C
om
p
u
t
e
r
S
c
ie
n
c
e
an
d
I
n
f
or
m
at
io
n
T
e
c
h
n
ol
ogi
e
s
V
ol
.
7
, N
o.
1
,
M
a
r
c
h
2026
, pp.
56
~
65
I
S
S
N
:
2722
-
3221
,
D
O
I
:
10.11591/cs
it
.
v
7
i
1
.
pp
56
-
65
56
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ia
e
s
pr
ime
.c
om
/i
nde
x
.php/c
s
it
O
p
t
i
m
i
z
i
n
g i
n
t
e
r
c
on
n
e
c
t
i
on
c
al
l
r
ou
t
i
n
g:
a m
ac
h
i
n
e
l
e
ar
n
i
n
g
ap
p
r
oac
h
f
or
c
ost
an
d
q
u
al
i
t
y e
f
f
i
c
i
e
n
c
y
I
vy A
n
e
s
u
M
u
d
ar
i,
M
ai
n
f
or
d
M
u
t
an
d
avar
i,
K
e
n
n
e
t
h
C
h
iwor
e
r
a
S
c
hool
of
I
nf
or
m
a
t
i
on S
c
i
e
nc
e
a
nd T
e
c
hnol
ogy, H
a
r
a
r
e
I
ns
t
i
t
ut
e
of
T
e
c
hnol
ogy
, H
a
r
a
r
e
, Z
i
m
ba
bw
e
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
J
un 8, 2025
R
e
vi
s
e
d
S
e
p 1, 2025
A
c
c
e
pt
e
d
N
ov 28, 2025
This
study
presents
the
design
and
development
of
an
automated
lea
st
cost
routing
(LCR)
model
for
telecommunications
interconnection
calls
using
machine
learning
.
Leveraging
a
random
forest
regressor
,
the
model
p
redicts
the
most
cost
-
effective
call
routing
path
based
on
pricing
and
n
etwork
latency
.
Trained
on
real
-
world
call
detail
records
(CDRs)
from
TelOne
Zimbabwe,
the
model
achieved
a
high
R²
score
of
0.851,
with
a
mean
absolut
e
error
(MAE)
of
$0.0482
per
minute.
Evaluation
results
demonstrate
an
averag
e
cost
reduction
of
46.75%
compared
to
tradition
al
r
outing
methods,
with
prediction
t
imes
under
0.1
seconds
and
latency
rem
aining
within
acceptable
thresholds.
This
work
provides
a
practical,
scalabl
e,
and
efficient
soluti
on
for
telecom.
operators
seeking
to
reduce
intercon
nection
costs
and
maintain
service
q
uality
through
intelligent
routing
auto
mation.
The
model
architecture
and
p
erformance
to
make
it
viable
for
integrat
ion
into real
-
time teleco
m
infrastruc
ture.
K
e
y
w
o
r
d
s
:
A
ut
om
a
te
d
C
a
ll
-
r
out
in
g
I
nt
e
r
c
onne
c
ti
on
L
e
a
s
t
c
os
t
r
out
in
g
M
a
c
hi
ne
l
e
a
r
ni
ng
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
I
vy A
ne
s
u M
uda
r
i
S
c
hool
of
I
nf
or
m
a
ti
on S
c
ie
nc
e
a
nd T
e
c
hnol
ogy, H
a
r
a
r
e
I
ns
ti
tu
te
of
T
e
c
hnol
ogy
H
a
r
a
r
e
, Z
im
ba
bw
e
E
m
a
il
:
h230942f
@
hi
t.
a
c
.z
w
1.
I
N
T
R
O
D
U
C
T
I
O
N
D
id
y
ou
k
no
w
th
a
t
t
e
l
e
c
om
m
uni
c
a
ti
on
s
c
om
p
a
n
ie
s
c
a
n
lo
s
e
m
i
ll
i
on
s
of
d
ol
l
a
r
s
a
y
e
a
r
b
e
c
a
u
s
e
in
t
e
r
c
on
ne
c
ti
on
c
a
l
l
s
a
r
e
r
out
e
d
p
oor
l
y
?
I
nt
e
r
c
onn
e
c
ti
on
c
a
l
ls
, w
hi
c
h go
th
r
ou
gh
t
w
o
or
m
or
e
t
e
l
e
c
om
ne
tw
or
k
s
,
a
r
e
b
e
c
om
i
ng
m
or
e
i
m
p
or
t
a
nt
to
th
e
gl
o
ba
l
t
e
l
e
c
om
m
uni
c
a
ti
on
s
in
f
r
a
s
tr
u
c
tu
r
e
.
A
s
in
t
e
r
na
ti
o
na
l
c
om
m
un
ic
a
ti
on
gr
ow
s
a
nd
d
ig
it
a
l
e
c
o
s
y
s
t
e
m
s
be
c
o
m
e
m
or
e
c
om
pl
e
x,
t
he
pr
ic
e
of
th
e
s
e
c
a
ll
s
c
a
n
c
ha
ng
e
a
lo
t
d
e
p
e
ndi
ng
on
w
h
e
r
e
t
h
e
c
a
ll
i
s
m
a
de
, g
ov
e
r
n
m
e
nt
r
ul
e
s
,
a
n
d
a
gr
e
e
m
e
nt
s
b
e
tw
e
e
n
op
e
r
a
to
r
s
[
1]
.
F
or
te
le
c
om
ope
r
a
to
r
s
,
m
a
in
ta
in
in
g
hi
gh
-
qua
li
ty
s
e
r
vi
c
e
w
hi
le
c
ont
r
ol
li
ng
ope
r
a
ti
ona
l
e
xpe
ndi
tu
r
e
s
r
e
m
a
in
s
a
pe
r
s
is
te
nt
c
h
a
ll
e
nge
.
A
c
r
it
ic
a
l
c
om
pone
nt
of
th
i
s
is
le
a
s
t
c
o
s
t
r
out
in
g
(
L
C
R
)
—
a
m
e
th
od
th
a
t
s
e
le
c
ts
th
e
m
os
t
c
os
t
-
e
f
f
e
c
ti
ve
pa
th
f
or
c
a
ll
te
r
m
in
a
ti
on
[
2]
.
T
r
a
di
ti
ona
ll
y,
r
out
in
g
de
c
is
io
ns
a
r
e
m
a
de
us
in
g
s
ta
ti
c
r
out
in
g
ta
bl
e
s
or
f
ix
e
d
c
onf
ig
ur
a
ti
ons
.
I
n
th
is
ba
s
e
li
ne
,
r
out
in
g
de
c
is
io
ns
a
r
e
m
a
d
e
us
in
g
s
ta
ti
c
r
out
in
g
ta
bl
e
s
,
c
onf
ig
ur
e
d
m
a
nua
ll
y
ba
s
e
d
on
th
e
lo
w
e
s
t
publ
i
s
he
d
ta
r
i
f
f
s
a
va
il
a
bl
e
a
t
th
e
ti
m
e
of
s
e
tu
p.
T
he
s
e
ta
bl
e
s
ty
pi
c
a
ll
y
r
e
m
a
in
f
ix
e
d
unt
il
upda
te
d
by
n
e
twor
k
e
ngi
ne
e
r
s
,
of
te
n
on
a
w
e
e
kl
y
or
m
ont
hl
y
ba
s
i
s
.
T
h
e
s
e
a
r
e
not
s
ui
ta
bl
e
f
or
e
nvi
r
onm
e
nt
s
w
he
r
e
pr
ic
e
s
a
nd
tr
a
f
f
ic
c
ondi
ti
on
s
c
ha
nge
[
3]
.
T
hi
s
of
te
n
r
e
s
ul
ts
in
te
le
c
om
ne
twor
ks
c
ont
in
ui
ng t
o us
e
e
xpe
n
s
iv
e
r
out
e
s
e
v
e
n w
he
n c
he
a
pe
r
, m
or
e
e
f
f
ic
ie
nt
a
lt
e
r
na
ti
ve
s
e
xi
s
t.
I
n
tr
a
di
ti
ona
l
L
C
R
,
r
out
in
g
de
c
i
s
io
ns
a
r
e
m
a
d
e
us
in
g
s
ta
ti
c
r
out
in
g
ta
bl
e
s
s
to
r
e
d
in
th
e
s
w
it
c
h
or
s
e
s
s
io
n
bor
de
r
c
ont
r
ol
le
r
(
S
B
C
)
.
E
a
c
h
r
ou
ti
ng
ta
bl
e
e
nt
r
y
c
ons
is
ts
of
:
i)
pr
e
f
ix
m
a
tc
h:
t
he
s
ys
te
m
c
he
c
ks
th
e
c
a
ll
e
d
num
be
r
(
di
a
le
d
num
be
r
id
e
nt
if
ic
a
ti
on
s
e
r
vi
c
e
)
a
nd
m
a
t
c
he
s
it
a
ga
in
s
t
th
e
lo
nge
s
t
d
e
s
ti
na
ti
on
pr
e
f
ix
(
e
.g.,
+
44)
in
th
e
ta
bl
e
;
ii
)
c
a
r
r
ie
r
li
s
t
pe
r
pr
e
f
ix
:
f
or
e
a
c
h
pr
e
f
ix
,
one
or
m
or
e
c
a
r
r
ie
r
s
a
r
e
li
s
te
d
in
pr
io
r
it
y
or
de
r
,
ty
pi
c
a
ll
y
s
or
te
d
by
lo
w
e
s
t
a
dve
r
ti
s
e
d
ta
r
if
f
;
a
nd
ii
i)
r
out
e
s
e
le
c
ti
on:
t
he
s
w
it
c
h s
e
le
c
t
s
th
e
f
ir
s
t
a
va
il
a
bl
e
Evaluation Warning : The document was created with Spire.PDF for Python.
C
om
put
S
c
i
I
nf
T
e
c
hnol
I
S
S
N
:
2722
-
3221
O
pt
imi
z
in
g i
nt
e
r
c
onne
c
ti
on c
al
l
r
out
in
g:
a m
ac
hi
ne
l
e
ar
ni
ng ap
pr
oac
h f
or
c
o
s
t
and
…
(
I
v
y
A
ne
s
u M
udar
i
)
57
c
a
r
r
ie
r
t
r
unk
i
n t
he
l
is
t
a
nd
r
out
e
s
t
he
c
a
ll
.
I
f
t
ha
t
c
a
r
r
ie
r
i
s
una
v
a
il
a
bl
e
(
c
onge
s
ti
on, no r
e
s
pons
e
)
, i
t
f
a
ll
s
ba
c
k
to
t
he
ne
xt
c
a
r
r
ie
r
i
n t
he
s
e
que
nc
e
.
U
pda
te
s
a
r
e
done
m
a
nua
ll
y
w
e
e
kl
y.
T
hi
s
m
e
a
n
s
r
out
in
g
de
c
is
io
ns
a
lwa
y
s
f
ol
lo
w
th
e
c
he
a
pe
s
t
publ
is
he
d
ta
r
if
f
a
t
th
e
ti
m
e
of
upda
te
,
w
it
hout
c
ons
id
e
r
in
g
r
e
a
l
-
ti
m
e
qua
li
ty
s
ig
na
ls
.
L
a
te
nc
y,
or
th
e
ti
m
e
it
ta
ke
s
f
or
da
ta
to
be
s
e
nt
,
is
a
not
he
r
im
por
ta
nt
f
a
c
to
r
in
us
e
r
s
a
ti
s
f
a
c
ti
on,
a
lo
ng
w
it
h
c
os
t.
L
ong
de
la
ys
c
a
n
m
a
ke
c
a
ll
s
s
ound
w
or
s
e
,
w
hi
c
h
c
a
n
m
a
ke
us
e
r
s
unha
ppy
a
nd
c
a
us
e
th
e
m
to
le
a
ve
[
4]
.
A
s
a
r
e
s
ul
t,
m
or
e
a
nd
m
or
e
pe
opl
e
a
r
e
in
te
r
e
s
te
d
in
s
m
a
r
t
r
out
in
g
s
tr
a
te
gi
e
s
th
a
t
f
in
d
a
ba
la
nc
e
be
twe
e
n
c
os
t
a
nd
qua
li
ty
of
s
e
r
vi
c
e
(
Q
oS
)
.
W
hi
le
th
e
tr
a
di
ti
ona
l
m
e
th
od
is
s
tr
a
ig
ht
f
or
w
a
r
d,
it
s
uf
f
e
r
s
f
r
om
s
e
ve
r
a
l
li
m
it
a
ti
ons
:
i)
la
c
k
of
a
da
pt
a
bi
li
ty
:
r
out
e
s
r
e
m
a
in
unc
ha
nge
d
e
ve
n
w
he
n
c
he
a
pe
r
a
l
te
r
na
ti
ve
s
be
c
om
e
a
va
il
a
bl
e
due
to
dyna
m
ic
pr
ic
in
g
or
tr
a
f
f
ic
f
lu
c
tu
a
ti
ons
;
ii
)
c
onf
ig
u
r
a
ti
on
ove
r
he
a
d:
m
a
nua
l
upda
te
s
in
tr
oduc
e
de
la
y
s
a
nd
a
r
e
pr
one
to
hum
a
n
e
r
r
or
;
a
nd
ii
i)
qua
li
ty
bl
in
dne
s
s
:
r
out
in
g
ta
bl
e
s
pr
io
r
it
iz
e
ta
r
if
f
r
a
te
s
w
it
hout
f
a
c
to
r
in
g
in
la
te
nc
y,
ji
tt
e
r
,
or
pa
c
ke
t
lo
s
s
, w
hi
c
h
c
a
n l
e
a
d t
o de
gr
a
de
d u
s
e
r
e
xpe
r
ie
nc
e
.
