I
AE
S
I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
Ar
t
if
icial
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
Vol.
14
,
No.
4
,
Augus
t
2025
,
pp.
3063
~
3073
I
S
S
N:
2252
-
8938
,
DO
I
:
10
.
11591/i
jai
.
v
14
.i
4
.
pp
30
63
-
3073
3063
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
ai
.
iaes
c
or
e
.
c
om
A
r
t
ific
ia
l
in
t
e
ll
ig
e
n
c
e
p
r
e
d
i
c
t
i
ve
m
od
e
li
n
g f
o
r
e
d
u
c
at
io
n
al
in
d
ic
at
o
r
s u
si
n
g
d
at
a
p
r
o
f
il
in
g t
e
c
h
n
i
q
u
e
s
S
ou
k
ain
a
Nai
1
,
B
ah
aa
E
d
d
in
e
E
lb
agh
az
aou
i
1
,
2
,
Am
al
Rif
ai
3
,
Abd
e
lal
i
m
S
ad
iq
1
1
L
a
bor
a
to
r
y of
C
omput
e
r
S
c
ie
nc
e
s
, F
a
c
ul
ty
of
S
c
ie
nc
e
s
, I
bn
T
o
f
a
il
U
ni
ve
r
s
it
y, K
e
ni
tr
a
, M
or
oc
c
o
2
N
a
ti
ona
l
S
c
hool
of
A
ppl
ie
d S
c
ie
nc
e
s
,
S
ul
ta
n M
oul
a
y S
li
ma
ne
U
ni
ve
r
s
it
y, B
e
ni
M
e
ll
a
l,
M
or
oc
c
o
3
L
a
bor
a
to
r
y of
R
e
s
e
a
r
c
h i
n E
ngi
ne
e
r
in
g of
C
omput
in
g E
nvi
r
on
me
nt
f
or
H
uma
n
L
e
a
r
ni
ng
,
R
e
gi
ona
l
C
e
nt
e
r
f
or
t
he
P
r
of
e
s
s
io
ns
of
E
duc
a
ti
on a
nd T
r
a
in
in
g
,
R
a
ba
t,
M
or
oc
c
o
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Oc
t
24,
2024
R
e
vis
e
d
J
un
16,
2025
Ac
c
e
pted
J
ul
10,
2025
In
Mo
r
o
cco
,
t
h
e
e
s
cal
a
t
i
n
g
c
h
al
l
en
g
es
i
n
t
h
e
ed
u
cat
i
o
n
s
ect
o
r
u
n
d
ers
c
o
re
t
h
e
n
eces
s
i
t
y
f
o
r
p
rec
i
s
e
p
re
d
i
c
t
i
o
n
s
an
d
i
n
fo
rme
d
d
e
ci
s
i
o
n
-
mak
i
n
g
.
E
ffect
i
v
e
man
ag
eme
n
t
o
f
t
h
e
ed
u
cat
i
o
n
s
y
s
t
em
d
e
p
en
d
s
o
n
ro
b
u
s
t
s
t
a
t
i
s
t
i
cal
d
at
a,
w
h
i
ch
i
s
cru
c
i
al
fo
r
g
u
i
d
i
n
g
d
eci
s
i
o
n
s
,
refi
n
i
n
g
p
o
l
i
ci
e
s
,
an
d
i
mp
r
o
v
i
n
g
b
o
t
h
t
h
e
q
u
al
i
t
y
an
d
acces
s
i
b
i
l
i
t
y
o
f
e
d
u
ca
t
i
o
n
.
Rel
i
a
b
l
e
i
n
d
i
ca
t
o
r
s
are
v
i
t
a
l
fo
r
en
s
u
ri
n
g
eff
i
ci
e
n
cy
,
e
q
u
i
t
y
,
an
d
acc
u
racy
i
n
ed
u
cat
i
o
n
al
p
l
a
n
n
i
n
g
an
d
d
eci
s
i
o
n
-
mak
i
n
g
.
W
i
t
h
o
u
t
d
ep
e
n
d
a
b
l
e
d
a
t
a,
i
mp
l
eme
n
t
i
n
g
effect
i
v
e
p
o
l
i
c
i
es
,
ad
d
re
s
s
i
n
g
t
h
e
n
ee
d
s
ap
p
ro
p
ri
a
t
el
y
,
an
d
ach
i
ev
i
n
g
p
o
s
i
t
i
v
e
o
u
t
c
o
m
es
b
eco
me
s
d
i
ff
i
cu
l
t
.
T
h
i
s
p
ap
er
ai
m
s
t
o
i
d
en
t
i
f
y
t
h
e
o
p
t
i
mal
mach
i
n
e
l
earn
i
n
g
mo
d
e
l
fo
r
an
al
y
zi
n
g
e
d
u
ca
t
i
o
n
a
l
i
n
d
i
cat
o
rs
b
y
c
o
mp
ari
n
g
a
ra
n
g
e
o
f
ad
v
a
n
ced
m
o
d
e
l
s
acr
o
s
s
a
co
m
p
reh
e
n
s
i
v
e
s
et
o
f
met
r
i
cs
.
T
h
e
o
b
j
ec
t
i
v
e
i
s
t
o
d
et
erm
i
n
e
t
h
e
mo
s
t
effect
i
v
e
mo
d
el
fo
r
p
r
o
fi
l
i
n
g
re
l
ev
an
t
i
n
fo
rma
t
i
o
n
an
d
ad
d
re
s
s
i
n
g
p
red
i
ct
i
v
e
c
h
al
l
en
g
es
w
i
t
h
h
i
g
h
accu
rac
y
.
K
e
y
w
o
r
d
s
:
Ac
a
de
mi
c
s
uppor
t
Ar
ti
f
icia
l
int
e
ll
igenc
e
Da
ta
pr
of
il
ing
E
duc
a
ti
on
M
a
c
hine
lea
r
ning
P
r
e
diction
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
S
ouka
ina
Na
i
L
a
bor
a
tor
y
of
C
omput
e
r
S
c
ienc
e
s
,
F
a
c
ult
y
o
f
S
c
ie
nc
e
s
,
I
bn
T
o
f
a
il
Unive
r
s
it
y
Ke
nit
r
a
14000,
M
or
oc
c
o
E
mail:
s
ouka
ina.
na
i@gm
a
il
.
c
om
1.
I
NT
RODU
C
T
I
ON
I
n
M
or
oc
c
o,
the
e
duc
a
ti
on
s
e
c
tor
f
a
c
e
s
numer
ous
c
ha
ll
e
nge
s
that
r
e
quir
e
p
r
e
c
is
e
,
da
ta
-
dr
iven
int
e
r
ve
nti
on
[
1
]
.
