
o
,
B.
W
a
n
g
,
B
.
V
i
s
w
a
n
a
t
h
,
H
.
Z
h
e
n
g
,
a
n
d
B.
Y
.
Z
h
a
o
,
“
Co
m
p
l
e
x
i
t
y
v
s
.
p
e
rfo
rm
a
n
c
e
:
e
m
p
i
ri
c
a
l
a
n
a
l
y
s
i
s
o
f
m
a
c
h
i
n
e
l
e
a
rn
i
n
g
a
s
a
s
e
rv
i
c
e
,
”
IM
C
’1
7
:
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
1
7
In
t
e
r
n
e
t
M
e
a
s
u
r
e
m
e
n
t
Co
n
f
e
r
e
n
c
e
,
p
p
.
3
8
4
–
3
9
7
,
2
0
1
7
,
d
o
i
:
1
0
.
1
1
4
5
/
3
1
3
1
3
6
5
.
3
1
3
1
3
7
2
.
[7
]
O
E
CD
,
P
e
n
s
i
o
n
m
a
r
k
e
t
s
i
n
f
o
c
u
s
.
P
a
ri
s
,
F
ra
n
c
e
:
O
E
C
D
,
2
0
2
0
.
[8
]
X
.
Y
a
o
a
n
d
C.
J
i
a
n
g
,
“
Re
s
e
a
rc
h
a
n
d
a
p
p
l
i
c
a
t
i
o
n
o
f
c
o
m
m
u
n
i
t
y
p
e
n
s
i
o
n
s
e
rv
i
c
e
d
a
t
a
a
n
a
l
y
s
i
s
s
y
s
t
e
m
b
a
s
e
d
o
n
d
a
t
a
m
i
n
i
n
g
a
l
g
o
ri
t
h
m
,
”
2
0
2
1
2
0
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
S
y
m
p
o
s
i
u
m
o
n
D
i
s
t
r
i
b
u
t
e
d
Co
m
p
u
t
i
n
g
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
f
o
r
B
u
s
i
n
e
s
s
E
n
g
i
n
e
e
r
i
n
g
a
n
d
S
c
i
e
n
c
e
(D
CA
B
E
S
)
,
p
p
.
2
9
–
3
2
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
D
CA
BE
S
5
2
9
9
8
.
2
0
2
1
.
0
0
0
1
4
.
[9
]
T
.
K
.
L
e
e
,
J
.
H
.
Ch
o
,
D
.
S
.
K
w
o
n
,
a
n
d
S
.
Y
.
S
o
h
n
,
“
G
l
o
b
a
l
s
t
o
c
k
m
a
rk
e
t
i
n
v
e
s
t
m
e
n
t
s
t
ra
t
e
g
i
e
s
b
a
s
e
d
o
n
fi
n
a
n
c
i
a
l
n
e
t
w
o
rk
i
n
d
i
c
a
t
o
rs
u
s
i
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
,
”
E
x
p
e
r
t
S
y
s
t
e
m
s
wi
t
h
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
1
7
,
p
p
.
2
2
8
–
2
4
2
,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
s
w
a
.
2
0
1
8
.
0
9
.
0
0
5
.
[1
0
]
M
.
A
.
W
a
l
l
e
r
a
n
d
S
.
E
.
F
a
w
c
e
t
t
,
“
D
a
t
a
s
c
i
e
n
c
e
,
p
re
d
i
c
t
i
v
e
a
n
a
l
y
t
i
c
s
,
a
n
d
b
i
g
d
a
t
a
:
a
re
v
o
l
u
t
i
o
n
t
h
a
t
w
i
l
l
t
ra
n
s
fo
rm
s
u
p
p
l
y
c
h
a
i
n
d
e
s
i
g
n
a
n
d
m
a
n
a
g
e
m
e
n
t
,
”
J
o
u
r
n
a
l
o
f
B
u
s
i
n
e
s
s
L
o
g
i
s
t
i
c
s
,
v
o
l
.
3
4
,
n
o
.
2
,
p
p
.
7
7
–
8
4
,
2
0
1
3
,
d
o
i
:
1
0
.
1
1
1
1
/
j
b
l
.
1
2
0
1
0
.
[1
1
]
M
.
Ch
e
n
,
S
.
M
a
o
,
a
n
d
Y
.