P
r
e
vi
ous
r
e
s
e
a
r
c
h
ha
s
lo
ok
e
d
a
t
L
C
R
a
nd
tr
a
f
f
ic
m
a
na
ge
m
e
nt
s
e
pa
r
a
te
ly
,
but
not
m
a
ny
ha
ve
c
om
bi
ne
d
r
e
a
l
-
w
or
ld
in
te
r
c
onne
c
ti
on
da
ta
w
it
h
m
a
c
hi
ne
le
a
r
ni
ng
to
c
r
e
a
te
r
out
in
g
m
ode
ls
th
a
t
c
a
n
pr
e
di
c
t
c
os
ts
.
T
hi
s
la
c
k
of
knowle
dge
in
bot
h
th
e
a
c
a
de
m
ic
a
nd
bu
s
in
e
s
s
w
or
ld
s
ope
ns
up
ne
w
pos
s
ib
il
it
ie
s
f
or
r
out
in
g
opt
im
iz
a
ti
on.
T
hi
s
pa
pe
r
a
ddr
e
s
s
e
s
th
a
t
ga
p
by
pr
opos
i
ng
a
m
a
c
hi
ne
l
e
a
r
ni
ng
-
ba
s
e
d
a
ut
om
a
te
d
L
C
R
m
ode
l,
le
ve
r
a
gi
ng
a
r
a
ndom
f
o
r
e
s
t
r
e
gr
e
s
s
or
.
T
he
m
ode
l
ut
il
iz
e
s
hi
s
to
r
ic
a
l
c
a
ll
r
e
c
or
ds
,
ope
r
a
to
r
pr
ic
in
g
li
s
ts
,
a
nd
ne
twor
k
c
onf
ig
ur
a
ti
ons
to
pr
e
di
c
t
th
e
m
os
t
c
os
t
-
e
f
f
e
c
ti
ve
in
te
r
c
onne
c
ti
on
r
out
e
s
a
nd
a
ls
o
c
ons
id
e
r
s
c
hoos
in
g r
out
e
s
w
it
h m
in
im
a
l
la
te
nc
y.
T
hi
s
pa
pe
r
'
s
r
e
m
a
in
in
g
s
e
c
ti
ons
a
r
e
or
ga
ni
z
e
d
a
s
f
ol
lo
w
s
:
s
e
c
ti
on
2
pr
ovi
de
s
a
r
e
vi
e
w
of
r
e
le
va
nt
r
e
s
e
a
r
c
h
on
m
a
c
hi
ne
le
a
r
ni
ng
a
ppl
ic
a
ti
on
s
in
t
e
le
c
om
m
u
ni
c
a
ti
ons
r
out
in
g
a
nd
L
C
R
.
T
he
m
od
e
l'
s
im
pl
e
m
e
nt
a
ti
on
a
nd
tr
a
in
in
g
pr
oc
e
dur
e
a
r
e
de
s
c
r
ib
e
d
in
de
t
a
i
l
in
s
e
c
ti
on
3
.
T
he
e
xpe
r
im
e
nt
a
l
r
e
s
ul
ts
a
nd
e
va
lu
a
ti
on
m
e
tr
ic
s
a
r
e
pr
e
s
e
nt
e
d
in
s
e
c
ti
on
4
.
S
e
c
ti
on
5
c
on
c
lu
de
s
th
e
s
tu
dy
a
nd
di
s
c
us
s
e
s
it
s
im
pl
ic
a
ti
ons
f
or
f
ut
ur
e
r
e
s
e
a
r
c
h a
nd r
e
a
l
-
w
or
ld
a
ppl
ic
a
ti
ons
.
2.
R
E
L
A
T
E
D
WORK
T
he
a
s
s
oc
ia
te
d
r
e
s
e
a
r
c
h
in
di
c
a
te
s
th
a
t
th
e
u
s
e
of
m
a
c
hi
ne
le
a
r
ni
ng
in
r
out
in
g
m
e
th
ods
is
a
r
a
pi
dl
y
e
xpa
ndi
ng
in
nova
ti
on
.
I
t
ha
s
be
e
n
pr
ove
n
to
be
bot
h
e
f
f
e
c
ti
ve
a
nd
hi
ghl
y
pr
e
c
is
e
.
V
a
r
io
us
r
e
s
e
a
r
c
he
r
s
ha
ve
a
ls
o pr
e
s
e
nt
e
d
s
uppor
ti
ng t
he
or
ie
s
t
ha
t
r
e
in
f
or
c
e
t
he
s
e
f
in
di
ngs
.
T
r
a
di
ti
ona
ll
y,
a
r
out
in
g
m
e
th
od'
s
r
out
in
g
ta
bl
e
,
w
hi
c
h
in
c
lu
de
d a
li
s
t
of
pot
e
nt
ia
l
r
out
e
s
a
nd
th
e
r
ul
e
s
f
or
c
hoos
in
g
th
e
be
s
t
one
f
or
a
pa
r
ti
c
ul
a
r
c
onne
c
ti
on
o
r
ba
ndw
id
th
-
a
ll
oc
a
ti
on
r
e
que
s
t,
s
e
r
ve
d
a
s
it
s
f
unda
m
e
nt
a
l
to
ol
.
T
he
s
e
r
ul
e
s
w
e
r
e
u
s
e
d
by
th
e
or
ig
in
a
ti
ng
nod
e
,
w
hi
c
h
w
a
s
in
c
ha
r
ge
of
m
a
ki
ng
r
e
que
s
t
s
,
to
f
in
d
a
n
a
dm
is
s
ib
le
r
out
e
th
a
t
s
a
ti
s
f
ie
d
th
e
r
e
que
s
t'
s
s
pe
c
if
ic
a
ti
ons
[
5]
.
T
he
r
out
in
g
ta
bl
e
s
w
e
r
e
c
onf
ig
ur
e
d
m
a
nua
ll
y,
s
o
if
th
e
r
e
w
a
s
hum
a
n
e
r
r
or
in
c
onf
ig
ur
a
ti
on,
it
c
ou
ld
im
pa
c
t
th
e
in
c
or
r
e
c
t
or
in
e
f
f
ic
ie
nt
r
out
in
g.
M
a
nua
l
upda
te
s
w
e
r
e
a
l
s
o
c
os
tl
y
a
nd
ti
m
e
-
c
ons
um
in
g,
a
nd
th
e
m
e
th
od
la
c
ke
d
a
d
a
pt
a
bi
li
ty
to
dyna
m
ic
ne
twor
k c
ondi
ti
ons
.
A
m
in
e
t
al
.
[
6]
lo
oke
d
a
t
how
th
r
e
e
m
a
in
ty
pe
s
of
m
a
c
hi
n
e
le
a
r
ni
ng
—
r
e
in
f
or
c
e
m
e
nt
le
a
r
ni
ng,
uns
upe
r
vi
s
e
d
le
a
r
ni
ng,
a
nd
s
upe
r
vi
s
e
d
le
a
r
ni
ng
—
c
a
n
h
e
lp
im
pr
ove
r
out
in
g
in
s
of
twa
r
e
-
de
f
in
e
d
ne
twor
ki
ng
(
S
D
N
)
.
T
he
ir
a
na
ly
s
is
le
d
th
e
m
to
th
e
c
onc
lu
s
io
n
th
a
t
th
e
r
e
ha
s
be
e
n
a
not
a
bl
e
in
c
r
e
a
s
e
in
th
e
a
ppl
ic
a
ti
on
of
m
a
c
hi
ne
le
a
r
ni
ng,
a
nd
m
or
e
e
s
p
e
c
ia
ll
y
de
e
p
r
e
in
f
or
c
e
m
e
nt
l
e
a
r
ni
ng,
f
or
S
D
N
r
out
in
g
opt
im
iz
a
ti
on.
T
he
y
c
r
e
di
te
d t
he
us
e
of
m
a
c
hi
ne
l
e
a
r
ni
ng f
or
r
out
in
g e
f
f
ic
ie
nc
y.
M
a
c
hi
n
e
l
e
a
r
ni
ng
te
c
hno
lo
g
y
w
a
s
in
tr
od
uc
e
d
in
D
o
u
e
t
al
.
[
7]
.
T
hi
s
w
a
s
f
ol
lo
w
e
d
b
y
r
o
ut
i
ng
a
lg
or
it
hm
s
th
a
t
m
a
d
e
u
s
e
of
di
f
f
e
r
e
nt
m
a
c
hi
n
e
l
e
a
r
ni
n
g
te
c
h
nol
o
gi
e
s
.
S
a
nt
os
e
t
al
.
[
8]
d
e
m
o
n
s
tr
a
t
e
d
th
e
s
i
gni
f
ic
a
n
t
p
ot
e
nt
i
a
l
of
m
a
c
h
in
e
l
e
a
r
ni
ng
te
c
h
ni
q
ue
s
in
im
pr
ov
i
ng
n
e
t
w
or
ki
n
g
t
a
s
k
s
.
T
h
e
a
bi
li
t
y
to
l
e
a
r
n
f
r
om
hi
s
to
r
ic
a
l
da
t
a
s
e
t
s
a
nd a
p
pl
y t
h
a
t
k
no
w
le
dg
e
c
on
tr
ib
ut
e
d
t
o
b
e
tt
e
r
out
c
om
e
s
.
F
or
f
u
tu
r
e
w
or
k
,
th
e
s
tu
dy pl
a
nn
e
d
to
in
c
o
r
por
a
t
e
Q
oS
c
r
i
te
r
ia
in
t
o
th
e
de
c
i
s
i
on
-
m
a
k
in
g
pr
o
c
e
s
s
to
e
n
s
ur
e
a
pp
li
c
a
ti
o
n
-
s
p
e
c
i
f
i
c
r
e
qui
r
e
m
e
n
t
s
.
A
dd
it
i
on
a
ll
y,
th
e
r
o
ut
i
ng
s
tr
a
t
e
g
y
w
a
s
i
nt
e
n
de
d
to
b
e
e
xt
e
nd
e
d
to
ot
h
e
r
m
a
c
hi
ne
l
e
a
r
ni
ng
a
lg
or
it
hm
s
,
w
it
h
a
de
t
a
i
le
d
c
om
pa
r
i
s
on
a
m
o
ng
th
e
m
.
G
e
l
e
n
be
[
9]
a
ls
o i
nt
r
odu
c
e
d
t
h
e
s
u
bj
e
c
t
i
n n
e
t
w
or
k r
out
in
g
.
T
he
a
ut
hor
s
c
onc
e
nt
r
a
te
d
on
in
ve
s
ti
ga
ti
ng
r
e
in
f
or
c
e
d
le
a
r
ni
ng
-
ba
s
e
d
a
nd
s
upe
r
vi
s
e
d
le
a
r
ni
ng
-
ba
s
e
d
r
out
in
g
a
lg
or
it
hm
s
[
10]
,
[
11]
.
M
a
c
hi
ne
le
a
r
ni
ng
-
ba
s
e
d
r
out
in
g
m
e
th
ods
w
e
r
e
m
or
e
a
da
pt
a
bl
e
th
a
n
tr
a
di
ti
ona
l
r
out
in
g
a
lg
or
it
hm
s
in
c
om
pl
e
x
a
nd
dyna
m
ic
ne
twor
k
s
c
e
na
r
io
s
[
12]
.
S
upe
r
vi
s
e
d
le
a
r
ni
ng
(
in
c
lu
di
ng
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
io
n)
pe
r
f
or
m
e
d w
e
ll
on t
r
a
f
f
ic
pr
e
di
c
ti
on.
U
s
in
g
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
io
n
,
Z
ha
o
[
5]
d
e
m
on
s
tr
a
t
e
d
th
e
e
f
f
e
c
ti
ve
ne
s
s
of
it
s
us
e
f
or
im
pr
ov
e
d
a
c
c
ur
a
c
y
a
nd
s
t
a
bi
li
t
y
of
lo
g
is
ti
c
s
c
os
t
pr
e
di
c
t
io
n
a
nd
pr
o
vi
de
d
m
or
e
e
f
f
e
c
ti
ve
s
uppor
t
f
or
c
o
s
t
c
ont
r
o
l
a
nd
ope
r
a
ti
on
opt
im
i
z
a
ti
o
n
of
a
lo
g
is
ti
c
s
e
nt
e
r
pr
i
s
e
.
B
e
n
e
f
it
s
of
u
s
i
ng
th
e
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
or
w
e
r
e
a
ut
om
a
ti
c
f
e
a
tu
r
e
s
e
le
c
ti
on,
e
f
f
ic
ie
nt
ove
r
f
it
ti
ng
pr
e
ve
n
ti
on,
a
nd
th
e
c
a
pa
c
it
y
to
h
a
ndl
e
hi
gh
-
di
m
e
ns
i
ona
l
da
t
a
[
6]
.
C
om
pa
r
e
d t
o L
a
s
s
o
r
e
gr
e
s
s
i
on w
it
h a
n
R
² of
0
.39,
r
a
ndom f
or
e
s
t
r
e
gr
e
s
s
io
n
pr
ov
e
d t
o
out
pe
r
f
or
m
w
it
h
a
n R
² of
0.86.