T
he
r
a
pid
r
is
e
o
f
is
s
ue
s
s
uc
h
a
s
e
duc
a
ti
ona
l
a
c
c
e
s
s
,
qua
li
ty
dis
pa
r
it
ies
,
a
nd
r
e
s
our
c
e
a
ll
oc
a
ti
on
ha
s
he
ight
e
ne
d
the
ne
e
d
f
or
a
c
c
ur
a
te
f
or
e
c
a
s
ti
ng
a
nd
pr
e
dictive
a
na
lys
is
to
inf
or
m
poli
c
y
d
e
c
is
ions
[
2]
.
E
f
f
e
c
ti
ve
mana
ge
ment
of
e
duc
a
ti
on
s
ys
tems
hinges
on
the
a
va
il
a
bil
it
y
of
r
e
li
a
ble
s
tatis
ti
c
a
l
da
ta
[
3]
,
whic
h
not
only
inf
o
r
ms
poli
c
ymake
r
s
but
a
ls
o
play
s
a
pivot
a
l
r
ole
in
c
r
a
f
ti
ng
s
tr
a
tegie
s
to
e
nha
nc
e
the
ove
r
a
ll
qua
li
ty
a
nd
e
quit
y
o
f
e
duc
a
ti
on
[
4]
.
How
e
ve
r
,
a
c
hieving
thes
e
goa
ls
de
pe
nds
on
the
r
obus
tnes
s
a
nd
r
e
li
a
bil
it
y
of
e
duc
a
ti
ona
l
indi
c
a
tor
s
[
5]
.
E
duc
a
ti
ona
l
indi
c
a
tor
s
,
s
uc
h
a
s
a
c
a
de
mi
c
s
uppor
t
[
6
]
,
dr
opout
[
7
]
,
e
nr
oll
ment
r
a
tes
[
8
]
,
li
te
r
a
c
y
leve
ls
[
9]
,
gove
r
nme
nt
e
xpe
ndit
ur
e
[
10
]
,
a
nd
s
tudent
pe
r
f
or
manc
e
metr
ics
[
11
]
,
a
r
e
f
unda
menta
l
in
a
s
s
e
s
s
ing
the
e
f
f
icie
nc
y
a
nd
e
quit
y
o
f
e
duc
a
ti
on
a
l
s
ys
tems
[
12]
.
T
he
a
bs
e
nc
e
of
de
pe
nda
ble
da
ta
c
a
n
s
e
ve
r
e
ly
hinder
the
a
bil
it
y
to
im
pleme
nt
e
f
f
e
c
t
ive
poli
c
ies
a
nd
make
inf
or
med
de
c
is
ions
[
13]
,
lea
ding
to
s
ubopti
mal
o
utcome
s
a
nd
mi
s
s
e
d
oppor
tuni
ti
e
s
f
or
im
pr
ove
ment
[
14]
.
T
hus
,
ther
e
is
a
p
r
e
s
s
ing
ne
e
d
f
or
a
dva
nc
e
d
a
na
lyt
i
c
a
l
tool
s
c
a
pa
ble
of
pr
of
i
li
ng
[
15]
,
a
nd
p
r
e
dictin
g
ke
y
t
r
e
nds
withi
n
the
e
duc
a
ti
ona
l
lands
c
a
pe
[
16]
.
T
his
pa
pe
r
e
xplor
e
s
the
a
ppli
c
a
ti
on
of
mac
hine
lea
r
ning
tec
hniques
to
e
duc
a
ti
ona
l
da
ta
a
na
lys
is
[
17]
,
with
the
pr
im
a
r
y
a
im
o
f
identif
ying
the
opt
im
a
l
p
r
e
dictive
model
f
or
a
na
lyzing
a
nd
f
or
e
c
a
s
ti
ng
e
du
c
a
ti
ona
l
s
tatis
ti
c
s
in
M
or
oc
c
o.
B
y
c
ompar
ing
a
va
r
iet
y
of
a
dva
nc
e
d
ma
c
hine
lea
r
n
ing
a
lgor
i
thm
s
a
c
r
os
s
a
c
ompr
e
he
ns
ive
s
e
t
of
e
duc
a
ti
ona
l
indi
c
a
tor
s
[
18]
–
[
20]
,
thi
s
s
tudy
s
e
e
ks
to
de
ter
mi
ne
whic
h
model
o
f
f
e
r
s
the
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
306
3
-
3073
3064
highes
t
a
c
c
ur
a
c
y
.
I
t
a
ls
o
a
im
s
to
a
s
s
e
s
s
the
r
e
li
a
bil
it
y
in
p
r
e
dicting
tr
e
nds
a
nd
guidi
ng
de
c
is
ion
-
maki
ng
in
the
e
duc
a
ti
on
s
e
c
tor
.
2.
RE
L
AT
E
D
WORKS
T
he
us
e
of
pr
e
dictive
modeling
a
nd
mac
hine
l
e
a
r
ning
tec
hniques
in
the
e
duc
a
ti
on
s
e
c
tor
ha
s
ga
r
ne
r
e
d
incr
e
a
s
ing
a
tt
e
nti
on
in
r
e
c
e
nt
ye
a
r
s
[
2
1]
.
R
e
s
e
a
r
c
he
r
s
ha
ve
e
xplor
e
d
va
r
ious
methodol
ogies
to
a
na
lyze
e
duc
a
ti
ona
l
da
ta
[
22]
,
de
ve
lop
pr
e
dictive
i
ns
ight
s
[
23]
,
a
nd
a
id
de
c
is
ion
-
make
r
s
in
a
ddr
e
s
s
ing
c
r
it
ica
l
c
ha
ll
e
nge
s
,
s
uc
h
a
s
r
e
s
our
c
e
a
ll
oc
a
ti
on
[
24]
,
s
tudent
pe
r
f
or
manc
e
[
25]
,
a
nd
e
duc
a
ti
ona
l
a
c
c
e
s
s
[
26]
.
T
his
s
e
c
ti
on
r
e
view
s
ke
y
s
tudi
e
s
a
nd
a
dva
nc
e
s
in
the
f
ield,
highl
ight
ing
r
e
leva
nt
mac
hine
lea
r
ning
models
a
nd
their
a
ppli
c
a
ti
on
to
e
duc
a
ti
ona
l
indi
c
a
tor
s
[
27]
.