L
i
u
,
“
Bi
g
d
a
t
a
:
a
s
u
rv
e
y
,
”
M
o
b
i
l
e
Ne
t
wo
r
k
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
9
,
n
o
.
2
,
p
p
.
1
7
1
–
2
0
9
,
A
p
r.
2
0
1
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
0
3
6
-
0
1
3
-
0
4
8
9
-
0.
[1
2
]
M
.
H
a
i
,
Y
.
Z
h
a
n
g
,
a
n
d
Y
.
Z
h
a
n
g
,
“
A
p
e
rfo
rm
a
n
c
e
e
v
a
l
u
a
t
i
o
n
o
f
c
l
a
s
s
i
fi
c
a
t
i
o
n
a
l
g
o
ri
t
h
m
s
fo
r
b
i
g
d
a
t
a
,
”
P
r
o
c
e
d
i
a
Co
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
2
2
,
p
p
.
1
1
0
0
–
1
1
0
7
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
p
ro
c
s
.
2
0
1
7
.
1
1
.
4
7
9
.
[1
3
]
Y
.
Y
.
-
Y
u
a
n
,
S
.
M
.
-
L
e
i
,
a
n
d
W
.
N
u
o
,
“
T
h
e
c
o
n
s
t
ru
c
t
i
o
n
o
f
a
c
l
o
u
d
c
o
m
p
u
t
i
n
g
-
b
a
s
e
d
i
n
t
e
l
l
i
g
e
n
t
p
e
n
s
i
o
n
s
e
rv
i
c
e
p
l
a
t
fo
rm
,
”
2
0
1
8
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
S
m
a
r
t
G
r
i
d
a
n
d
E
l
e
c
t
r
i
c
a
l
A
u
t
o
m
a
t
i
o
n
(ICS
G
E
A
)
,
p
p
.
2
9
5
–
2
9
8
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
0
9
/
ICS
G
E
A
.
2
0
1
8
.
0
0
0
8
0
.
[1
4
]
S
.
K
a
b
a
d
e
,
A
.
S
h
a
rm
a
,
a
n
d
A
.
K
a
g
a
l
k
a
r,
“
S
e
c
u
ri
n
g
p
e
n
s
i
o
n
s
y
s
t
e
m
s
w
i
t
h
A
I
-
d
ri
v
e
n
ri
s
k
a
n
a
l
y
t
i
c
s
a
n
d
c
l
o
u
d
-
n
a
t
i
v
e
m
a
c
h
i
n
e
l
e
a
rn
i
n
g
a
rc
h
i
t
e
c
t
u
re
s
,
”
In
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
m
e
r
g
i
n
g
R
e
s
e
a
r
c
h
i
n
E
n
g
i
n
e
e
r
i
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
5
,
n
o
.
2
,
p
p
.
5
2
–
6
4
,
2
0
2
4
,
d
o
i
:
1
0
.
6
3
2
8
2
/
3
0
5
0
-
9
2
2
X
.
IJ
E
RE
T
-
V
5
I2
P
1
0
6
.
[1
5
]
Be
n
e
fi
t
s
a
n
d
P
e
n
s
i
o
n
s
M
o
n
i
t
o
r,
“
H
o
w
A
I
i
s
t
ra
n
s
fo
rm
i
n
g
p
e
n
s
i
o
n
m
a
n
a
g
e
m
e
n
t
,
b
e
n
e
fi
t
s
a
n
d
p
e
n
s
i
o
n
s
m
o
n
i
t
o
r,
”
Be
n
e
fi
t
s
a
n
d
P
e
n
s
i
o
n
s
M
o
n
i
t
o
r.
[O
n
l
i
n
e
].
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
b
e
n
e
fi
t
s
a
n
d
p
e
n
s
i
o
n
s
m
o
n
i
t
o
r.
c
o
m
/
n
e
w
s
/
i
n
d
u
s
t
ry
-
n
e
w
s
/
h
o
w
-
ai
-
is
-
t
ra
n
s
fo
r
m
i
n
g
-
p
e
n
s
i
o
n
-
m
a
n
a
g
e
m
e
n
t
/
3
8
0
6
5
5
[1
6
]
H
.
M
a
rk
o
w
i
t
z
,
“
P
o
rt
fo
l
i
o
s
e
l
e
c
t
i
o
n
,
”
T
h
e
J
o
u
r
n
a
l
o
f
F
i
n
a
n
c
e
,
v
o
l
.