T
h
e
a
u
th
or
c
onc
l
ude
d
t
ha
t
t
he
r
a
nd
om
f
or
e
s
t
r
e
gr
e
s
s
or
w
a
s
hi
ghl
y
e
f
f
ic
i
e
nt
f
or
pr
e
di
c
ti
ng t
he
c
o
s
t.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2722
-
3221
C
om
put
S
c
i
I
nf
T
e
c
hnol
,
V
ol
. 7, No. 1, M
a
r
c
h 2026
:
56
-
65
58
B
out
a
ba
e
t
al
.
[
13]
c
om
pi
le
d
a
w
e
a
lt
h
of
da
ta
on
m
a
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
s
to
a
s
s
i
s
t
ne
twor
ki
ng.
T
he
y
a
dde
d
v
a
lu
a
bl
e
in
f
or
m
a
ti
on
a
bout
th
e
m
e
th
ods
us
e
d,
th
e
ir
a
dva
nt
a
ge
s
,
di
s
a
dva
nt
a
ge
s
,
a
nd
vi
a
bi
li
ty
in
a
c
tu
a
l
ne
twor
ki
ng
s
it
ua
ti
ons
to
th
e
c
onve
r
s
a
ti
on.
N
e
ur
a
l
ne
two
r
ks
,
w
hi
c
h
a
r
e
us
e
d
in
pr
e
di
c
ti
ng
f
ut
u
r
e
tr
a
f
f
ic
ba
s
e
d on pa
s
t
tr
a
f
f
ic
da
ta
, w
e
r
e
f
ound
t
o be
ve
r
y a
c
c
ur
a
te
f
or
b
ot
h s
hor
t
-
te
r
m
a
nd l
ong
-
te
r
m
pr
e
di
c
ti
ons
w
hi
le
be
in
g
s
im
pl
e
a
nd
r
e
qui
r
in
g
f
e
w
f
e
a
tu
r
e
s
.
T
h
e
pa
pe
r
“
opt
im
a
l
ne
twor
k
r
out
e
e
s
ti
m
a
to
r
us
in
g
pr
e
di
c
ti
on
a
lg
or
it
hm
s
”
,
pr
opos
e
d
a
pr
e
di
c
ti
on
-
ba
s
e
d
r
out
in
g
m
ode
l
th
a
t
id
e
nt
if
ie
s
opt
im
a
l
ne
twor
k
pa
th
s
to
s
uppor
t
e
f
f
ic
ie
nt
da
ta
tr
a
ns
m
is
s
io
n
f
r
om
s
our
c
e
to
de
s
ti
na
ti
on
I
P
a
ddr
e
s
s
e
s
,
th
e
r
e
by
im
pr
ovi
ng
ove
r
a
ll
tr
a
ns
m
is
s
io
n
e
f
f
ic
ie
nc
y i
n t
e
r
m
s
of
qua
li
ty
-
of
-
s
e
r
vi
c
e
m
e
tr
ic
s
[
14]
.
H
a
il
a
n
a
nd
A
ls
ha
he
e
n
[
15]
r
e
vi
e
w
e
d
a
nd
e
xa
m
in
e
d
th
e
c
onc
e
pt
s
of
ne
twor
k
pl
a
nni
ng
a
nd
opt
im
iz
a
ti
on, t
r
a
f
f
ic
e
ngi
ne
e
r
in
g, a
nd
Q
oS
to
f
in
d p
r
a
c
ti
c
a
l
w
a
ys
t
o i
m
pr
ove
ne
twor
k pe
r
f
o
r
m
a
nc
e
a
nd r
e
duc
e
la
te
nc
y
a
nd
pa
c
ke
t
lo
s
s
.
T
he
s
tu
dy
e
m
pha
s
i
z
e
d
th
e
im
por
ta
nc
e
of
a
dva
nc
e
d
tr
a
f
f
ic
e
ngi
ne
e
r
in
g
s
tr
a
te
gi
e
s
to
m
a
na
ge
th
e
r
is
in
g
ne
twor
k
c
om
pl
e
xi
ty
a
nd
th
e
gr
ow
in
g
de
m
a
nd
f
or
hi
gh
-
s
pe
e
d
c
onne
c
ti
vi
ty
a
nd
di
ve
r
s
e
s
e
r
vi
c
e
s
.
I
t
id
e
nt
if
ie
d
ke
y
di
r
e
c
ti
ons
f
or
f
ut
ur
e
r
e
s
e
a
r
c
h,
in
c
lu
di
ng
th
e
de
ve
lo
pm
e
nt
of
dyna
m
ic
a
nd
a
da
pt
iv
e
tr
a
f
f
ic
m
a
na
ge
m
e
nt
a
ppr
oa
c
he
s
l
e
ve
r
a
gi
ng ma
c
hi
ne
l
e
a
r
ni
ng.
A
s
tu
dy
to
e
xa
m
in
e
how
s
e
r
vi
c
e
qua
li
ty
,
tr
us
t,
a
nd
c
om
m
it
m
e
nt
in
f
lu
e
nc
e
c
us
to
m
e
r
lo
ya
lt
y
a
m
ong
te
le
c
om
s
e
r
vi
c
e
us
e
r
s
in
I
ndi
a
,
w
hi
le
a
ls
o
a
s
s
e
s
s
in
g
th
e
m
od
e
r
a
ti
ng
e
f
f
e
c
ts
of
g
e
nde
r
,
m
a
r
it
a
l
s
ta
tu
s
,
a
nd
c
onne
c
ti
on
ty
pe
.
U
s
in
g
s
ur
ve
y
r
e
s
pon
s
e
s
,
th
e
s
tu
dy
f
ound
t
ha
t
r
e
s
pons
iv
e
n
e
s
s
,
a
s
s
ur
a
nc
e
,
a
nd
e
m
pa
th
y
pos
it
iv
e
ly
a
f
f
e
c
t
bot
h
tr
us
t
a
nd
c
om
m
it
m
e
nt
,
w
he
r
e
a
s
ta
ngi
bi
li
ty
in
f
lu
e
nc
e
s
tr
us
t
onl
y.
C
om
m
it
m
e
nt
a
nd
tr
us
t
w
e
r
e
s
how
n t
o pos
it
iv
e
ly
i
m
pa
c
t
lo
ya
lt
y
[
16]
.
T
he
li
te
r
a
tu
r
e
il
lu
s
tr
a
te
d
th
e
w
id
e
-
r
a
ngi
ng
us
e
of
di
f
f
e
r
e
nt
a
ppr
oa
c
he
s
a
nd
te
c
hni
que
s
to
im
pe
d
e
ne
twor
k
r
out
in
g
opt
im
iz
a
ti
on.
I
t
m
a
de
us
e
of
th
e
pot
e
nt
ia
l
f
or
m
a
c
hi
ne
le
a
r
ni
ng
to
br
id
ge
th
e
ga
p
c
a
us
e
d
by
tr
a
di
ti
ona
l
m
a
nua
l
a
ppr
oa
c
he
s
. R
e
s
e
a
r
c
he
r
s
f
ound tha
t
s
upe
r
vi
s
e
d l
e
a
r
ni
ng a
lg
or
it
hm
s
w
e
r
e
t
he
m
os
t
e
f
f
e
c
ti
ve
in
th
e
s
e
opt
im
iz
a
ti
on
s
e
tt
in
gs
.
T
h
e
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
or
a
ls
o
pr
ove
d
to
ha
ve
hi
gh
pr
e
di
c
ti
on a
c
c
ur
a
c
y
w
it
h
m
in
im
a
l
e
r
r
or
. T
he
r
e
s
e
a
r
c
h di
d not i
nc
or
por
a
te
r
e
a
l
-
w
or
ld
t
e
le
c
om
da
ta
i
nt
e
gr
a
ti
on.
3.
M
E
T
H
O
D
T
hi
s
s
tu
dy
w
il
l
c
onc
e
nt
r
a
te
on
c
r
e
a
ti
ng
a
nd
a
s
s
e
s
s
in
g
a
m
a
c
h
in
e
le
a
r
ni
ng
-
ba
s
e
d
le
a
s
t
-
c
os
t
r
out
in
g
m
ode
l
th
a
t
ta
ke
s
in
to
a
c
c
ount
c
r
uc
ia
l
f
a
c
to
r
s
li
ke
c
a
ll
c
o
s
ts
a
nd
ne
twor
k
c
ondi
ti
ons
us
in
g
r
e
a
l
-
w
or
ld
te
le
c
om
da
ta
s
e
ts
.
T
he
pur
pos
e
of
th
e
s
tu
dy
is
to
a
ut
om
a
te
c
a
ll
r
out
i
ng
w
hi
ls
t
m
in
im
iz
in
g
c
a
ll
r
out
in
g
c
os
ts
w
it
h
e
nha
nc
e
d
n
e
twor
k
ut
il
iz
a
ti
on.
T
h
e
e
f
f
e
c
ti
ve
ne
s
s
of
th
e
m
ode
l
w
il
l
be
c
onf
ir
m
e
d
th
r
ough
c
om
pa
r
is
on
w
it
h
tr
a
di
ti
ona
l
r
out
in
g
te
c
hni
que
s
.
T
e
le
c
om
op
e
r
a
to
r
s
w
il
l
ga
in
a
c
om
pe
ti
ti
ve
e
dge
w
it
h
th
e
s
ugge
s
te
d
m
ode
l
be
c
a
us
e
it
a
ll
ow
s
th
e
m
to
a
tt
r
a
c
t
a
nd
r
e
ta
in
c
ons
um
e
r
s
by
pr
ovi
di
ng
hi
gh
s
e
r
vi
c
e
r
e
li
a
bi
li
ty
a
nd
r
e
duc
in
g
in
te
r
c
onne
c
ti
on
c
os
ts
.
i)
to
dr
a
w
in
a
nd
ke
e
p
c
ons
um
e
r
s
,
bus
in
e
s
s
e
s
m
us
t
pr
ovi
de
hi
gh
s
e
r
vi
c
e
r
e
li
a
bi
li
ty
a
nd
r
e
duc
e
d
in
te
r
c
onne
c
ti
on
c
os
ts
;
ii
)
te
le
c
om
c
om
pa
ni
e
s
c
a
n
m
or
e
e
f
f
ic
ie
nt
ly
a
ll
oc
a
te
r
e
s
our
c
e
s
th
a
nks
to
a
ut
om
a
ti
on,
w
hi
c
h
lo
w
e
r
s
ope
r
a
ti
ona
l
ove
r
he
a
d
;
a
nd
ii
i)
by
c
ut
ti
ng
e
xpe
ns
e
s
,
in
c
r
e
a
s
in
g
pr
oduc
ti
vi
ty
,
a
nd
gua
r
a
nt
e
e
in
g
th
e
be
s
t
pos
s
ib
le
c
a
ll
qua
li
ty
f
or
c
us
to
m
e
r
s
,
a
m
a
c
hi
ne
le
a
r
ni
ng
-
pow
e
r
e
d
s
ys
te
m
gi
ve
s
te
le
c
om
ope
r
a
to
r
s
a
c
om
pe
ti
ti
ve
a
dva
nt
a
ge
.
H
ow
th
e
m
a
c
hi
ne
le
a
r
ni
ng
m
ode
l
f
it
s
w
it
hi
n
a
te
le
c
om
r
out
in
g
s
ta
c
k
c
a
n
be
s
e
e
n
in
F
ig
ur
e
1
.
U
s
in
g
a
n a
ppl
ie
d qua
nt
it
a
ti
ve
r
e
s
e
a
r
c
h de
s
ig
n, t
hi
s
w
or
k a
im
s
t
o de
ve
l
op a
nd va
li
da
te
a
m
a
c
hi
ne
l
e
a
r
ni
ng mode
l
th
a
t
c
a
n
pr
e
di
c
t
th
e
be
s
t
in
te
r
c
onne
c
ti
on
r
out
e
s
f
or
voi
c
e
c
a
ll
s
.
T
o
opt
im
iz
e
f
or
c
a
ll
c
os
t
a
nd
la
te
nc
y,
th
e
s
tu
dy
f
oc
us
e
s
on
c
r
e
a
ti
ng
a
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
or
m
ode
l.
T
e
l
O
ne
Z
im
ba
bw
e
pr
ovi
de
s
hi
s
to
r
ic
a
l
te
le
c
om
c
onne
c
ti
vi
ty
da
ta
f
or
t
r
a
in
in
g a
nd e
va
lu
a
ti
ng t
he
m
ode
l.
F
ig
ur
e
1.
M
L
m
ode
l
in
te
gr
a
ti
on w
it
hi
n t
he
t
e
le
c
om
c
a
ll
r
out
in
g
s
ta
c
k
Evaluation Warning : The document was created with Spire.PDF for Python.