One
pr
omi
ne
nt
a
r
e
a
of
r
e
s
e
a
r
c
h
f
oc
us
e
s
on
the
p
r
e
dictive
modeling
o
f
s
tudent
pe
r
f
or
manc
e
[
28]
.
S
tudi
e
s
s
uc
h
a
s
thos
e
by
[
29]
,
[
30]
ha
ve
e
xplor
e
d
mac
hine
lea
r
ning
a
lgor
it
hms
,
including
de
c
is
ion
tr
e
e
s
,
r
a
ndom
f
o
r
e
s
ts
,
a
nd
s
uppor
t
ve
c
tor
mac
hines
,
to
p
r
e
dict
a
c
a
de
mi
c
outcome
s
ba
s
e
d
on
f
a
c
tor
s
s
uc
h
a
s
s
tudent
de
mogr
a
phics
,
pr
ior
a
c
a
de
mi
c
his
tor
y,
a
nd
s
oc
io
-
e
c
onomi
c
s
tatus
.
T
he
s
e
models
ha
ve
de
mons
tr
a
ted
c
ons
ider
a
ble
a
c
c
ur
a
c
y
in
f
or
e
c
a
s
ti
ng
s
tudent
s
uc
c
e
s
s
a
nd
identif
ying
a
t
-
r
is
k
s
tudents
,
pr
ovidi
ng
va
luable
ins
ight
s
f
or
e
duc
a
tor
s
a
nd
a
dmi
nis
tr
a
to
r
s
.
Anothe
r
s
igni
f
ica
nt
li
ne
o
f
r
e
s
e
a
r
c
h
invol
ve
s
us
ing
mac
hine
lea
r
ning
to
p
r
e
dict
s
c
hool
dr
opout
r
a
tes
.
F
or
e
xa
mpl
e
,
Ala
m
e
t
al
.
[
31]
uti
li
z
e
d
logi
s
ti
c
r
e
g
r
e
s
s
ion
a
nd
ne
ur
a
l
ne
twor
ks
to
identif
y
s
tudents
a
t
r
is
k
of
dr
opping
out
of
s
c
hool.
T
he
s
tudy
f
ound
that
mac
h
ine
lea
r
ning
models
c
ould
e
f
f
e
c
ti
ve
ly
pr
e
dict
d
r
op
out
r
a
tes
by
a
na
lyzing
his
tor
ica
l
e
nr
oll
ment
da
ta,
a
t
ten
da
nc
e
r
e
c
or
ds
,
a
nd
s
oc
io
-
e
c
onomi
c
f
a
c
tor
s
.
S
im
il
a
r
ly,
Kule
to
e
t
al
.
[
32]
e
xpa
nde
d
thi
s
wor
k
by
incor
p
or
a
ti
ng
a
ddit
ional
pr
e
dictor
s
s
uc
h
a
s
s
tudent
e
nga
ge
ment
metr
ics
a
nd
tea
c
he
r
a
s
s
e
s
s
ment
s
,
f
ur
ther
im
pr
ovin
g
the
a
c
c
ur
a
c
y
of
dr
opout
pr
e
dictions
.
R
e
s
e
a
r
c
h
ha
s
a
ls
o
inves
ti
ga
ted
the
opti
mi
z
a
ti
on
of
r
e
s
our
c
e
a
ll
oc
a
ti
on
a
nd
e
xpe
ndit
u
r
e
in
the
e
duc
a
ti
on
s
e
c
tor
.
T
he
r
e
s
e
a
r
c
h
in
[
33
]
,
[
34]
h
a
ve
e
mpl
oye
d
mac
hine
lea
r
ning
tec
hniques
to
a
na
lyze
e
duc
a
ti
ona
l
e
xpe
ndit
ur
e
s
,
s
taf
f
ing
pa
tt
e
r
ns
,
a
nd
s
c
hool
inf
r
a
s
tr
uc
tur
e
ne
e
ds
.
T
he
s
e
wor
ks
de
mon
s
tr
a
te
that
pr
e
dictive
modeling
c
a
n
s
uppor
t
mo
r
e
e
f
f
icie
nt
a
nd
e
quit
a
ble
dis
tr
ibut
ion
o
f
e
duc
a
ti
ona
l
r
e
s
our
c
e
s
,
e
ns
ur
ing
that
f
inanc
ial
a
nd
hu
man
c
a
pit
a
l
is
a
ll
oc
a
ted
whe
r
e
it
is
mos
t
ne
e
de
d
[
35]
.
I
n
the
c
ontext
of
e
duc
a
ti
ona
l
poli
c
y
planning,
mac
hine
lea
r
ning
models
ha
ve
be
e
n
a
ppli
e
d
to
f
or
e
c
a
s
t
tr
e
nds
s
uc
h
a
s
e
nr
oll
ment
r
a
tes
,
li
ter
a
c
y
l
e
ve
ls
,
a
nd
the
de
mand
f
or
tea
c
he
r
s
.
F
or
ins
tanc
e
,
gr
oup
of
r
e
s
e
a
r
c
he
r
de
ve
loped
a
ti
me
-
s
e
r
ies
f
or
e
c
a
s
ti
ng
mo
de
l
us
ing
r
e
c
ur
r
e
nt
ne
ur
a
l
ne
twor
ks
(
R
NN
)
to
pr
e
d
ict
f
utur
e
s
tudent
e
nr
oll
ment
a
c
r
os
s
dif
f
e
r
e
nt
e
duc
a
ti
ona
l
l
e
ve
ls
[
36]
.
T
he
ir
model
pr
ov
ided
a
c
ti
ona
ble
ins
i
ghts
int
o
e
xpe
c
ted
c
ha
nge
s
in
e
nr
oll
ment
tr
e
nds
,
a
ll
owing
p
oli
c
ymake
r
s
to
a
djus
t
s
tr
a
tegie
s
a
n
d
r
e
s
our
c
e
s
a
c
c
or
dingl
y.
S
im
il
a
r
ly,
Ka
ngiwa
e
t
al.
[
37
]
e
xplo
r
e
d
how
da
ta
-
dr
ive
n
de
c
is
ion
-
making
c
ould
e
nha
nc
e
the
e
f
f
e
c
ti
v
e
ne
s
s
of
e
duc
a
ti
ona
l
int
e
r
ve
nti
ons
,
s
ugge
s
ti
ng
that
pr
e
dictive
models
c
ould
s
e
r
ve
a
s
powe
r
f
ul
tool
s
f
or
s
ha
ping
na
ti
ona
l
e
duc
a
ti
ona
l
poli
c
ies
.