7
,
n
o
.
1
,
p
p
.
7
7
–
9
1
,
M
a
r.
1
9
5
2
,
d
o
i
:
1
0
.
1
1
1
1
/
j
.
1
5
4
0
-
6
2
6
1
.
1
9
5
2
.
t
b
0
1
5
2
5
.
x
.
[1
7
]
V
.
S
.
Bh
a
ra
t
h
i
,
D
.
P
ra
m
o
d
,
a
n
d
R.
Ra
m
a
n
,
“
In
t
e
n
t
i
o
n
t
o
u
s
e
a
rt
i
fi
c
i
a
l
i
n
t
e
l
l
i
g
e
n
c
e
s
e
rv
i
c
e
s
i
n
fi
n
a
n
c
i
a
l
i
n
v
e
s
t
m
e
n
t
d
e
c
i
s
i
o
n
s
,
”
2
0
2
2
In
t
e
r
n
a
t
i
o
n
a
l
Co
n
f
e
r
e
n
c
e
o
n
D
e
c
i
s
i
o
n
A
i
d
S
c
i
e
n
c
e
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
(D
A
S
A
)
,
v
o
l
.
1
8
,
n
o
.
2
,
p
p
.
5
6
0
–
5
6
6
,
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
D
A
S
A
5
4
6
5
8
.
2
0
2
2
.
9
7
6
5
1
8
3
.
[1
8
]
B.
G
.
M
a
l
k
i
e
l
a
n
d
E
.
F
.
F
a
m
a
,
“
E
ff
i
c
i
e
n
t
c
a
p
i
t
a
l
m
a
rk
e
t
s
:
a
re
v
i
e
w
o
f
t
h
e
o
ry
a
n
d
e
m
p
i
ri
c
a
l
w
o
rk
,
”
T
h
e
J
o
u
r
n
a
l
o
f
F
i
n
a
n
c
e
,
v
o
l
.
2
5
,
n
o
.
2
,
p
p
.
3
8
3
–
4
1
7
,
M
a
y
1
9
7
0
,
d
o
i
:
1
0
.
1
1
1
1
/
j
.
1
5
4
0
-
6
2
6
1
.
1
9
7
0
.
t
b
0
0
5
1
8
.
x
.
[1
9
]
T
.
F
i
s
c
h
e
r
a
n
d
C.
K
ra
u
s
s
,
“
D
e
e
p
l
e
a
rn
i
n
g
w
i
t
h
l
o
n
g
s
h
o
rt
-
t
e
rm
m
e
m
o
ry
n
e
t
w
o
rk
s
fo
r
fi
n
a
n
c
i
a
l
m
a
rk
e
t
p
re
d
i
c
t
i
o
n
s
,
”
E
u
r
o
p
e
a
n
J
o
u
r
n
a
l
o
f
O
p
e
r
a
t
i
o
n
a
l
R
e
s
e
a
r
c
h
,
v
o
l
.
2
7
0
,
n
o
.
2
,
p
p
.
6
5
4
–
6
6
9
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
j
o
r.
2
0
1
7
.
1
1
.
0
5
4
.
[2
0
]
I.
G
o
o
d
fe
l
l
o
w
,
Y
.
Be
n
g
i
o
,
a
n
d
A
.
Co
u
rv
i
l
l
e
,
D
e
e
p
l
e
a
r
n
i
n
g
.
Ca
m
b
ri
d
g
e
,
U
n
i
t
e
d
S
t
a
t
e
s
:
M
IT
P
re
s
s
,
2
0
1
6
.
[2
1
]
CM
S
,
“
M
a
c
h
i
n
e
l
e
a
rn
i
n
g
a
p
p
l
i
c
a
t
i
o
n
s
i
n
F
i
n
a
n
c
e
,
”
CM
S
:
Ca
n
a
d
i
a
n
M
a
t
h
e
m
a
t
i
c
a
l
S
o
c
i
e
t
y
,
2
0
1
7
.
[2
2
]
A
.
E
.
K
h
a
n
d
a
n
i
,
A
.
J
.
K
i
m
,
a
n
d
A
.
W
.