C
om
put
S
c
i
I
nf
T
e
c
hnol
I
S
S
N
:
2722
-
3221
O
pt
imi
z
in
g i
nt
e
r
c
onne
c
ti
on c
al
l
r
out
in
g:
a m
ac
hi
ne
l
e
ar
ni
ng ap
pr
oac
h f
or
c
o
s
t
and
…
(
I
v
y
A
ne
s
u M
udar
i
)
59
3.1. P
la
n
n
in
g a
n
d
d
at
a
p
r
e
p
ar
at
io
n
T
he
pr
oc
e
s
s
of
de
s
ig
ni
ng
a
nd
d
e
ve
lo
pi
ng
a
r
e
s
e
a
r
c
h
c
onc
e
pt
ua
l
f
r
a
m
e
w
or
k
in
vol
ve
d
ga
th
e
r
in
g
th
e
r
e
qui
r
e
d
c
a
ll
de
ta
il
r
e
c
or
d
s
(
C
D
R
s
)
,
da
ta
tr
a
ns
f
or
m
a
ti
on,
a
nd
a
na
ly
z
in
g
th
e
d
a
ta
f
or
th
e
m
od
e
l.
T
h
e
C
D
R
s
w
e
r
e
r
e
a
di
ly
a
va
il
a
bl
e
a
t
no
c
os
t.
N
o
hum
a
n
s
ubj
e
c
t
s
w
e
r
e
d
ir
e
c
tl
y
in
vol
ve
d
in
th
is
r
e
s
e
a
r
c
h.
W
e
e
ns
ur
e
d
e
th
ic
a
l
c
le
a
r
a
nc
e
a
nd pr
iv
a
c
y c
om
pl
ia
nc
e
by i
n
c
lu
di
ng his
to
r
ic
a
l
in
te
r
c
onne
c
ti
on c
a
ll
l
ogs
i
n t
he
da
ta
s
e
t.
3.2.
D
at
a an
d
p
re
-
p
r
oc
e
s
s
in
g
T
he
da
ta
a
nd
pr
e
-
pr
oc
e
s
s
in
g
s
ta
g
e
in
vol
ve
s
e
ve
r
a
l
s
te
ps
to
pr
e
pa
r
e
th
e
C
D
R
s
f
or
m
ode
l
de
ve
lo
pm
e
nt
, a
s
out
li
ne
d
f
ol
lo
w
s
:
i)
C
D
R
da
ta
c
ol
le
c
ti
on
:
t
he
pr
oc
e
s
s
of
c
ol
le
c
ti
ng
C
D
R
da
ta
s
a
m
pl
e
s
f
r
om
th
e
da
ta
ba
s
e
of
pr
e
vi
ous
ly
c
a
pt
ur
e
d
C
D
R
f
il
e
s
w
a
s
don
e
th
r
ough
th
e
D
e
pa
r
tm
e
nt
of
T
e
lOne
I
nt
e
r
c
onne
c
ti
on.
T
he
d
a
ta
ha
s
th
e
f
ol
lo
w
in
g c
ol
um
ns
:
D
a
ta
s
e
t
C
ol
um
n
s
c
a
ll
in
g_numbe
r
c
a
ll
e
d_numbe
r
c
a
ll
_dur
a
ti
on_mi
n
or
ig
in
a
ti
ng_c
a
r
r
ie
r
de
s
ti
na
ti
on_c
a
r
r
ie
r
de
s
ti
na
ti
on_c
ount
r
y
r
out
e
_us
e
d
c
os
t_
pe
r
_m
in
_us
d
la
te
nc
y_m
s
is
_pe
a
k_hour
to
ta
l_
c
os
t_
us
d
da
y_of
_w
e
e
k
c
a
ll
_hour
ii)
A
nonymi
z
a
ti
on:
i
n
c
om
pl
ia
nc
e
w
it
h
th
e
da
ta
pr
ot
e
c
ti
on
a
c
t
[
17]
,
c
a
ll
e
r
num
be
r
s
w
e
r
e
c
onve
r
te
d
in
to
ne
w
num
be
r
s
be
f
or
e
us
e
.
T
hi
s
is
due
to
th
e
e
xt
r
e
m
e
s
e
ns
it
iv
it
y
of
th
e
te
le
c
om
m
uni
c
a
ti
ons
da
ta
[
18]
, [
19]
. T
hi
s
a
ls
o he
lp
e
d t
o s
e
p
a
r
a
te
i
de
nt
it
ie
s
of
c
u
s
to
m
e
r
s
’
a
nd t
he
ir
pe
r
s
ona
l
bus
in
e
s
s
.
3.3.
F
e
at
u
r
e
s
e
le
c
t
io
n
T
h
e
f
ol
l
ow
i
ng
t
h
e
s
e
l
e
c
ti
on
a
n
d
e
xt
r
a
c
ti
o
n
of
s
i
gn
if
i
c
a
nt
f
e
a
tu
r
e
s
,
s
e
v
e
r
a
l
f
e
a
t
ur
e
s
w
e
r
e
id
e
nt
if
i
e
d.
T
h
e
s
e
f
e
a
t
ur
e
s
w
e
r
e
f
o
und
to
be
th
e
m
o
s
t
p
e
r
ti
ne
nt
i
n
d
e
t
e
r
m
i
ni
ng
c
a
ll
r
ou
ti
n
g
c
o
s
t
s
.
T
he
r
e
s
u
lt
s
a
r
e
s
how
n
in
T
a
bl
e
1.
T
a
bl
e
1.
D
a
ta
f
ie
ld
s
a
nd t
he
ir
r
e
le
va
nc
e
f
or
pr
e
di
c
ti
on
F
i
e
l
d
R
e
l
e
va
nc
e
or
i
gi
na
t
i
ng_c
a
r
r
i
e
r
T
he
or
i
gi
na
t
i
ng_c
a
r
r
i
e
r
he
l
ps
i
de
nt
i
f
y t
he
i
ni
t
i
a
l
ne
t
w
or
k, w
hi
c
h i
nf
l
ue
nc
e
s
r
out
i
ng de
c
i
s
i
ons
.
de
s
t
i
na
t
i
on_c
a
r
r
i
e
r
D
e
s
t
i
na
t
i
on_c
a
r
r
i
e
r
s
e
r
ve
s
a
s
a
n i
ndi
c
a
t
or
of
t
he
t
a
r
ge
t
ne
t
w
or
k, i
nf
l
ue
nc
i
ng pr
i
c
i
ng a
nd i
nt
e
r
c
onne
c
t
i
on
a
gr
e
e
m
e
nt
s
.
de
s
t
i
na
t
i
on_c
ount
r
y
E
s
s
e
nt
i
a
l
f
or
f
i
gur
i
ng out
c
os
t
s
, r
ul
e
s
, a
nd r
out
i
ng r
out
e
s
.
r
out
e
_us
e
d
T
he
di
s
pl
a
y of
t
he
c
ur
r
e
nt
r
out
i
ng pa
t
h i
s
e
s
s
e
nt
i
a
l
f
or
t
r
a
i
ni
ng a
nd va
l
i
da
t
i
on pur
pos
e
s
.
c
os
t
_pe
r
_m
i
n_us
d
T
he
obj
e
c
t
i
ve
s
of
L
C
R
opt
i
m
i
z
a
t
i
on a
r
e
c
l
o
s
e
l
y r
e
l
a
t
e
d t
o c
os
t
.
l
a
t
e
nc
y_m
s
T
he
ne
e
d t
o ba
l
a
nc
e
c
os
t
a
nd qua
l
i
t
y m
a
y i
nf
l
ue
nc
e
r
out
e
s
e
l
e
c
t
i
on.
i
s
_pe
a
k_hour
N
e
t
w
or
k c
onge
s
t
i
on m
a
y i
m
pa
c
t
bot
h a
va
i
l
a
bl
e
r
out
e
s
a
nd c
os
t
s
.
da
y_of
_w
e
e
k
I
t
m
a
y i
m
pa
c
t
c
a
l
l
pa
t
t
e
r
ns
but
ha
s
l
e
s
s
of
a
di
r
e
c
t
c
os
t
i
m
pa
c
t
.
c
a
l
l
_hour
T
he
i
m
pa
c
t
on ne
t
w
or
k l
oa
d a
nd pr
i
c
i
ng m
ode
l
s
m
a
ke
s
i
t
m
ode
r
a
t
e
l
y r
e
l
e
va
nt
.
3.
4
.
M
at
e
r
ia
ls
an
d
t
ool
s
T
he
m
a
t
e
r
ia
ls
a
nd
to
ol
s
us
e
d
in
th
is
s
tu
dy
in
c
lu
de
th
e
P
yt
hon
f
or
pr
ogr
a
m
m
in
g
la
ngua
ge
.
T
he
l
ib
r
a
r
ie
s
us
e
d a
r
e
S
c
ik
it
-
le
a
r
n, P
a
nda
s
, N
um
P
y, a
nd M
a
tp
lo
tl
ib
, a
lo
ng w
it
h t
he
r
a
ndom f
or
e
s
t
r
e
gr
e
s
s
or
f
or
t
he
m
ode
l.
T
he
p
la
tf
or
m
us
e
d i
s
a
l
oc
a
l
m
a
c
hi
ne
w
it
h s
uf
f
ic
ie
nt
c
o
m
put
a
ti
ona
l
pow
e
r
f
or
t
r
a
in
in
g.
3.5.
M
od
e
l
d
e
ve
lo
p
m
e
n
t
T
he
m
a
in
obj
e
c
ti
ve
is
to
pr
e
di
c
t
th
e
c
a
r
r
ie
r
or
r
out
e
w
it
h
th
e
lo
w
e
s
t
pr
oj
e
c
te
d
c
os
t
w
hi
le
m
e
e
ti
ng
Q
oS
r
e
qui
r
e
m
e
nt
s
.
T
o
a
c
hi
e
ve
th
is
,
a
s
upe
r
vi
s
e
d
le
a
r
ni
ng
a
lg
o
r
it
hm
w
a
s
us
e
d
to
de
ve
lo
p
a
p
r
e
di
c
ti
ve
m
ode
l
th
a
t
c
a
n
id
e
nt
if
y
pa
tt
e
r
ns
in
da
ta
,
f
or
e
c
a
s
t
f
ut
ur
e
tr
e
nds
,
a
nd
m
a
ke
f
a
c
t
-
ba
s
e
d
de
c
i
s
io
ns
[
20]
.
T
h
e
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
or
w
a
s
t
r
a
in
e
d t
o f
or
e
c
a
s
t
th
e
m
o
s
t
e
c
onomi
c
a
l
pa
t
h. T
he
pr
opos
e
d f
r
a
m
e
w
or
k f
or
ge
ne
r
a
l
m
ode
l
tr
a
in
in
g
a
nd
lo
gi
c
pr
ogr
a
m
m
in
g
w
a
s
de
ve
lo
pe
d
in
P
yt
hon. T
he
m
ode
l
w
a
s
tr
a
in
e
d
to
m
in
im
iz
e
m
e
a
n a
bs
ol
ut
e
e
r
r
or
(
M
A
E
)
a
nd me
a
n s
qua
r
e
d e
r
r
or
(
M
S
E
)
, pr
e
di
c
ti
ng t
he
l
ow
e
s
t
-
c
os
t
r
out
e
gi
ve
n a
s
e
t
of
i
nput
c
ondi
ti
ons
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2722
-
3221
C
om
put
S
c
i
I
nf
T
e
c
hnol
,
V
ol
. 7, No. 1, M
a
r
c
h 2026
:
56
-
65
60
3.5.1.
R
an
d
om
f
or
e
s
t
r
e
gr
e
s
s
o
r
T
he
a
lg
or
it
hm
us
e
s
t
h
e
m
a
th
e
m
a
ti
c
a
l
e
qu
a
ti
on s
how
n
a
s
(
1)
.
̂
=
1
∑
(
)
=
1
(
1)
W
he
r
e
,
ŷ
is
pr
e
di
c
te
d c
o
s
t
,
N
is
num
be
r
of
t
r
e
e
s
, a
nd
(
)
is
pr
e
di
c
ti
on f
r
om
t
r
e
e
i
f
or
in
put
x
.
U
s
in
g
va
r
io
us
da
ta
s
ubs
e
ts
,
r
a
ndom
f
or
e
s
ts
c
ons
tr
uc
t
m
ul
ti
pl
e
de
c
is
io
n
tr
e
e
s
be
f
or
e
a
ve
r
a
gi
ng
th
e
out
c
om
e
s
.
T
he
m
os
t
in
f
lu
e
nt
ia
l
f
e
a
tu
r
e
s
a
f
f
e
c
ti
ng
c
a
ll
c
os
t
s
c
a
n
be
id
e
nt
if
ie
d
us
in
g
r
a
ndom
f
or
e
s
t
s
to
c
a
lc
ul
a
te
f
e
a
tu
r
e
im
por
ta
nc
e
s
c
or
e
s
[
21]
,
[
22]
,
a
s
il
lu
s
tr
a
te
d
i
n
F
ig
ur
e
2.
T
hi
s
b
e
ne
f
it
s
bus
in
e
s
s
in
s
ig
ht
s
a
s
w
e
ll
a
s
m
ode
l
tr
a
ns
pa
r
e
nc
y.
O
ne
m
e
th
od
th
a
t
c
a
n
r
e
duc
e
ove
r
f
it
ti
ng
is
r
a
ndom
f
or
e
s
ts
[
23]
,
[
24
]
.
B
e
c
a
us
e
it
us
e
s
f
e
w
e
r
pa
r
a
m
e
te
r
s
th
a
n
ot
he
r
e
ns
e
m
bl
e
a
lg
or
it
hm
s
to
pr
oc
e
s
s
la
r
ge
da
ta
s
e
t
s
,
it
is
a
ls
o
qui
c
k
a
nd
e
f
f
ic
ie
nt
[
25]
.