De
s
pit
e
the
gr
owing
body
of
r
e
s
e
a
r
c
h
in
thi
s
f
i
e
ld,
ther
e
r
e
main
c
ha
ll
e
nge
s
in
the
a
c
c
ur
a
te
a
nd
r
e
li
a
ble
pr
e
diction
of
e
duc
a
ti
ona
l
indi
c
a
tor
s
[
38]
,
pa
r
ti
c
ular
ly
whe
n
de
a
li
ng
with
la
r
ge
a
nd
diver
s
e
da
tas
e
ts
.
I
ntegr
a
ti
ng
da
ta
pr
o
f
il
ing
with
mac
hine
lea
r
ning
ha
s
the
potential
to
im
pr
ove
the
a
c
c
ur
a
c
y
a
nd
us
a
bi
li
ty
of
pr
e
dictive
models
in
the
e
duc
a
ti
on
s
e
c
tor
[
39]
.
A
s
the
a
va
il
a
bil
it
y
of
e
duc
a
ti
ona
l
da
ta
c
onti
nue
s
to
e
xpa
nd,
r
e
s
e
a
r
c
he
r
s
a
r
e
incr
e
a
s
ingl
y
f
oc
us
ing
on
r
e
f
ini
ng
t
he
s
e
models
to
a
ddr
e
s
s
c
ompl
e
x
is
s
ue
s
s
uc
h
a
s
in
e
qua
li
ty
[
40]
,
dr
opout
r
a
tes
[
41]
,
a
nd
the
e
f
f
icie
nt
a
ll
oc
a
ti
o
n
of
r
e
s
our
c
e
s
[
42]
.
T
his
pa
pe
r
buil
ds
on
thi
s
body
of
wor
k
by
e
mpl
oying
a
c
ompar
a
ti
ve
a
ppr
oa
c
h
to
mac
hine
lea
r
ning
models
[
43]
,
with
the
goa
l
o
f
identif
ying
the
mos
t
s
uit
a
ble
tec
hniques
f
or
pr
of
il
ing
a
nd
pr
e
dicting
e
d
uc
a
ti
ona
l
indi
c
a
tor
s
in
M
or
oc
c
o
[
44]
.
B
y
a
na
lyzing
a
wide
r
a
nge
of
indi
c
a
tor
s
,
thi
s
s
tudy
a
im
s
to
c
ontr
ibut
e
to
the
ongoing
de
ve
lopm
e
nt
of
AI
-
dr
iven
s
olut
ions
.
T
he
s
e
s
ol
uti
ons
c
a
n
a
ddr
e
s
s
the
c
r
it
ica
l
c
ha
ll
e
nge
s
f
a
c
e
d
by
the
e
duc
a
ti
on
s
e
c
tor
[
45]
,
pa
r
ti
c
ular
ly
in
the
c
ontext
o
f
de
ve
lopi
ng
c
ountr
ies
[
46
]
.
3.
PR
OB
L
E
M
AT
I
C
I
n
M
or
oc
c
o,
the
e
duc
a
ti
on
s
e
c
tor
f
a
c
e
s
dyna
mi
c
c
ha
ll
e
nge
s
that
de
mand
pr
e
c
i
s
e
a
nd
f
or
wa
r
d
-
looki
ng
da
ta
to
guide
poli
c
y
a
nd
im
pr
ove
outco
mes
[
47]
.
R
e
li
a
ble
f
or
e
c
a
s
ti
ng
o
f
e
duc
a
ti
ona
l
ind
ica
tor
s
is
e
s
s
e
nti
a
l
f
or
e
f
f
e
c
ti
ve
de
c
is
ion
-
making
[
48]
,
pa
r
ti
c
ular
ly
in
a
ddr
e
s
s
ing
is
s
ue
s
r
e
late
d
to
a
c
c
e
s
s
,
qua
li
ty,
a
nd
r
e
s
our
c
e
a
ll
oc
a
ti
on
[
49
]
.
How
e
ve
r
,
the
s
e
lec
ti
on
of
a
n
a
pp
r
opr
iate
mac
hine
lea
r
ning
model
f
or
pr
e
dicting
f
utur
e
e
duc
a
ti
ona
l
va
lues
pr
e
s
e
nts
a
c
ompl
e
x
pr
obl
e
m
[
50]
.
T
he
c
or
e
is
s
ue
li
e
s
in
de
ter
mi
ning
whic
h
mac
hine
lea
r
ning
model
c
a
n
de
li
ve
r
the
mos
t
a
c
c
ur
a
te
a
nd
de
pe
nda
ble
f
or
e
c
a
s
ts
ba
s
e
d
on
his
tor
ica
l
e
duc
a
ti
o
na
l
da
ta.
T
his
s
tudy
a
im
s
to
a
ddr
e
s
s
thi
s
pr
oblem
by
a
na
ly
z
ing
a
r
a
nge
o
f
e
duc
a
ti
ona
l
indi
c
a
to
r
s
li
ter
a
c
y
lev
e
ls
,
a
nd
a
c
a
de
mi
c
pe
r
f
or
manc
e
metr
ics
,
d
r
opout,
a
c
a
de
mi
c
s
uppor
t
a
nd
dis
c
us
s
ing
their
mea
nings
a
nd
im
pl
ica
ti
ons
,
a
nd
a
pplyi
ng
va
r
ious
a
dva
nc
e
d
mac
hine
lea
r
ning
models
to
pr
e
dict
f
u
tur
e
t
r
e
nds
.
T
he
r
e
s
e
a
r
c
h
will
invol
ve
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
A
r
ti
fi
c
ial
int
e
ll
igenc
e
pr
e
dictive
mode
li
ng
for
e
duc
ati
onal
indi
c
ator
s
us
ing
data
…
(
Souk
aina
N
ai
)
3065
c
ompar
ing
models
on
their
p
r
e
dictive
a
c
c
ur
a
c
y
a
nd
s
uit
a
bil
it
y
f
or
dif
f
e
r
e
nt
types
o
f
e
duc
a
ti
o
na
l
da
ta.
Additi
ona
ll
y,
the
s
tudy
a
im
s
to
identif
y
c
ha
ll
e
nge
s
r
e
late
d
to
da
ta
qua
li
ty
a
nd
model
pe
r
f
or
manc
e
,
of
f
e
r
ing
ins
ight
s
int
o
the
pr
a
c
ti
c
a
l
a
ppli
c
a
ti
on
of
mac
hine
lea
r
ning
in
e
duc
a
ti
ona
l
planning.