L
o
,
“
Co
n
s
u
m
e
r
c
re
d
i
t
-
ri
s
k
m
o
d
e
l
s
v
i
a
m
a
c
h
i
n
e
-
l
e
a
rn
i
n
g
a
l
g
o
ri
t
h
m
s
,
”
J
o
u
r
n
a
l
o
f
B
a
n
k
i
n
g
&
F
i
n
a
n
c
e
,
v
o
l
.
3
4
,
n
o
.
1
1
,
p
p
.
2
7
6
7
–
2
7
8
7
,
N
o
v
.
2
0
1
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
b
a
n
k
fi
n
.
2
0
1
0
.
0
6
.
0
0
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Com
pu
t
S
c
i
Inf
T
e
c
h
nol
IS
S
N
:
2722
-
3221
Cl
oud
-
b
as
e
d
pr
e
d
i
c
t
i
v
e
ana
l
y
t
i
c
s
f
or
p
e
ns
i
on
f
u
nd
pe
r
f
or
m
an
c
e
opt
i
m
i
z
a
t
i
on
(
B
e
au
t
y
G
ar
aba
)
55
[2
3
]
P
.
J
o
ri
o
n
,
“
V
a
l
u
e
a
t
ri
s
k
:
t
h
e
n
e
w
b
e
n
c
h
m
a
rk
fo
r
m
a
n
a
g
i
n
g
fi
n
a
n
c
i
a
l
ri
s
k
,
”
J
o
u
r
n
a
l
o
f
Ch
e
m
i
c
a
l
In
f
o
r
m
a
t
i
o
n
a
n
d
M
o
d
e
l
i
n
g
,
v
o
l
.
5
3
,
n
o
.
9
,
p
p
.
1
–
8
2
0
,
2
0
0
7
.
[2
4
]
P
.
W
i
r
a
w
a
n
,
“
L
e
v
e
ra
g
i
n
g
p
re
d
i
c
t
i
v
e
a
n
a
l
y
t
i
c
s
i
n
fi
n
a
n
c
i
n
g
d
e
c
i
s
i
o
n
-
m
a
k
i
n
g
fo
r
c
o
m
p
a
ra
t
i
v
e
a
n
a
l
y
s
i
s
a
n
d
o
p
t
i
m
i
z
a
t
i
o
n
,
”
A
d
v
a
n
c
e
s
i
n
M
a
n
a
g
e
m
e
n
t
&
F
i
n
a
n
c
i
a
l
R
e
p
o
r
t
i
n
g
,
v
o
l
.
1
,
n
o
.
3
,
S
e
p
.
2
0
2
3
,
d
o
i
:
1
0
.
6
0
0
7
9
/
a
m
fr.
v
1
i
3
.
2
0
9
.
[2
5
]
E
.
W
.
T
.
N
g
a
i
,
Y
.
H
u
,
Y
.
H
.
W
o
n
g
,
Y
.
C
h
e
n
,
a
n
d
X
.
S
u
n
,
“
T
h
e
a
p
p
l
i
c
a
t
i
o
n
o
f
d
a
t
a
m
i
n
i
n
g
t
e
c
h
n
i
q
u
e
s
i
n
fi
n
a
n
c
i
a
l
fra
u
d
d
e
t
e
c
t
i
o
n
:
a
c
l
a
s
s
i
fi
c
a
t
i
o
n
fra
m
e
w
o
rk
a
n
d
a
n
a
c
a
d
e
m
i
c
re
v
i
e
w
o
f
l
i
t
e
ra
t
u
re
,
”
D
e
c
i
s
i
o
n
S
u
p
p
o
r
t
S
y
s
t
e
m
s
,
v
o
l
.
5
0
,
n
o
.
3
,
p
p
.
5
5
9
–
5
6
9
,
F
e
b
.
2
0
1
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
d
s
s
.
2
0
1
0
.
0
8
.
0
0
6
.
BI
O
G
R
A
P
H
I
ES
O
F
A
U
T
H
O
R
S
B
e
au
t
y
G
a
r
ab
a
i
s
c
u
r
r
e
n
t
l
y
pu
r
s
ui
n
g
a
m
a
s
t
e
r
’
s
de
gr
e
e
i
n
c
l
ou
d
c
om
pu
t
i
n
g
w
i
t
h
t
he
H
a
r
a
r
e
I
ns
t
i
t
u
t
e
o
f
T
e
c
hn
ol
o
gy
(
H
I
T
)
.