T
he
f
or
m
ul
a
w
a
s
us
e
d
to
de
te
r
m
in
e
th
e
a
ve
r
a
ge
l
a
te
nc
y
e
xpe
r
ie
nc
e
d
w
he
n
ut
il
iz
in
g
th
e
c
hos
e
n
r
out
e
, a
s
w
e
ll
a
s
pot
e
nt
ia
l
r
out
e
s
t
ha
t
m
ig
ht
pr
ovi
de
l
ow
e
r
c
a
ll
t
e
r
m
in
a
ti
on r
a
te
s
.
T
he
m
ode
l'
s
pe
r
f
or
m
a
nc
e
w
a
s
va
li
da
te
d,
a
nd
it
s
ge
ne
r
a
li
z
a
bi
li
t
y
w
a
s
e
ns
ur
e
d
u
s
in
g
a
70/
30
tr
a
in
-
te
s
t
s
pl
it
a
nd
5
-
f
ol
d
c
r
os
s
-
va
li
da
ti
on.
R
a
ndom
f
or
e
s
t
w
a
s
c
hos
e
n
f
or
it
s
a
bi
li
ty
to
ba
la
nc
e
s
tr
ong
pr
e
di
c
ti
ve
pe
r
f
or
m
a
nc
e
w
it
h
in
te
r
pr
e
ta
bi
li
ty
,
m
a
ki
ng
it
w
e
ll
-
s
ui
te
d
f
or
a
ppl
ic
a
ti
ons
w
he
r
e
unde
r
s
ta
ndi
ng
m
ode
l
de
c
is
io
ns
is
im
por
ta
nt
.
T
he
s
tu
dy
ga
in
e
d
pr
a
c
ti
c
a
l
s
ig
ni
f
ic
a
nc
e
,
a
nd
th
e
m
ode
l'
s
va
li
di
ty
w
a
s
im
pr
ove
d
by
us
in
g
r
e
a
l
-
w
or
ld
te
le
c
om
da
ta
f
r
om
a
n
ope
r
a
ti
ona
l
s
e
tt
in
g,
w
hi
c
h
r
e
f
le
c
te
d
a
c
tu
a
l
r
out
in
g
a
nd
c
o
s
t
c
ondi
ti
ons
.
T
o m
a
ke
s
ur
e
t
he
m
ode
l
in
c
lu
de
d t
he
m
os
t
pe
r
ti
ne
nt
e
le
m
e
nt
s
a
f
f
e
c
ti
ng c
a
ll
c
os
t
a
nd qua
li
ty
,
f
e
a
tu
r
e
s
e
le
c
ti
on
w
a
s
gui
de
d
by
dom
a
in
e
xp
e
r
ie
nc
e
a
nd
c
onc
e
nt
r
a
te
d
on
va
r
ia
b
le
s
s
tr
ongl
y
r
e
la
te
d
to
r
out
in
g
e
c
onomi
c
s
a
nd
ne
twor
k r
e
s
tr
ic
ti
ons
.
F
ig
ur
e
2. R
a
ndom f
or
e
s
t
f
e
a
tu
r
e
i
m
por
ta
nc
e
f
or
l
e
a
s
t
-
c
os
t
r
out
in
g
3.6.
E
val
u
at
io
n
m
e
t
r
ic
s
T
o
de
te
r
m
in
e
th
e
di
s
ti
nc
ti
on
a
nd
im
por
ta
nc
e
,
th
e
m
ode
l'
s
pr
e
c
is
io
n
a
nd
e
f
f
e
c
ti
ve
ne
s
s
in
f
or
e
c
a
s
ti
ng
th
e
be
s
t
pa
th
w
e
r
e
c
ont
r
a
s
te
d
w
it
h
c
onve
nt
io
na
l
r
out
in
g
te
c
hni
que
s
.
K
e
y
pe
r
f
or
m
a
nc
e
in
di
c
a
to
r
s
s
uc
h
a
s
a
c
c
e
pt
a
bl
e
la
te
nc
y,
le
a
s
t
-
c
os
t
r
out
e
id
e
nt
if
ic
a
ti
on,
a
nd
pr
e
di
c
ti
on
ti
m
e
w
e
r
e
a
na
ly
z
e
d.
T
he
f
ol
lo
w
in
g
e
va
lu
a
ti
on me
tr
ic
s
w
e
r
e
m
e
a
s
ur
e
d:
i)
C
os
t
p
e
r
f
or
m
a
nc
e
:
M
A
E
a
nd M
S
E
, w
hi
c
h r
e
pr
e
s
e
nt
t
he
di
s
c
r
e
p
a
nc
y be
twe
e
n e
xp
e
c
te
d a
nd
a
c
tu
a
l
c
o
s
ts
.
ii)
L
a
te
nc
y c
om
pa
r
is
on a
c
r
os
s
pr
e
di
c
te
d ve
r
s
u
s
t
r
a
di
ti
ona
l
r
out
e
s
(
th
r
e
s
hol
d 300 ms
)
.
iii)
P
r
e
di
c
ti
on
ti
m
e
:
r
eal
-
ti
m
e
f
e
a
s
ib
il
it
y
w
a
s
e
va
lu
a
te
d
by
a
c
c
ou
nt
in
g
f
or
th
e
a
ve
r
a
ge
ti
m
e
ta
ke
n
by
th
e
m
ode
l
to
pr
e
di
c
t
th
e
opt
im
a
l
r
out
e
.
iv
)
O
pe
r
a
ti
ona
l
e
f
f
ic
ie
nc
y:
a
c
om
pa
r
is
on
of
th
e
ov
e
r
a
ll
c
os
t
s
a
vi
ng
s
f
r
om
th
e
m
a
c
hi
ne
le
a
r
ni
ng
m
ode
l
w
it
h
th
e
t
r
a
di
ti
ona
l
L
C
R
te
c
hni
que
s
, w
hi
c
h r
e
li
e
d on
s
ta
ti
c
t
a
bl
e
s
.
T
he
s
tu
dy'
s
c
onc
lu
s
io
ns
m
a
y
not
b
e
a
s
br
oa
dl
y
a
ppl
ic
a
bl
e
to
ot
he
r
ne
twor
ks
or
ge
ogr
a
phi
c
a
l
a
r
e
a
s
be
c
a
us
e
it
is
ba
s
e
d
on
in
te
r
c
onn
e
c
ti
on
da
ta
f
r
om
a
s
in
gl
e
te
le
c
o
m
ope
r
a
to
r
,
T
e
lOne
.
F
ur
th
e
r
m
or
e
,
be
c
a
us
e
th
e
m
ode
l
is
tr
a
in
e
d
us
in
g
hi
s
to
r
ic
a
l
pr
ic
in
g
da
ta
,
it
m
a
y
be
c
om
e
le
s
s
s
uc
c
e
s
s
f
ul
in
e
nvi
r
onm
e
nt
s
w
it
h
pr
ic
in
g
s
tr
uc
tu
r
e
s
th
a
t
c
ha
ng
e
of
te
n,
r
e
qui
r
in
g
r
e
tr
a
in
in
g
on
a
r
e
gul
a
r
ba
s
is
.
A
ddi
ti
ona
ll
y,
e
ve
n
th
ough
la
te
n
c
y
da
ta
w
a
s
c
a
pt
ur
e
d
f
or
e
ve
r
y
c
onve
r
s
a
ti
on,
it
onl
y
r
e
pr
e
s
e
nt
s
th
e
a
ve
r
a
ge
c
ir
c
um
s
ta
nc
e
s
a
t
th
e
ti
m
e
of
th
e
c
a
ll
a
nd
m
ig
ht
not
a
c
c
ur
a
te
ly
r
e
f
le
c
t
ne
twor
k os
c
il
la
ti
ons
or
br
ie
f
de
la
ys
t
ha
t
oc
c
ur
i
n r
e
a
l
ti
m
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
C
om
put
S
c
i
I
nf
T
e
c
hnol
I
S
S
N
:
2722
-
3221
O
pt
imi
z
in
g i
nt
e
r
c
onne
c
ti
on c
al
l
r
out
in
g:
a m
ac
hi
ne
l
e
ar
ni
ng ap
pr
oac
h f
or
c
o
s
t
and
…
(
I
v
y
A
ne
s
u M
udar
i
)
61
3.7.
O
ve
r
al
l
m
od
e
l
ar
c
h
it
e
c
t
u
r
e
O
pt
im
a
l
r
out
e
pr
e
di
c
ti
on
w
or
kf
lo
w
is
il
lu
s
tr
a
te
d
in
F
ig
ur
e
3.
T
hi
s
f
ig
ur
e
pr
ovi
de
s
a
n
e
xa
m
pl
e
of
th
e
pr
oc
e
s
s
th
a
t
w
il
l
be
us
e
d
to
m
a
ke
th
e
pr
e
di
c
ti
on.
T
he
w
or
kf
lo
w
a
ls
o
hi
ghl
ig
ht
s
how
e
a
c
h
s
ta
ge
c
ont
r
ib
ut
e
s
to
ge
ne
r
a
ti
ng t
he
f
in
a
l
r
out
in
g output
.
F
ig
ur
e
3.
O
ve
r
a
ll
w
or
kf
lo
w
of
t
he
opt
im
a
l
r
out
e
pr
e
di
c
ti
on mod
e
l
4.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
4.1. M
od
e
l
t
r
ai
n
in
g r
e
s
u
lt
s
T
he
t
r
a
in
e
d m
ode
l
s
how
s
s
tr
ong pe
r
f
or
m
a
nc
e
, a
c
hi
e
vi
ng a
n R
²
of
0.851. I
t
a
ls
o pr
oduc
e
s
a
l
ow
M
A
E
of
$0.0482
pe
r
m
in
ut
e
,
in
d
ic
a
ti
ng
s
m
a
ll
a
ve
r
a
ge
pr
e
di
c
ti
on
e
r
r
o
r
s
.
A
ddi
ti
ona
ll
y,
th
e
M
S
E
of
0.0032
c
onf
ir
m
s
th
a
t
th
e
m
ode
l
pr
ovi
de
s
r
e
li
a
bl
e
a
nd a
c
c
ur
a
t
e
e
s
ti
m
a
te
s
.
4.2.
O
ve
r
al
l
m
od
e
l
r
e
s
u
lt
s
4.2.1.
R
ou
t
e
s
e
le
c
t
io
n
(
in
p
u
t
)
T
he
r
out
e
s
e
le
c
ti
on
pr
oc
e
s
s
r
e
qui
r
e
s
s
e
v
e
r
a
l
in
put
pa
r
a
m
e
te
r
s
th
a
t
in
f
lu
e
nc
e
th
e
m
ode
l’
s
de
c
is
io
n.
T
he
s
e
in
put
s
pl
a
y
a
c
r
uc
ia
l
r
ol
e
in
de
te
r
m
in
in
g
th
e
opt
im
a
l
r
out
e
f
or
e
a
c
h
c
a
ll
.
A
s
s
how
n
in
F
ig
ur
e
4,
th
e
in
te
r
f
a
c
e
a
ll
ow
s
us
e
r
s
t
o
s
pe
c
if
y t
he
d
e
s
ti
na
ti
on c
ount
r
y
, p
e
a
k/
o
ff
-
pe
a
k
, d
a
y of
w
e
e
k
, a
nd h
our
of
da
y
.
F
ig
ur
e
4
.
U
s
e
r
i
nt
e
r
f
a
c
e
f
or
s
e
le
c
ti
ng c
a
ll
r
out
in
g i
nput
pa
r
a
m
e
te
r
s
4.2.2.
R
ou
t
in
g r
e
c
om
m
e
n
d
at
io
n
r
e
s
u
lt
s
(
ou
t
p
u
t
)
T
he
m
ode
l
pr
oduc
e
s
s
e
ve
r
a
l
out
put
pa
r
a
m
e
te
r
s
th
a
t
de
s
c
r
ib
e
th
e
r
e
c
om
m
e
nde
d
r
out
e
a
nd
it
s
pe
r
f
or
m
a
nc
e
c
ha
r
a
c
te
r
is
ti
c
s
T
he
s
e
out
put
s
pr
ovi
de
in
s
ig
ht
in
to
bot
h
th
e
e
f
f
ic
ie
nc
y
a
nd
r
e
li
a
bi
li
ty
of
th
e
s
e
le
c
te
d
r
out
e
.
A
s
il
lu
s
tr
a
te
d
in
F
ig
ur
e
5
,
th
e
s
ys
te
m
di
s
pl
a
ys
th
e
r
e
c
om
m
e
nde
d
r
out
e
a
lo
ng
w
it
h
ke
y
m
e
tr
ic
s
s
uc
h a
s
r
out
e
c
a
r
r
ie
r
s
,
la
te
nc
y
,
pr
e
di
c
te
d c
o
s
t
,
s
uc
c
e
s
s
r
a
te
, a
nd
r
out
e
I
D
.
F
ig
ur
e
5
.
M
ode
l
out
put
di
s
pl
a
yi
ng t
he
r
e
c
om
m
e
nde
d r
out
e
a
nd
it
s
pe
r
f
or
m
a
nc
e
m
e
tr
ic
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2722
-
3221
C
om
put
S
c
i
I
nf
T
e
c
hnol
,
V
ol
. 7, No. 1, M
a
r
c
h 2026
:
56
-
65
62
4.3.