B
y
pin
point
ing
the
mos
t
e
f
f
e
c
ti
ve
mac
hine
lea
r
ning
model
f
or
t
his
pur
pos
e
,
the
s
tudy
will
c
ontr
ibut
e
to
mor
e
i
nf
or
med
de
c
is
ion
-
making
a
nd
s
tr
a
tegic
planning
in
M
or
oc
c
o’
s
e
duc
a
ti
on
s
e
c
tor
,
ult
im
a
tely
s
uppor
ti
ng
e
f
f
or
ts
to
e
nha
nc
e
e
duc
a
ti
ona
l
qua
li
ty
a
nd
e
quit
y.
4.
M
E
T
HO
D
OL
OG
Y
4
.
1.
Anal
ys
is
of
e
d
u
c
at
ion
al
d
a
t
a
an
d
in
d
icat
or
s
T
his
s
tudy
a
im
s
to
a
na
lyze
his
tor
ica
l
e
duc
a
ti
ona
l
da
ta
a
nd
a
pply
va
r
ious
a
dva
nc
e
d
mac
hine
lea
r
ning
models
to
pr
e
dict
f
utur
e
e
duc
a
ti
ona
l
indi
c
a
to
r
s
in
M
or
oc
c
o,
including
li
te
r
a
c
y
leve
ls
,
a
c
a
de
mi
c
pe
r
f
or
manc
e
metr
ics
,
dr
opout
r
a
tes
,
a
nd
pa
r
ti
c
ipation
in
s
upp
o
r
t
pr
ogr
a
ms
.
T
he
objec
ti
ve
is
to
e
va
luate
a
nd
c
o
mpar
e
the
pr
e
dictive
a
c
c
ur
a
c
y
a
nd
r
e
li
a
bil
it
y
o
f
d
if
f
e
r
e
nt
mac
hine
lea
r
ning
models
in
va
r
ious
e
duc
a
ti
ona
l
c
ontexts
.
Additi
ona
ll
y,
the
r
e
s
e
a
r
c
h
will
e
xplor
e
c
ha
ll
e
nge
s
r
e
late
d
to
da
ta
qua
li
ty
a
nd
model
pe
r
f
or
manc
e
,
of
f
e
r
ing
a
c
ti
ona
ble
ins
ight
s
f
or
e
nha
nc
ing
de
c
is
ion
-
making
a
nd
s
tr
a
tegic
planning
withi
n
M
or
oc
c
o’
s
e
duc
a
ti
on
s
e
c
tor
.
Ulti
mate
ly,
th
is
s
tudy
s
e
e
ks
to
im
p
r
ove
e
duc
a
ti
o
na
l
qua
li
ty
a
nd
e
quit
y
by
s
uppor
ti
ng
the
de
ve
lop
ment
of
inf
or
med,
da
ta
-
dr
iven
poli
c
i
e
s
.
4
.
2.
Dat
as
e
t
I
n
r
e
c
e
nt
ye
a
r
s
,
the
W
or
ld
B
a
nk
ha
s
made
s
i
gnif
ica
nt
s
tr
ides
in
e
nha
nc
ing
tr
a
ns
pa
r
e
nc
y
a
nd
a
c
c
e
s
s
ibi
li
ty
of
e
duc
a
ti
ona
l
da
ta
globally.
As
pa
r
t
of
thi
s
ini
ti
a
ti
ve
,
the
W
or
ld
B
a
nk
ha
s
publi
s
he
d
a
r
a
nge
of
s
tatis
ti
c
s
s
pe
c
if
ica
ll
y
f
oc
us
e
d
on
e
duc
a
ti
on
in
M
or
oc
c
o
[
18]
.
T
he
s
e
s
tatis
ti
c
s
,
a
va
il
a
ble
thr
oug
h
t
he
ir
da
ta
r
e
pos
it
or
y
a
t
W
or
ld
B
a
nk
e
duc
a
ti
on
indi
c
a
tor
s
f
or
M
or
oc
c
o,
pr
ovidi
ng
a
va
luable
r
e
s
our
c
e
f
o
r
s
take
holder
s
int
e
r
e
s
ted
in
unde
r
s
tanding
the
dyna
mi
c
s
o
f
M
or
oc
c
o’
s
e
duc
a
ti
on
s
ys
tem.
4
.
3
.
I
n
d
icat
or
s
T
he
da
tas
e
t
include
s
de
tailed
indi
c
a
tor
s
on
va
r
io
us
a
s
pe
c
ts
of
e
duc
a
ti
on
(
253
indi
c
a
tor
s
)
,
s
uc
h
a
s
e
nr
oll
ment
r
a
tes
,
e
duc
a
ti
ona
l
a
t
tainment,
a
nd
the
a
va
il
a
bil
it
y
of
e
duc
a
ti
ona
l
r
e
s
our
c
e
s
.
T
he
s
e
indi
c
a
tor
s
a
r
e
c
r
uc
ial
f
or
e
va
luating
the
pe
r
f
or
manc
e
of
the
e
d
uc
a
ti
on
s
ys
tem,
identif
ying
tr
e
nds
,
a
nd
making
inf
or
med
de
c
is
ions
to
e
nha
nc
e
e
duc
a
ti
ona
l
poli
c
ies
a
nd
pr
a
c
ti
c
e
s
.
Among
the
253
indi
c
a
tor
s
,
we
f
oc
us
e
d
on
thos
e
r
e
late
d
to
a
c
a
de
mi
c
s
uppor
t.
T
o
a
c
hieve
thi
s
,
w
e
e
xtr
a
c
ted
the
r
e
leva
nt
da
ta
f
r
om
the
main
f
i
le.
T
a
ble
1
pr
e
s
e
nts
the
r
e
s
ult
s
of
thi
s
e
xtr
a
c
ti
o
n.
I
ndica
tor
s
f
or
a
s
s
e
s
s
ing
e
duc
a
ti
ona
l
outcome
s
include
s
e
ve
r
a
l
ke
y
a
r
e
a
s
.
F
i
r
s
tl
y,
no
e
duc
a
ti
on
/
dr
opout
indi
c
a
tor
s
mea
s
ur
e
the
pe
r
c
e
ntage
of
indi
v
iduals
who
ha
ve
e
it
he
r
r
e
c
e
ived
n
o
f
or
mal
e
duc
a
ti
on
or
a
r
e
not
c
ur
r
e
ntl
y
e
nr
oll
e
d
in
s
c
hool,
r
e
f
lec
ti
ng
the
pr
e
va
lenc
e
of
e
duc
a
ti
ona
l
dis
e
nga
ge
ment.