S
h
e
ho
l
ds
a
B
a
c
he
l
or
of
T
e
c
hno
l
og
y
de
g
r
e
e
i
n
I
nf
or
m
a
t
i
on
T
e
c
hno
l
ogy
f
r
o
m
t
h
e
s
a
m
e
i
ns
t
i
t
u
t
i
on.
B
e
a
u
t
y
ha
s
s
e
ve
r
a
l
y
e
a
r
s
’
e
xpe
r
i
e
nc
e
w
i
t
h
a
pe
n
s
i
on
f
und
,
w
he
r
e
s
he
ha
s
l
e
d
t
he
i
m
p
l
e
m
e
n
t
a
t
i
on
of
t
h
e
c
o
m
p
a
ny’
s
d
i
g
i
t
a
l
s
t
r
a
t
e
gy.
S
he
i
s
a
l
s
o
a
c
e
r
t
i
f
i
e
d
da
t
a
pr
o
t
e
c
t
i
on
of
f
i
c
e
r
.
H
e
r
o
t
h
e
r
i
n
t
e
r
e
s
t
s
a
r
e
i
n
m
e
nt
o
r
i
ng
g
i
r
l
s
a
n
d
w
o
m
e
n
i
n
I
C
T
.
S
h
e
c
ur
r
e
n
t
l
y
s
e
r
ve
s
on
t
he
W
o
m
e
n
i
n
I
C
T
Z
i
m
ba
bw
e
C
om
m
i
t
t
e
e
.
S
h
e
c
a
n
be
c
ont
a
c
t
e
d
a
t
e
m
a
i
l
:
h230
787
a
@h
i
t
.
a
c
.
z
w
.
M
ai
n
f
o
r
d
M
u
t
an
d
avar
i
i
s
a
P
h
.
D
.
S
c
hol
a
r
a
t
S
R
M
I
S
T
U
ni
ve
r
s
i
t
y
,
I
ndi
a
,
a
n
d
a
L
e
c
t
ur
e
r
a
nd
P
os
t
gr
a
dua
t
e
S
t
udi
e
s
C
o
or
d
i
n
a
t
or
a
t
t
he
H
a
r
a
r
e
I
ns
t
i
t
ut
e
of
T
e
c
hn
ol
ogy
(
H
I
T
)
,
Z
i
m
ba
bw
e
.
W
i
t
h
a
dva
nc
e
d
d
e
g
r
e
e
s
i
n
C
o
m
pu
t
e
r
S
c
i
e
nc
e
a
n
d
S
t
r
a
t
e
gy
&
I
nno
va
t
i
o
n,
h
i
s
r
e
s
e
a
r
c
h
s
p
a
ns
d
a
t
a
a
na
l
yt
i
c
s
,
c
ybe
r
s
e
c
u
r
i
t
y
,
I
oT
,
A
I
,
a
n
d
c
l
o
ud
c
o
m
put
i
ng
.
H
e
i
s
a
m
e
m
be
r
of
H
I
T
’
s
c
ybe
r
s
e
c
u
r
i
t
y
a
nd
A
I
r
e
s
e
a
r
c
h
g
r
oup
s
a
nd
a
c
t
i
v
e
l
y
c
on
t
r
i
bu
t
e
s
t
o
na
t
i
o
na
l
I
C
T
s
t
a
n
da
r
ds
t
hr
ough
t
he
S
t
a
nda
r
ds
A
s
s
o
c
i
a
t
i
on
o
f
Z
i
m
b
a
bw
e
.
He
ha
s
pub
l
i
s
h
e
d
w
i
de
l
y
on
t
op
i
c
s
s
uc
h
a
s
da
t
a
l
os
s
pr
e
ve
n
t
i
on
s
ys
t
e
m
s
,
d
i
g
i
t
a
l
l
e
a
r
ni
n
g
i
nf
r
a
s
t
r
uc
t
ur
e
,
a
nd
e
-
he
a
l
t
h
s
e
c
ur
i
t
y.
H
i
s
w
or
k
br
i
dge
s
a
c
a
de
m
i
c
r
e
s
e
a
r
c
h
w
i
t
h
i
nd
us
t
r
y
a
pp
l
i
c
a
t
i
ons
,
f
oc
us
i
ng
o
n
pr
a
c
t
i
c
a
l
d
i
g
i
t
a
l
s
o
l
u
t
i
o
ns
f
or
e
du
c
a
t
i
on,
t
e
l
e
c
o
m
m
un
i
c
a
t
i
on
s
,
a
nd
he
a
l
t
hc
a
r
e
i
n
Z
i
m
ba
bw
e
.