P
r
e
d
ic
t
io
n
t
im
e
T
he
pr
e
di
c
ti
on
ti
m
e
e
va
lu
a
ti
on
m
e
a
s
ur
e
s
how
f
a
s
t
th
e
m
od
e
l
c
a
n
ge
ne
r
a
te
a
r
out
in
g
de
c
is
io
n.
A
s
s
how
n
in
F
ig
ur
e
6,
th
e
P
yt
hon
c
ode
s
ni
ppe
t
c
a
lc
ul
a
te
s
th
e
a
ve
r
a
ge
pr
e
di
c
ti
on
ti
m
e
a
c
r
os
s
m
ul
ti
pl
e
r
uns
.
T
he
r
e
s
ul
ts
i
ndi
c
a
te
t
ha
t
th
e
m
ode
l
c
a
n pr
e
di
c
t
th
e
opt
im
a
l
r
out
e
i
n a
n a
ve
r
a
ge
of
0.080993 s
e
c
ond
s
.
F
ig
ur
e
6.
P
yt
hon s
c
r
ip
t
us
e
d t
o m
e
a
s
ur
e
t
he
m
ode
l’
s
a
ve
r
a
g
e
pr
e
di
c
ti
on t
im
e
4
.4.
C
os
t
c
om
p
ar
is
on
T
he
c
os
t
c
om
pa
r
is
on
hi
ghl
ig
ht
s
th
e
di
f
f
e
r
e
nc
e
be
twe
e
n
tr
a
di
ti
ona
ll
y
r
out
e
d
c
os
ts
a
nd
th
e
m
ode
l
-
ge
ne
r
a
te
d
r
out
in
g
c
os
t
s
.
A
s
s
ho
w
n
in
F
ig
ur
e
7
,
th
e
m
od
e
l
c
ons
i
s
te
nt
ly
pr
oduc
e
s
lo
w
e
r
c
os
ts
a
c
r
os
s
a
ll
de
s
ti
na
ti
ons
c
om
pa
r
e
d
to
tr
a
di
ti
ona
l
r
out
in
g.
F
u
r
th
e
r
m
or
e
,
T
a
bl
e
2
s
how
s
th
a
t
th
e
im
pl
e
m
e
nt
a
ti
on
of
th
e
m
ode
l
r
e
s
ul
ts
in
a
n
a
ve
r
a
ge
c
os
t
r
e
duc
ti
on
o
f
a
ppr
oxi
m
a
te
ly
46.75%
,
de
m
ons
tr
a
ti
ng
it
s
e
f
f
e
c
ti
ve
ne
s
s
in
im
pr
ovi
ng r
out
in
g e
f
f
ic
ie
nc
y.
F
ig
ur
e
7
.
C
om
pa
r
is
on of
t
r
a
di
ti
ona
l
c
os
t
vs
m
ode
l
c
os
t
T
a
bl
e
2
.
A
ve
r
a
ge
p
e
r
c
e
nt
a
ge
of
c
o
s
t
di
f
f
e
r
e
nc
e
D
e
s
t
i
na
t
i
on
R
out
e
us
e
d
T
r
a
di
t
i
ona
l
-
c
os
t
M
ode
l
-
c
os
t
C
os
t
di
f
f
e
r
e
nc
e
/
m
i
n
%
D
i
f
f
e
r
e
nc
e
N
G
(
N
i
ge
r
i
a
)
T
i
go
-
T
e
l
e
c
e
l
0.132
0.0843
0.0477
36.13636
Z
A
(
S
out
h A
f
r
i
c
a
)
T
e
l
e
c
e
l
-
O
r
a
nge
0.143
0.0746
0.0684
47.83217
U
S
(
U
ni
t
e
d S
t
a
t
e
s
)
T
e
l
e
c
e
l
-
Z
a
m
t
e
l
0.199
0.0786
0.1204
60.50251
G
H
(
G
ha
na
)
Z
a
m
t
e
l
-
Z
a
m
t
e
l
0.169
0.0713
0.0977
57.81065
K
E
(
K
e
nya
)
T
e
l
e
c
e
l
-
O
r
a
nge
0.158
0.0888
0.0692
43.79747
I
N
(
I
ndi
a
)
T
e
l
e
c
e
l
-
V
oda
0.106
0.0695
0.0365
34.4339
A
ve
r
a
ge
%
c
os
t
di
f
f
e
r
e
nc
e
46.75218
4
.5.
L
at
e
n
c
y
an
al
ys
is
T
he
de
la
y
a
s
s
oc
i
a
te
d
w
it
h
e
a
c
h
r
out
e
s
e
le
c
te
d
f
or
e
a
c
h
c
o
unt
r
y
w
a
s
e
va
lu
a
te
d
to
va
li
da
te
th
e
e
f
f
ic
ie
nc
y
of
th
e
s
e
le
c
te
d
opt
im
a
l
r
out
e
,
a
s
s
how
n
in
F
ig
ur
e
8
.
T
hi
s
a
na
ly
s
is
he
lp
s
to
a
s
s
e
s
s
how
qui
c
kl
y
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
C
om
put
S
c
i
I
nf
T
e
c
hnol
I
S
S
N
:
2722
-
3221
O
pt
imi
z
in
g i
nt
e
r
c
onne
c
ti
on c
al
l
r
out
in
g:
a m
ac
hi
ne
l
e
ar
ni
ng ap
pr
oac
h f
or
c
o
s
t
and
…
(
I
v
y
A
ne
s
u M
udar
i
)
63
m
ode
l
c
a
n pr
oc
e
s
s
r
out
in
g de
c
is
io
n
s
c
om
pa
r
e
d t
o t
r
a
di
ti
ona
l
m
e
th
ods
.
F
r
om
th
e
f
ig
ur
e
, i
t
w
a
s
hi
ghl
ig
ht
e
d t
ha
t
th
e
m
a
xi
m
um
l
a
te
nc
y w
a
s
unde
r
300 mi
ll
is
e
c
onds
.
F
ig
ur
e
8
.
C
om
pa
r
is
on of
t
r
a
di
ti
ona
l
la
te
nc
y vs
m
ode
l
la
te
nc
y
5.6.
D
is
c
u
s
s
io
n
T
he
m
ode
l
ha
s
e
xpl
a
na
to
r
y
pow
e
r
,
a
c
c
ount
in
g
f
or
85.1%
of
t
he
va
r
ia
nc
e
in
c
a
ll
c
os
t
pe
r
m
in
ut
e
,
w
it
h a
n a
ve
r
a
ge
a
bs
ol
ut
e
de
vi
a
ti
on of
4.82 c
e
nt
s
pe
r
m
in
ut
e
be
twe
e
n t
he
m
ode
l'
s
pr
e
di
c
te
d c
os
t
a
nd t
he
a
c
tu
a
l
c
os
t.
T
he
M
S
E
be
twe
e
n
th
e
m
ode
l'
s
pr
e
di
c
ti
ons
a
nd
th
e
a
c
tu
a
l
e
xpe
ns
e
s
i
s
0.0032,
s
ig
ni
f
yi
ng
c
om
m
e
nda
bl
e
m
ode
l
pe
r
f
or
m
a
nc
e
.
T
he
r
e
s
e
a
r
c
h
il
lu
s
tr
a
te
s
th
e
e
f
f
ic
a
c
y
of
a
m
a
c
hi
ne
le
a
r
ni
ng
-
ba
s
e
d
a
ppr
oa
c
h
f
or
opt
im
iz
in
g
le
a
s
t
-
c
os
t
r
out
in
g
in
in
te
r
c
onne
c
ti
on
c
a
ll
s
.
T
he
m
e
th
odol
ogy
e
nha
nc
e
s
r
out
in
g
de
c
is
io
n
s
to
m
in
im
iz
e
c
os
ts
a
nd
e
n
s
ur
e
s
upe
r
io
r
s
e
r
vi
c
e
by
le
ve
r
a
gi
ng
pr
e
v
io
us
da
ta
.
T
he
m
a
c
hi
ne
le
a
r
ni
ng
m
e
th
odol
ogy
de
c
r
e
a
s
e
d c
o
s
ts
by a
n a
ve
r
a
ge
of
46.75%
r
e
la
ti
ve
t
o c
onve
nt
io
n
a
l
r
out
in
g m
e
th
ods
. C
onve
nt
io
na
l
r
out
e
d c
os
ts
(
bl
ue
)
e
xc
e
e
d m
ode
l
-
r
out
in
g
c
os
ts
(
gr
e
e
n)
, w
it
h
a
s
ta
ti
s
ti
c
a
ll
y s
ig
ni
f
ic
a
nt
di
f
f
e
r
e
nc
e
, a
s
i
ll
us
tr
a
te
d i
n
F
ig
ur
e
7
.
G
iv
e
n
th
a
t
th
e
m
ode
l'
s
la
te
nc
y
w
a
s
s
a
ti
s
f
a
c
to
r
y,
th
e
Q
oS
r
e
m
a
in
e
d
uni
m
pe
de
d,
r
e
s
ul
ti
ng
in
a
n
ove
r
a
ll
e
qui
li
br
iu
m
be
twe
e
n
c
os
t
a
nd
qua
li
ty
. T
he
m
ode
l
pr
oc
e
s
s
e
d
r
ou
ti
ng
de
c
is
io
ns
in
0.08
s
e
c
onds
,
pr
ovi
ng
vi
a
bl
e
f
or
r
e
a
l
-
ti
m
e
de
pl
oym
e
nt
.
T
he
m
ode
l
in
te
r
f
a
c
e
s
w
it
h
hi
s
to
r
ic
a
l
s
y
s
te
m
s
th
r
ough
P
yt
hon
A
P
I
s
to
f
a
c
il
it
a
te
f
unc
ti
ona
li
ty
in
th
e
s
w
it
c
h
s
ys
te
m
s
th
a
t
gove
r
n
c
a
ll
r
out
in
g.
F
ut
u
r
e
r
e
a
l
-
ti
m
e
a
da
pt
a
bi
li
ty
m
a
y
in
ve
s
ti
ga
te
r
e
in
f
or
c
e
m
e
nt
le
a
r
ni
n
g
f
or
dyna
m
ic
pr
ic
in
g.
T
he
s
tu
dy'
s
li
m
it
a
ti
ons
e
n
c
om
pa
s
s
r
e
li
a
n
c
e
on
th
e
pr
e
c
is
io
n
a
nd
c
om
pr
e
he
n
s
iv
e
ne
s
s
of
hi
s
to
r
ic
a
l
da
ta
a
nd
th
e
uni
que
c
h
a
r
a
c
te
r
is
ti
c
s
of
pr
ic
in
g
a
gr
e
e
m
e
nt
s
,
w
hi
c
h
m
a
y
di
f
f
e
r
s
ig
ni
f
ic
a
nt
ly
a
m
ong
pr
ovi
de
r
s
a
nd l
oc
a
ti
ons
.
5.
C
O
N
C
L
U
S
I
O
N
T
hi
s
r
e
s
e
a
r
c
h s
uc
c
e
s
s
f
ul
ly
de
ve
lo
pe
d a
nd e
va
lu
a
te
d a
m
a
c
hi
ne
l
e
a
r
ni
ng
-
ba
s
e
d
L
C
R
m
ode
l
us
in
g r
e
a
l
in
te
r
c
onne
c
ti
on
da
ta
f
r
om
a
m
a
jo
r
te
le
c
om
ope
r
a
to
r
.
T
he
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
or
m
ode
l
de
m
ons
tr
a
te
d
s
tr
ong
pr
e
di
c
ti
ve
a
c
c
ur
a
c
y,
e
na
bl
in
g
up
to
60
%
c
os
t
s
a
vi
ngs
w
hi
le
m
a
in
ta
in
in
g
a
c
c
e
pt
a
bl
e
c
a
ll
qua
li
ty
w
it
h
la
te
nc
y
unde
r
300
m
s
.
W
it
h
a
n
a
ve
r
a
ge
pr
e
di
c
ti
on
ti
m
e
of
ju
s
t
0.08
s
e
c
onds
,
th
e
m
ode
l
pr
ove
s
f
e
a
s
ib
le
f
or
r
e
a
l
-
ti
m
e
de
pl
oym
e
nt
.
T
he
s
e
r
e
s
ul
ts
hi
ghl
ig
ht
th
e
pr
a
c
ti
c
a
l
va
lu
e
of
a
p
pl
yi
ng
m
a
c
hi
ne
le
a
r
ni
ng
to
opt
im
iz
e
te
le
c
om
r
out
in
g
de
c
is
io
ns
—
r
e
duc
in
g
ope
r
a
ti
ona
l
c
os
t
s
,
e
nha
nc
in
g
e
f
f
ic
ie
nc
y,
a
nd
s
uppor
ti
ng
hi
gh
s
e
r
vi
c
e
r
e
li
a
bi
li
ty
.
T
he
us
e
of
hi
s
to
r
ic
a
l
te
l
e
c
om
da
ta
a
nd
m
e
a
s
ur
a
bl
e
pe
r
f
or
m
a
nc
e
m
e
tr
ic
s
e
n
s
ur
e
s
th
a
t
th
e
s
ol
ut
io
n
is
bot
h
r
obus
t
a
nd
gr
ounde
d
in
r
e
a
l
-
w
or
ld
c
ondi
ti
ons
.