S
e
c
ondly,
tea
c
he
r
t
r
a
ini
ng
indi
c
a
tor
s
a
s
s
e
s
s
the
qua
li
f
ica
ti
ons
a
nd
pr
of
e
s
s
i
ona
l
de
ve
lopm
e
nt
of
e
duc
a
tor
s
,
whic
h
dir
e
c
tl
y
im
pa
c
t
the
qua
li
ty
of
a
c
a
de
mi
c
s
uppor
t
pr
ovided
to
s
tudents
.
E
duc
a
ti
on
e
xpe
ndit
ur
e
i
ndica
tor
s
qua
nti
f
y
the
f
inanc
ial
r
e
s
our
c
e
s
a
ll
oc
a
ted
to
e
duc
a
ti
on,
inf
luenc
ing
the
a
va
il
a
bil
it
y
o
f
e
s
s
e
nti
a
l
r
e
s
our
c
e
s
a
nd
s
up
por
t
withi
n
s
c
hools
.
Additi
ona
ll
y
,
be
ne
f
it
ing
f
r
om
s
uppor
t
p
r
ogr
a
ms
t
r
a
c
ks
the
pr
opo
r
ti
on
of
s
tudents
r
e
c
e
ivi
ng
s
uppleme
ntar
y
a
c
a
de
mi
c
a
s
s
is
tanc
e
,
whic
h
c
a
n
e
nha
nc
e
lea
r
ning
outcome
s
.
F
inally
,
int
e
r
na
ti
ona
l
a
s
s
e
s
s
ment
s
e
va
luate
s
tudent
pe
r
f
or
manc
e
on
gl
oba
l
s
c
a
les
,
pr
ovidi
ng
ins
ight
s
int
o
the
e
f
f
e
c
ti
v
e
ne
s
s
of
a
c
a
de
mi
c
s
uppor
t
pr
ogr
a
ms
in
im
pr
oving
e
duc
a
ti
ona
l
a
c
hieve
ments
.
T
a
ble
1.
L
is
t
of
the
indi
c
a
tor
s
r
e
late
d
to
s
c
hool
dr
o
pout
a
nd
a
c
a
de
mi
c
s
uppor
t
I
ndi
c
a
to
r
t
ype
I
ndi
c
a
to
r
c
ode
(
C
S
V
f
il
e
)
D
e
s
c
r
ip
ti
on
N
o e
duc
a
ti
on
(
S
c
hool
dr
opout)
B
A
R
.N
O
E
D
.1519.F
E
.Z
S
, B
A
R
.N
O
E
D
.1519.Z
S
,
B
A
R
.N
O
E
D
.15U
P
.F
E
.Z
S
.
P
e
r
c
e
nt
a
ge
of
in
di
vi
dua
ls
w
it
h
no
e
duc
a
ti
on
in
va
r
io
us
a
ge
gr
oups
, i
mpl
yi
ng s
c
hool
dr
opout.
O
ut
of
s
c
hool
(
D
r
opout)
S
E
.L
P
V
.P
R
I
M
.O
O
S
.F
E
, S
E
.
L
P
V
.P
R
I
M
.O
O
S
.M
A
P
e
r
c
e
nt
a
ge
of
pr
im
a
r
y
s
tu
de
nt
s
w
ho
a
r
e
out
of
s
c
hool
, r
e
f
le
c
ti
ng pote
nt
ia
l
dr
opouts
.
U
ne
nr
ol
lm
e
nt
(
D
r
opout)
S
E
.P
R
M
.U
N
E
R
.Z
S
P
e
r
c
e
nt
a
ge
of
s
tu
de
nt
s
une
nr
ol
le
d
a
t
th
e
pr
im
a
r
y
le
ve
l,
i
mpl
yi
ng a
r
e
la
ti
ons
hi
p t
o
s
c
hool
dr
opout.
T
e
a
c
h
e
r
t
r
a
in
in
g
(
A
c
a
de
mi
c
s
uppor
t)
S
E
.P
R
E
.T
C
A
Q
.Z
S
, S
E
.
P
R
M
.T
C
A
Q
.Z
S
,
S
E
.S
E
C
.T
C
A
Q
.L
O
.Z
S
, S
E
.S
E
C
.T
C
A
Q
.U
P
.Z
S
P
e
r
c
e
nt
a
ge
of
te
a
c
he
r
s
w
ho
a
r
e
tr
a
in
e
d/
qua
li
f
ie
d
a
t
pr
e
-
pr
im
a
r
y, pr
i
ma
r
y, a
nd s
e
c
onda
r
y l
e
ve
ls
.
E
duc
a
ti
on e
xpe
ndi
tu
r
e
(
S
uppor
t)
S
E
.X
P
D
.C
U
R
.T
O
T
L
.Z
S
, S
E
.X
P
D
.P
R
I
M
.Z
S
,
S
E
.X
P
D
.S
E
C
O
.Z
S
, S
E
.X
P
D
.T
E
R
T
.Z
S
P
ubl
ic
e
xpe
ndi
tu
r
e
on
e
duc
a
ti
on
a
s
a
pe
r
c
e
nt
a
ge
of
to
ta
l
gove
r
nme
nt
e
xpe
ndi
tu
r
e
, by e
duc
a
ti
on l
e
ve
l.
B
e
ne
f
it
in
g f
r
om
s
uppor
t
pr
ogr
a
ms
S
E
.L
P
V
.P
R
I
M
.B
M
P
.F
E
, S
E
.L
P
V
.P
R
I
M
.B
M
P
.M
A
P
e
r
c
e
nt
a
ge
of
pr
im
a
r
y
s
tu
de
nt
s
be
ne
f
it
in
g
f
r
om
a
c
a
de
mi
c
s
uppor
t
pr
ogr
a
ms
.
I
nt
e
r
na
ti
ona
l
a
s
s
e
s
s
me
nt
s
(
S
uppor
t)
L
O
.P
I
R
L
S
.R
E
A
.
I
N
T
, L
O
.T
I
M
S
S
.M
A
T
8.I
N
T
P
e
r
f
or
ma
nc
e
in
in
te
r
na
ti
ona
l
a
s
s
e
s
s
m
e
nt
s
(
tr
e
nds
in
in
te
r
na
ti
ona
l
ma
th
e
ma
ti
c
s
a
nd
s
c
ie
nc
e
s
tu
dy
(
T
I
M
S
S
)
,
pr
ogr
e
s
s
in
in
te
r
na
ti
ona
l
r
e
a
di
ng
li
te
r
a
c
y
s
tu
dy
(
P
I
R
L
S
)
)
,
in
d
ir
e
c
tl
y
r
e
f
le
c
ti
ng
a
c
a
de
mi
c
s
uppor
t.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
306
3
-
3073
3066
4
.