H
e
i
s
a
l
s
o
i
n
vol
ve
d
i
n
c
ur
r
i
c
ul
u
m
de
v
e
l
o
p
m
e
n
t
,
pos
t
g
r
a
du
a
t
e
s
up
e
r
vi
s
i
o
n,
a
nd
bu
i
l
di
ng
a
c
a
de
m
i
c
-
i
ndu
s
t
r
y
pa
r
t
ne
r
s
h
i
p
s
.
H
e
c
a
n
be
c
o
nt
a
c
t
e
d
a
t
e
m
a
i
l
:
m
m
u
t
a
n
da
v
a
r
i
@h
i
t
.
a
c
.
z
w
.
Je
r
i
t
a
C
h
i
b
h
ab
h
a
i
s
a
l
e
c
t
u
r
e
r
a
t
t
h
e
H
a
r
a
r
e
I
ns
t
i
t
u
t
e
o
f
T
e
c
hn
ol
ogy
(
H
I
T
)
i
n
t
he
D
e
pa
r
t
m
e
n
t
of
I
n
f
o
r
m
a
t
i
on
T
e
c
hno
l
og
y.
W
i
t
h
a
s
t
r
o
ng
a
c
a
de
m
i
c
f
ounda
t
i
on
t
h
a
t
i
n
c
l
ude
s
a
ba
c
h
e
l
o
r
’
s
de
g
r
e
e
i
n
i
n
f
o
r
m
a
t
i
o
n
s
ys
t
e
m
s
,
a
m
a
s
t
e
r
’
s
i
n
da
t
a
s
c
i
e
n
c
e
a
nd
i
nf
o
r
m
a
t
i
c
s
,
a
nd
a
M
a
s
t
e
r
o
f
T
e
c
hno
l
og
y
i
n
s
of
t
w
a
r
e
e
ng
i
ne
e
r
i
ng
,
s
h
e
i
s
pa
s
s
i
on
a
t
e
a
bou
t
br
i
d
gi
n
g
t
h
e
o
r
e
t
i
c
a
l
c
onc
e
pt
s
w
i
t
h
p
r
a
c
t
i
c
a
l
a
p
pl
i
c
a
t
i
ons
i
n
t
he
f
i
e
l
d
s
of
da
t
a
s
c
i
e
nc
e
a
nd
s
of
t
w
a
r
e
e
n
gi
n
e
e
r
i
ng.
S
he
s
pe
c
i
a
l
i
z
e
s
i
n
d
a
t
a
a
na
l
y
t
i
c
s
,
m
a
c
hi
ne
l
e
a
r
n
i
ng
,
a
nd
s
o
f
t
w
a
r
e
e
ngi
ne
e
r
i
n
g,
w
i
t
h
a
p
a
r
t
i
c
u
l
a
r
f
oc
u
s
o
n
p
r
a
c
t
i
c
a
l
a
pp
l
i
c
a
t
i
ons
t
ha
t
a
ddr
e
s
s
c
on
t
e
m
po
r
a
r
y
c
ha
l
l
e
ng
e
s
.
T
he
i
r
r
e
s
e
a
r
c
h
e
xp
l
o
r
e
s
i
nno
va
t
i
v
e
m
e
t
ho
ds
f
o
r
l
e
v
e
r
a
gi
n
g
b
i
g
da
t
a
a
c
r
os
s
v
a
r
i
ou
s
s
e
c
t
or
s
,
d
e
v
e
l
o
pi
n
g
u
s
e
r
-
f
r
i
e
n
dl
y
s
of
t
w
a
r
e
s
o
l
u
t
i
ons
,
a
nd
e
nha
nc
i
ng
de
c
i
s
i
o
n
-
m
a
ki
ng
t
h
r
ou
gh
da
t
a
-
dr
i
ve
n
i
ns
i
gh
t
s
.
S
h
e
c
a
n
b
e
c
ont
a
c
t
e
d
a
t
e
m
a
i
l
:
j
c
hi
bha
b
ha
@
hi
t
.
a
c
.
z
w
.
Evaluation Warning : The document was created with Spire.PDF for Python.