L
ooki
ng
a
he
a
d,
th
is
m
ode
l
c
a
n
be
e
xt
e
nde
d
to
m
ul
ti
-
ope
r
a
to
r
e
nvi
r
onm
e
nt
s
,
in
te
gr
a
te
d
w
it
h
5G
/VoI
P
s
ys
te
m
s
,
a
nd
e
nha
n
c
e
d
us
in
g
r
e
in
f
or
c
e
m
e
nt
l
e
a
r
ni
ng
f
or
dyna
m
ic
a
da
pt
a
ti
on
to
f
lu
c
tu
a
ti
ng
pr
ic
in
g
a
nd
ne
twor
k
c
ondi
ti
ons
.
A
ddi
ti
ona
ll
y,
in
c
or
por
a
ti
ng
bl
oc
kc
ha
in
-
ba
s
e
d
s
e
tt
le
m
e
nt
s
m
a
y
f
ur
th
e
r
s
tr
e
a
m
li
ne
in
te
r
-
ope
r
a
to
r
c
os
t
r
e
c
on
c
il
ia
ti
on.
O
ve
r
a
ll
,
th
is
s
tu
dy
unde
r
s
c
or
e
s
th
e
gr
ow
in
g
r
e
le
va
nc
e
of
A
I
-
dr
iv
e
n
a
ut
om
a
ti
on
in
te
le
c
om
m
uni
c
a
ti
ons
a
nd
la
y
s
th
e
f
ounda
ti
on
f
or
s
m
a
r
te
r
,
c
os
t
-
e
f
f
ic
ie
nt
, a
nd a
da
pt
iv
e
r
out
in
g s
ys
te
m
s
.
A
C
K
N
O
WL
E
D
G
M
E
N
T
S
W
e
w
a
nt
to
s
in
c
e
r
e
ly
th
a
nk
T
e
lOne
P
vt
.
L
td
.
f
o
r
he
lp
in
g
w
i
th
th
e
m
ode
l'
s
r
e
qui
r
e
m
e
nt
s
a
nd
f
o
r
s
uppl
yi
ng
th
e
hi
s
to
r
ic
a
l
C
D
R
s
th
a
t
e
na
bl
e
d
th
i
s
s
tu
dy,
a
s
w
e
ll
a
s
f
or
he
lp
in
g
w
it
h
a
nonymi
z
in
g
c
onf
id
e
nt
ia
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2722
-
3221
C
om
put
S
c
i
I
nf
T
e
c
hnol
,
V
ol
. 7, No. 1, M
a
r
c
h 2026
:
56
-
65
64
da
ta
on
th
e
da
ta
s
e
t
gui
d
e
d
by
th
e
la
w
r
e
ga
r
di
ng
c
u
s
to
m
e
r
da
ta
pr
ot
e
c
ti
on.
T
he
ir
knowle
dg
e
,
pe
r
c
e
pt
io
ns
,
a
nd
he
lp
f
ul
c
r
it
ic
is
m
ha
ve
s
e
r
ve
d a
s
t
he
f
ounda
ti
on f
or
t
hi
s
s
tu
dy.
F
U
N
D
I
N
G
I
N
F
O
R
M
A
T
I
O
N
T
he
a
ut
hor
s
s
t
a
te
t
ha
t
no f
undi
ng i
s
i
nvol
ve
d i
n doing t
hi
s
r
e
s
e
a
r
c
h.
A
U
T
H
O
R
C
O
N
T
R
I
B
U
T
I
O
N
S
S
T
A
T
E
M
E
N
T
T
hi
s
jo
ur
na
l
us
e
s
th
e
C
ont
r
ib
ut
or
R
ol
e
s
T
a
xonomy
(
C
R
e
d
iT
)
to
r
e
c
ogni
z
e
in
di
vi
dua
l
a
ut
hor
c
ont
r
ib
ut
io
ns
, r
e
duc
e
a
ut
hor
s
hi
p di
s
put
e
s
,
a
nd f
a
c
il
it
a
te
c
ol
la
bo
r
a
ti
on.
N
am
e
o
f
A
u
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
I
vy A
ne
s
u M
uda
r
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
M
a
in
f
or
d M
ut
a
nda
va
r
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
K
e
nne
th
C
hi
w
or
e
r
a
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
onc
e
pt
ua
l
i
z
a
t
i
on
M
:
M
e
t
hodol
ogy
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
da
t
i
on
Fo
:
Fo
r
m
a
l
a
na
l
ys
i
s
I
:
I
nve
s
t
i
ga
t
i
on
R
:
R
e
s
our
c
e
s
D
:
D
a
t
a
C
ur
a
t
i
on
O
:
W
r
i
t
i
ng
-
O
r
i
gi
na
l
D
r
a
f
t
E
:
W
r
i
t
i
ng
-
R
e
vi
e
w
&
E
di
t
i
ng
Vi
:
Vi
s
ua
l
i
z
a
t
i
on
Su
:
Su
pe
r
vi
s
i
on
P
:
P
r
oj
e
c
t
a
dm
i
ni
s
t
r
a
t
i
on
Fu
:
Fu
ndi
ng a
c
qui
s
i
t
i
on
C
O
N
F
L
I
C
T
O
F
I
N
T
E
R
E
S
T
S
T
A
T
E
M
E
N
T
T
he
a
ut
h
o
r
s
s
ta
te
s
t
ha
t
no
kn
ow
n
c
o
m
p
e
t
in
g
f
i
na
nc
ia
l
in
te
r
e
s
ts
or
pe
r
s
ona
l
r
e
la
t
io
ns
hi
ps
c
ou
ld
ha
ve
a
pp
e
a
r
e
d
to
ha
v
e
a
n
i
m
pa
c
t
on
th
e
w
or
k
r
e
p
or
te
d
i
n
th
is
pa
pe
r
.
N
o
c
o
n
f
l
ic
ts
o
f
in
t
e
r
e
s
t
a
r
e
d
is
c
l
os
e
d
by
th
e
w
r
i
te
r
s
.
E
T
H
I
C
A
L
A
P
P
R
O
V
A
L
T
hi
s
s
tu
dy
c
om
pl
ie
s
w
it
h
e
th
ic
a
l
s
ta
nd
a
r
ds
f
or
te
le
c
om
m
uni
c
a
ti
ons
r
e
s
e
a
r
c
h.
A
ll
pe
r
s
ona
l
id
e
nt
if
ie
r
s
w
e
r
e
r
e
m
ove
d,
a
nd
a
ll
da
ta
us
e
d
w
a
s
tr
a
n
s
f
or
m
e
d
[
24]
.
N
o
pr
iv
a
te
da
ta
w
a
s
r
e
ve
a
l
e
d
[
25]
.
T
he
H
a
r
a
r
e
I
ns
ti
tu
te
of
T
e
c
hnol
ogy ga
ve
i
ts
i
n
s
ti
tu
ti
ona
l
a
ppr
ova
l
f
or
t
hi
s
s
tu
dy.
D
A
T
A
A
V
A
I
L
A
B
I
L
I
T
Y
D
ue
to
it
s
pr
opr
ie
ta
r
y
na
tu
r
e
a
nd
th
e
in
c
lu
s
io
n
of
c
om
m
e
r
c
ia
ll
y
s
e
ns
it
iv
e
da
ta
,
th
e
da
ta
s
e
t'
s
a
c
c
e
s
s
ib
il
it
y
is
li
m
it
e
d.
H
ow
e
ve
r
,
upon
r
e
que
s
t,
th
e
a
ut
hor
[
I
A
M
]
m
a
y
pr
ovi
de
th
e
a
nonymi
z
e
d
da
ta
s
e
t
us
e
d
f
or
m
ode
l
e
va
lu
a
ti
on.
R
E
F
E
R
E
N
C
E
S
[
1]
H
.
I
nt
ve
n,
J
.
O
l
i
ve
r
,
a
nd
E
.
S
e
pul
ve
da
,
T
e
l
e
c
om
m
uni
c
at
i
ons
r
e
gul
at
i
on
handbook
.
W
a
s
hi
ngt
on,
D
.C
.,
U
ni
t
e
d
S
t
a
t
e
s
:
T
he
W
or
l
d
B
a
nk, 2000.
[
2]
S
. I
t
i
ve
, “
O
pe
r
a
t
i
ona
l
e
xc
e
l
l
e
nc
e
s
t
r
a
t
e
gi
e
s
i
n t
he
t
e
l
e
c
om
m
uni
c
a
t
i
on i
ndu
s
t
r
y,”
I
nt
e
r
nat
i
onal
J
our
nal
of
R
e
s
e
ar
c
h P
ubl
i
c
at
i
on and
R
e
v
i
e
w
s
, vol
. 5, no. 12, pp. 1022
–
1027, 2024.
[
3]
D
.
M
i
t
r
a
,
S
.
S
a
r
ka
r
,
D
.
H
a
t
i
,
a
nd
D
.
M
i
t
r
a
,
“
A
c
om
pa
r
a
t
i
ve
s
t
udy
of
r
out
i
ng
pr
ot
oc
ol
s
,”
I
nt
e
r
nat
i
onal
R
e
s
e
a
r
c
h
J
ou
r
nal
o
f
A
dv
anc
e
d E
ngi
ne
e
r
i
ng and Sc
i
e
nc
e
, vol
. 2, no. 1, pp. 46
–
50, 2016.
[
4]
I
.
A
r
a
pa
ki
s
,
S
.
P
a
r
k,
a
nd
M
.
P
i
e
l
ot
,
“
I
m
pa
c
t
of
r
e
s
pons
e
l
a
t
e
nc
y
on
us
e
r
be
ha
vi
our
i
n
m
obi
l
e
w
e
b
s
e
a
r
c
h,”
ar
X
i
v
:
2101.09086
,
2021.
[
5]
S
.
Z
ha
o,
“
R
e
s
e
a
r
c
h
on
l
ogi
s
t
i
c
s
c
os
t
pr
e
di
c
t
i
on
ba
s
e
d
on
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
i
on
m
ode
l
i
ng,”
T
r
ans
ac
t
i
on
s
on
E
c
onom
i
c
s
,
B
us
i
ne
s
s
and M
anage
m
e
nt
R
e
s
e
a
r
c
h
, vol
. 11, pp. 367
–
376, N
ov. 2024, doi
:
10.
62051/
m
v8p5485.
[
6]
R
.
A
m
i
n,
E
.
R
oj
a
s
,
A
.
A
qdus
,
S
.
R
a
m
z
a
n,
D
.
C
.
-
P
e
r
e
z
,
a
nd
J
.
M
.
A
r
c
o,
“
A
s
ur
ve
y
on
m
a
c
hi
ne
l
e
a
r
ni
ng
t
e
c
hni
que
s
f
or
r
out
i
ng
opt
i
m
i
z
a
t
i
on i
n S
D
N
,”
I
E
E
E
A
c
c
e
s
s
, vol
. 9, pp. 104582
–
104611, 2021, doi
:
10.1109/
A
C
C
E
S
S
.2021.3099092.
[
7]
X
.
D
ou,
Z
.
W
a
ng,
S
.
L
i
,
a
nd
X
.
Z
ha
ng,
“
A
s
ur
ve
y
on
r
out
i
ng
a
l
gor
i
t
hm
s
b
a
s
e
d
on
m
a
c
hi
ne
l
e
a
r
ni
ng,”
I
ns
i
ght
s
i
n
C
om
put
e
r
,
Si
gnal
s
and Sy
s
t
e
m
s
, vol
. 1, no. 1, pp. 79
–
87, 2024, doi
:
10.70088/
a
z
t
vz
m
93.
[
8]
C
.
L
.
D
.
S
a
nt
o
s
,
A
.
M
.
M
e
z
he
r
,
J
.
P
.
A
.
L
e
on,
J
.
C
.
B
a
r
r
e
r
a
,
E
.
C
.
G
ue
r
r
a
,
a
nd
J
.
M
e
ng,
“
M
L
-
R
P
L
:
m
a
c
hi
ne
l
e
a
r
ni
ng
-
ba
s
e
d
r
out
i
ng
pr
ot
oc
ol
f
o
r
w
i
r
e
l
e
s
s
s
m
a
r
t
gr
i
d
ne
t
w
or
ks
,”
I
E
E
E
A
c
c
e
s
s
,
vol
.
11,
pp.
57401
–
57414,
2023,
doi
:
10.1109/
A
C
C
E
S
S
.2023.3283208.
[
9]
E
.
G
e
l
e
nbe
,
“
M
a
c
hi
ne
l
e
a
r
ni
ng
f
or
ne
t
w
or
k
r
out
i
ng,”
2020
9t
h
M
e
di
t
e
r
r
ane
an
C
onf
e
r
e
nc
e
on
E
m
be
dde
d
C
om
put
i
ng
(
M
E
C
O
)
,
pp. 1
–
1, 2020, doi
:
10.1109/
M
E
C
O
49872.2020.9134073.
[
10]
C
.
Z
ha
o,
M
.
Y
e
,
X
.
X
ue
,
J
.
L
v,
Q
.
J
i
a
ng,
a
nd
Y
.
W
a
ng,
“
D
R
L
-
M
4M
R
:
a
n
i
nt
e
l
l
i
ge
nt
m
ul
t
i
c
a
s
t
r
out
i
ng
a
ppr
oa
c
h
ba
s
e
d
on
D
Q
N
de
e
p r
e
i
nf
or
c
e
m
e
nt
l
e
a
r
ni
ng i
n S
D
N
,”
P
hy
s
i
c
al
C
om
m
uni
c
at
i
on
, vol
. 55, 2022, doi
:
10.1016/
j
.phyc
om
.2022.101919.