4
.
I
m
p
or
t
an
c
e
E
duc
a
ti
ona
l
indi
c
a
tor
s
a
r
e
vit
a
l
tool
s
f
or
mea
s
ur
ing
pr
ogr
e
s
s
,
identif
ying
tr
e
nds
,
a
nd
a
s
s
e
s
s
ing
the
e
f
f
e
c
ti
ve
ne
s
s
of
poli
c
ies
a
nd
int
e
r
ve
nti
ons
in
the
e
duc
a
ti
on
s
e
c
tor
.
T
he
y
p
r
ovide
e
s
s
e
nti
a
l
da
ta
on
ke
y
a
r
e
a
s
s
uc
h
a
s
li
ter
a
c
y
r
a
tes
,
dr
opout
r
a
tes
,
a
nd
a
c
a
de
mi
c
pe
r
f
or
manc
e
,
a
ll
owing
pol
icyma
ke
r
s
a
nd
e
duc
a
tor
s
to
unde
r
s
tand
the
c
ur
r
e
nt
s
tate
of
the
s
ys
tem.
T
he
s
e
i
ndica
tor
s
he
lp
t
r
a
c
k
im
p
r
ove
ments
or
c
ha
ll
e
nge
s
o
ve
r
ti
me
,
e
na
bli
ng
tar
ge
ted
int
e
r
ve
nti
ons
whe
r
e
ne
e
de
d.
F
or
e
xa
mpl
e
,
moni
tor
ing
dr
opou
t
r
a
tes
c
a
n
identif
y
v
ulner
a
ble
r
e
gions
or
populations
that
r
e
quir
e
a
ddit
ional
s
upp
or
t,
whi
le
li
ter
a
c
y
r
a
tes
he
lp
e
va
luate
the
im
pa
c
t
of
na
ti
ona
l
a
nd
loca
l
s
tr
a
tegie
s
.
Additi
ona
ll
y,
indi
c
a
tor
s
e
na
ble
da
ta
-
dr
iven
de
c
is
i
on
-
making,
e
ns
ur
ing
that
r
e
s
our
c
e
s
a
r
e
a
ll
oc
a
ted
e
f
f
icie
ntl
y
a
nd
e
quit
a
bly
.
T
he
y
a
ls
o
p
r
omot
e
a
c
c
ountabili
ty
a
nd
c
onti
nuous
im
pr
ove
ment
in
e
d
uc
a
ti
ona
l
outcome
s
.
Ulti
mate
ly,
e
duc
a
ti
ona
l
indi
c
a
tor
s
a
r
e
c
r
uc
ial
f
or
guidi
ng
s
tr
a
tegic
planning
a
nd
a
c
hievi
ng
goa
ls
r
e
late
d
to
e
duc
a
ti
ona
l
qua
li
ty
,
e
quit
y,
a
nd
a
c
c
e
s
s
.
Ana
lyzing
thes
e
s
tatis
ti
c
s
c
a
n
o
f
f
e
r
c
r
i
ti
c
a
l
ins
ig
hts
int
o
a
r
e
a
s
whe
r
e
im
pr
ove
ments
a
r
e
ne
e
de
d,
he
lp
mon
it
or
the
e
f
f
e
c
ti
ve
ne
s
s
of
c
ur
r
e
nt
e
duc
a
ti
ona
l
poli
c
ies
,
a
nd
guide
f
utur
e
planning
a
nd
int
e
r
ve
nti
ons
.
L
e
ve
r
a
ging
mac
hine
lea
r
ning
tec
hniques
to
a
na
lyze
thi
s
da
ta
c
a
n
f
ur
ther
e
nha
nc
e
the
a
c
c
ur
a
c
y
a
nd
de
pth
of
thes
e
ins
ight
s
,
lea
ding
to
mo
r
e
in
f
or
med
de
c
is
ion
-
ma
king
a
nd
tar
ge
ted
poli
c
y
a
djus
tm
e
nts
.
5.
I
M
P
L
E
M
E
NT
AT
I
ON
5.
1.
E
n
vironm
e
n
t
I
n
th
is
s
tudy,
we
will
us
e
P
ython
a
s
the
pr
im
a
r
y
pr
og
r
a
mm
ing
e
nvir
onment
due
to
it
s
powe
r
f
ul
li
br
a
r
ies
a
nd
tool
s
f
or
da
ta
a
na
lys
is
a
nd
mac
hine
lea
r
ning.
S
pe
c
if
ica
ll
y
,
we
will
leve
r
a
ge
s
c
iki
t
-
lea
r
n
(
s
klea
r
n)
,
a
ve
r
s
a
ti
le
a
nd
us
e
r
-
f
r
iendly
mac
hine
le
a
r
ning
li
b
r
a
r
y
withi
n
P
ython
,
whic
h
of
f
e
r
s
a
wide
r
a
nge
o
f
a
lgor
it
hms
f
or
p
r
e
dictive
modeling.
T
his
e
nvir
onment
a
ll
ows
f
or
s
e
a
ml
e
s
s
da
ta
pr
e
pr
oc
e
s
s
ing
,
model
tr
a
ini
ng,
a
nd
e
va
luation,
e
ns
ur
ing
that
we
c
a
n
e
f
f
i
c
iently
c
ompar
e
va
r
ious
mac
hine
lea
r
ning
models
in
ter
ms
of
their
a
c
c
ur
a
c
y
a
nd
pe
r
f
o
r
manc
e
.
P
y
thon’
s
r
ic
h
e
c
os
ys
tem,
c
ombi
ne
d
with
s
klea
r
n’
s
r
obus
tnes
s
,
pr
ovides
the
idea
l
platf
or
m
f
o
r
a
na
lyzing
e
duc
a
ti
ona
l
ind
ica
tor
s
a
nd
buil
ding
r
e
li
a
ble
p
r
e
dictive
models
to
a
dd
r
e
s
s
the
c
ha
ll
e
nge
s
in
M
or
oc
c
o’
s
e
duc
a
ti
on
s
e
c
tor
.
5.
2.
Dat
a
vis
u
ali
z
at
ion
W
e
a
im
to
pr
e
s
e
nt
t
he
e
xtr
a
c
ted
da
ta
dis
c
us
s
e
d
in
s
e
c
ti
on
4.
3
.
A
pa
r
ti
c
ular
f
oc
us
is
plac
e
d
on
a
c
a
de
mi
c
s
uppor
t
inf
o
r
mation
in
M
or
oc
c
o.