[
11]
S
.
Y
a
o,
X
.
L
i
n,
J
.
W
a
ng,
Q
.
Z
ha
ng,
a
nd
Z
.
W
a
ng,
“
R
e
t
hi
nki
ng
s
upe
r
vi
s
e
d
l
e
a
r
ni
ng
-
ba
s
e
d
ne
ur
a
l
c
om
bi
na
t
or
i
a
l
opt
i
m
i
z
a
t
i
on
f
or
r
out
i
ng pr
obl
e
m
,”
A
C
M
T
r
ans
ac
t
i
ons
on E
v
ol
ut
i
onar
y
L
e
ar
ni
ng and O
pt
i
m
i
z
at
i
on
, 2024, doi
:
10.1145/
3694690.
Evaluation Warning : The document was created with Spire.PDF for Python.
C
om
put
S
c
i
I
nf
T
e
c
hnol
I
S
S
N
:
2722
-
3221
O
pt
imi
z
in
g i
nt
e
r
c
onne
c
ti
on c
al
l
r
out
in
g:
a m
ac
hi
ne
l
e
ar
ni
ng ap
pr
oac
h f
or
c
o
s
t
and
…
(
I
v
y
A
ne
s
u M
udar
i
)
65
[
12]
D
.
T
ur
l
ykoz
ha
ye
va
e
t
al
.
,
“
E
va
l
ua
t
i
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
-
ba
s
e
d
r
out
i
ng
a
l
gor
i
t
hm
s
on
va
r
i
ous
w
i
r
e
l
e
s
s
ne
t
w
or
k
t
opol
ogi
e
s
,”
i
n
P
hot
oni
c
s
A
ppl
i
c
at
i
ons
i
n
A
s
t
r
ono
m
y
,
C
om
m
uni
c
at
i
ons
,
I
ndus
t
r
y
,
and
H
i
gh
E
ne
r
gy
P
hy
s
i
c
s
E
x
pe
r
i
m
e
nt
s
,
D
e
c
.
2024
.
doi
:
10.1117/
12.3058676.
[
13]
R
.
B
out
a
ba
e
t
al
.
,
“
A
c
om
pr
e
he
ns
i
ve
s
ur
ve
y
on
m
a
c
hi
ne
l
e
a
r
ni
ng
f
or
n
e
t
w
or
ki
ng:
e
vol
ut
i
on,
a
ppl
i
c
a
t
i
ons
a
nd
r
e
s
e
a
r
c
h
oppor
t
uni
t
i
e
s
,”
J
our
nal
of
I
nt
e
r
ne
t
Se
r
v
i
c
e
s
and A
ppl
i
c
at
i
ons
, vol
. 9, no. 1, 201
8, doi
:
10.1186/
s
13174
-
018
-
0087
-
2.
[
14]
B
.
S
a
ha
na
,
A
.
A
ga
r
w
a
l
,
A
.
A
m
e
t
a
,
A
.
B
ha
t
i
,
B
.
S
a
dha
na
,
a
nd
A
.
D
e
s
ha
pa
nde
,
“
O
pt
i
m
a
l
ne
t
w
or
k
r
out
e
e
s
t
i
m
a
t
or
us
i
ng
pr
e
di
c
t
i
on
a
l
gor
i
t
hm
s
,”
2020
I
E
E
E
I
nt
e
r
nat
i
onal
C
onf
e
r
e
n
c
e
on
E
l
e
c
t
r
oni
c
s
,
C
om
put
i
n
g
and
C
om
m
uni
c
at
i
on
T
e
c
hnol
ogi
e
s
(
C
O
N
E
C
C
T
)
,
pp. 1
–
5, J
ul
. 2020, doi
:
10.1109/
C
O
N
E
C
C
T
50063.2020.9198468.
[
15]
A
.
M
.
H
a
i
l
a
n
a
nd
H
.
A
l
s
ha
he
e
n,
“
S
ur
ve
y
of
t
r
a
f
f
i
c
e
ngi
ne
e
r
i
ng
s
ol
ut
i
ons
f
or
a
t
e
l
e
c
om
m
uni
c
a
t
i
on
ne
t
w
or
k
opt
i
m
i
z
a
t
i
on,”
T
i
k
r
i
t
J
our
nal
of
E
ngi
ne
e
r
i
ng Sc
i
e
nc
e
s
, vol
. 31, no. 2, pp. 128
–
137, 2024, doi
:
10.25130/
t
j
e
s
.31.2.12.
[
16]
K
.
Y
um
a
nd
B
.
Y
oo,
“
T
he
i
m
pa
c
t
of
s
e
r
vi
c
e
qua
l
i
t
y
on
c
u
s
t
om
e
r
l
oya
l
t
y
t
hr
ough
c
us
t
om
e
r
s
a
t
i
s
f
a
c
t
i
on
i
n
m
obi
l
e
s
oc
i
a
l
m
e
di
a
,
”
Sus
t
ai
nabi
l
i
t
y
, vol
. 15, no. 14, J
ul
. 2023, doi
:
10.3390/
s
u151411214.
[
17]
Z
i
m
ba
bw
e
a
n
G
ove
r
nm
e
nt
G
a
z
e
t
t
e
,
C
y
b
e
r
and
dat
a
pr
ot
e
c
t
i
on
(
l
i
c
e
ns
i
ng
of
d
at
a
c
ont
r
ol
l
e
r
s
and
appoi
nt
m
e
nt
of
dat
a
p
r
ot
e
c
t
i
on
of
f
i
c
e
r
s
)
r
e
gul
at
i
ons
,
S
t
a
t
ut
or
y I
ns
t
r
um
e
nt
155 of
2024
, v
ol
. C
II
, no. 72
,
2024.
[
18]
Z
i
m
ba
bw
e
T
e
a
c
he
r
s
A
s
s
o
c
i
a
t
i
on,
E
duc
at
i
on ac
t
[
c
hapt
e
r
25:
04]
. 2016.
[
19]
Z
i
m
ba
bw
e
T
e
a
c
he
r
s
A
s
s
o
c
i
a
t
i
on,
D
at
a pr
ot
e
c
t
i
on ac
t
[
c
hapt
e
r
11:
12]
. 2021.
[
20]
A
.
B
e
l
l
o,
I
.
A
j
i
ba
de
,
I
.
W
onuol
a
,
a
nd
D
.
E
kw
e
l
i
,
“
D
e
ve
l
opi
ng
a
pr
e
di
c
t
i
ve
m
ode
l
f
or
s
t
ude
nt
a
c
a
de
m
i
c
pe
r
f
or
m
a
nc
e
us
i
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
t
e
c
hni
que
s
,”
I
nt
e
r
nat
i
onal
J
our
nal
of
Sc
i
e
nc
e
and
R
e
s
e
ar
c
h
A
r
c
hi
v
e
,
vol
.
16,
no.
3,
pp.
219
–
234,
2025,
doi
:
10.30574/
i
j
s
r
a
.2025.16.3.2539.
[
21]
X
.
Y
ua
n,
S
.
L
i
u,
W
.
F
e
ng,
a
nd
G
.
D
a
uphi
n,
“
F
e
a
t
ur
e
i
m
por
t
a
nc
e
r
a
nki
ng
of
r
a
ndom
f
or
e
s
t
-
ba
s
e
d
e
nd
-
to
-
e
nd
l
e
a
r
ni
ng
a
l
gor
i
t
hm
,”
R
e
m
ot
e
Se
ns
i
ng
, vol
. 15, no. 21, 2023, doi
:
10.3390/
r
s
15215203.
[
22]
G
. M
a
, J
. C
he
n, J
. L
i
, H
. S
hi
, a
nd Y
. C
h
e
n, “
R
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
i
on m
a
y be
c
om
e
t
he
opt
i
m
a
l
r
e
gr
e
s
s
i
on m
ode
l
f
or
os
t
e
oa
r
t
hr
i
t
i
s
of
t
he
kne
e
i
n
e
l
de
r
l
y,
i
n
t
he
c
ont
e
xt
of
e
m
bodi
e
d
c
ogni
t
i
on
a
nd
ps
yc
hos
om
a
t
i
c
m
e
di
c
i
ne
,”
J
our
nal
of
M
ul
t
i
di
s
c
i
pl
i
nar
y
H
e
al
t
hc
ar
e
, vol
. 18, pp. 4219
–
4232, 2025, doi
:
10.2147/
J
M
D
H
.S
519195.
[
23]
F
.
P
e
dr
e
gos
a
e
t
al
.
,
“
S
c
i
ki
t
-
l
e
a
r
n:
m
a
c
hi
ne
l
e
a
r
ni
ng
i
n
P
yt
hon,”
J
ou
r
nal
of
M
ac
hi
ne
L
e
ar
ni
ng
R
e
s
e
ar
c
h
,
vol
.
12,
no.
85,
pp. 2825
–
2830, 2011.
[
24]
L
.
B
a
r
r
e
ña
da
,
P
.
D
hi
m
a
n,
D
.
T
i
m
m
e
r
m
a
n,
A
.
-
L
.
B
oul
e
s
t
e
i
x,
a
nd
B
.
V
a
n
C
a
l
s
t
e
r
,
“
U
nde
r
s
t
a
ndi
ng
ove
r
f
i
t
t
i
ng
i
n
r
a
ndom
f
or
e
s
t
f
or
pr
oba
bi
l
i
t
y
e
s
t
i
m
a
t
i
on:
a
vi
s
ua
l
i
z
a
t
i
on
a
nd
s
i
m
ul
a
t
i
on
s
t
udy,”
D
i
agno
s
t
i
c
and
P
r
ognos
t
i
c
R
e
s
e
a
r
c
h
,
vol
.
8,
no.
1,
2024
,
doi
:
10.1186/
s
41512
-
024
-
00177
-
1.
[
25]
H
.
A
.
Z
e
i
ni
,
D
.
A
l
-
J
e
z
na
w
i
,
H
.
I
m
r
a
n,
L
.
F
.
A
.
B
e
r
na
r
do,
Z
.
A
l
-
K
ha
f
a
j
i
,
a
nd
K
.
A
.
O
s
t
r
ow
s
ki
,
“
R
a
ndom
f
or
e
s
t
a
l
gor
i
t
hm
f
o
r
t
he
s
t
r
e
ngt
h pr
e
di
c
t
i
on of
ge
opol
ym
e
r
s
t
a
bi
l
i
z
e
d c
l
a
ye
y s
oi
l
,”
Sus
t
ai
nabi
l
i
t
y
, vol
. 1
5, no. 2, J
a
n. 2023, doi
:
10.3390/
s
u15021408.
B
I
O
G
R
A
P
H
I
E
S
O
F
A
U
T
H
O
R
S
Ivy
Anesu
Mudari
is
a
Bachelor
of
Technology
holder
in
Softwar
e
Engineering
from
the
School
of
Information
Science
and
Technology,
Harare
Insti
tute
of
Technology
.
She
is curren
tly pursuing a
master
of technology degree
in cloud comp
utin
g. She also w
orks with
a
telecommunic
ations
company,
which
has
driven
her
to
consider
inter
connection
as
an
area
of
interest. She
can be
contacte
d at email: h
230942f@hit.ac.zw.
Mainford
Mutand
avari
is
a
Ph
.
D
.
scholar
at
SRMIST
University,
India,
and
a
lecturer
and
postgradua
te
studies
coordinato
r
at
the
Harare
Institute
of
Technology
(HIT),
Zimbabwe.
With
advanced
degrees
in
computer
science
and
strate
gy
and
innovation,
his
research
spans
data
analytics,
cybersecurity,
IoT,
AI,
and
cloud
comp
uting.
He
is
a
member
of
HIT’s Cybersecurity and AI
research groups and actively cont
ributes t
o national ICT s
tandards
through
the
Standards
Association
of
Zimbabwe.
He
has
published
widely
on
topics
such
as
data
loss
preventions,
digital
learning
infrastructure,
and
e
-
health
sec
urity.
His
work
bridges
academic
research
with
indust
ry
applicati
ons
by
focusing
on
practi
cal
digital
soluti
ons
for
education
,
telecomm
unicatio
ns,
and
healthcare
in
Zimbabwe.
H
e
is
also
involved
in
curriculum
developm
ent,
postgrad
uate
supervis
ion,
and
build
ing
academic
-
industry
partnerships.
He can be contacted at email:
mmutandavar
i@
hit.ac.zw
.
Kenneth
Chiworera
is
a
lecturer
at
the
Harare
Institute
of
Technology
(HIT),
where
he
coordinates
the
HIT200
projects
and
supe
rvises
undergradu
ate
student’s
bachelor
of
technology
(B
.
Tech
.
)
and
a
master
of
technology
(M
.
Tech
.
)
in
so
ftware
engineering
.
His
research
interests
include
cloud
computing
,
cybersecurity,
artificial
intellig
ence
(AI),
and
blockchain
technology.
He
is
dedicated
to
mentoring
students,
fo
stering
innovation,
and
enhancing
practical
skills
through
project
-
based
learning.
He
actively
participates
in
academic
conferences
and
workshop
s,
contribu
ting
to
advancement
s
in
technolo
gy.
He
can
be
contacted
at email
:
kchiworera@hit.ac.zw
.
Evaluation Warning : The document was created with Spire.PDF for Python.