T
h
is
include
s
the
a
na
lys
is
of
da
ta
on
s
uppor
t
pr
ogr
a
m
pa
r
ti
c
ipation
a
s
s
hown
in
F
igur
e
1
,
a
nd
int
e
r
na
ti
on
a
l
a
s
s
e
s
s
ments
a
s
s
hown
in
F
igur
e
2.
F
igur
e
1.
Da
ta
vis
ua
li
z
a
ti
on
f
o
r
be
ne
f
it
ing
f
r
om
s
u
ppor
t
pr
og
r
a
ms
T
he
gr
a
ph
in
F
igu
r
e
1
il
lus
tr
a
tes
the
"
B
e
ne
f
it
ing
f
r
om
s
uppor
t
p
r
ogr
a
ms
"
indi
c
a
tor
s
f
r
o
m
2002
to
2016
with
two
li
ne
s
r
e
pr
e
s
e
nti
ng
dis
ti
nc
t
metr
ics
.
T
he
blue
li
ne
(
S
E
.
L
P
V.
P
R
I
M
.
B
M
P
.
F
E
)
b
e
gins
a
t
a
ppr
oxim
a
tely
63
in
2002,
r
is
e
s
to
a
r
ound
73
in
2
010,
a
nd
then
de
c
li
ne
s
s
ha
r
ply
to
a
bout
58
by
20
16.
T
his
indi
c
a
tes
a
n
ini
ti
a
l
im
pr
ove
ment
f
oll
owe
d
by
a
s
igni
f
ica
nt
dr
op
.
C
onve
r
s
e
ly,
the
or
a
nge
li
ne
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
A
r
ti
fi
c
ial
int
e
ll
igenc
e
pr
e
dictive
mode
li
ng
for
e
duc
ati
onal
indi
c
ator
s
us
ing
data
…
(
Souk
aina
N
ai
)
3067
(
S
E
.
L
P
V
.
P
R
I
M
.
B
M
P
.
M
A)
s
tar
ts
a
t
a
r
ound
70
in
2002,
s
tea
dil
y
c
li
mbs
to
a
pe
a
k
o
f
a
bout
80
in
2
012,
a
nd
then
f
a
ll
s
to
r
ough
ly
70
in
2016.
T
h
is
s
hows
a
c
o
nti
nuous
incr
e
a
s
e
unti
l
2012
,
f
oll
owe
d
by
a
de
c
r
e
a
s
e
.
T
he
pe
a
k
be
twe
e
n
2010
a
nd
2012
c
or
r
e
s
ponds
with
th
e
im
pleme
ntation
of
pr
ojec
t
E
1
.
P
5
unde
r
the
2009
to
2012
e
mer
ge
nc
y
pr
ogr
a
m
,
whic
h
a
im
e
d
to
c
ombat
s
c
hool
r
e
pe
ti
ti
on
a
nd
dr
opout
r
a
tes
thr
ough
pe
r
s
ona
li
z
e
d
moni
tor
ing
a
nd
s
uppor
t
[
6]
T
he
gr
a
ph
e
f
f
e
c
ti
ve
ly
c
ontr
a
s
ts
the
tr
e
nds
of
both
indi
c
a
tor
s
,
highl
ight
ing
their
r
is
e
a
nd
s
ubs
e
que
nt
de
c
li
ne
ove
r
ti
me.
T
he
g
r
a
ph
in
F
igur
e
2
tr
a
c
ks
two
int
e
r
na
ti
ona
l
a
s
s
e
s
s
ment
indi
c
a
tor
s
ove
r
ti
me:
pr
og
r
e
s
s
in
int
e
r
na
ti
ona
l
r
e
a
ding
li
ter
a
c
y
s
tudy
(
L
O.
P
I
R
L
S
.
R
E
A.
I
NT
)
a
nd
tr
e
nds
in
int
e
r
na
ti
ona
l
mathe
matics
a
nd
s
c
ienc
e
s
tudy
f
or
8th
g
r
a
de
r
s
in
mathe
matics
(
L
O.
T
I
M
S
S
.
M
AT
8
.
I
NT
)
.
T
he
blue
li
ne
f
or
L
O.
P
I
R
L
S
.
R
E
A.
I
NT
s
hows
a
s
igni
f
ica
nt
de
c
li
ne
f
r
om
2001,
r
e
a
c
hing
it
s
lowe
s
t
a
r
ound
2011,
then
r
is
ing
s
ha
r
ply
unti
l
2016
be
f
or
e
s
li
ghtl
y
de
c
li
ning
by
20
19.
I
n
c
ontr
a
s
t,
the
or
a
nge
li
ne
f
or
L
O.
T
I
M
S
S
.
M
AT
8.
I
N
T
dis
plays
a
s
tea
dy
incr
e
a
s
e
f
r
om
2001
to
2006
,
a
mi
nor
de
c
li
ne
unti
l
2011
,
a
nd
then
a
s
tabili
z
a
ti
on
with
a
s
li
ght
de
c
r
e
a
s
e
by
2016.
T
he
gr
a
ph
r
e
f
lec
ts
s
tudent
pe
r
f
or
manc
e
tr
e
nds
in
thes
e
a
s
s
e
s
s
ments
,
r
e
ve
a
li
ng
the
lowe
s
t
pe
r
f
or
manc
e
be
twe
e
n
2010
a
nd
2012.
T
h
is
dr
op
is
li
nke
d
to
the
im
pleme
ntation
o
f
a
n
a
lt
e
r
na
ti
ve
a
c
a
de
mi
c
s
uppor
t
pr
ogr
a
m
du
r
ing
that
pe
r
iod
.
F
igur
e
2.
Da
ta
vis
ua
li
z
a
ti
on
f
o
r
int
e
r
na
ti
ona
l
a
s
s
e
s
s
ments
5
.
3
.
M
ac
h
in
e
lear
n
in
g
r
e
gr
e
s
s
ion
algorit
h
m
s
T
o
make
p
r
e
diction
a
bout
the
number
of
s
tudents
who
ne
e
d
be
ne
f
it
ing
f
r
om
the
s
uppor
t
pr
ogr
a
ms
,
mac
hine
lea
r
ning
a
lgor
it
hms
a
r
e
e
s
s
e
nti
a
l
be
c
a
us
e
they
c
a
n
p
r
oc
e
s
s
lar
ge
a
nd
c
ompl
e
x
da
tas
e
ts
to
identif
y
pa
tt
e
r
ns
a
nd
t
r
e
nds
that
tr
a
dit
ional
methods
mi
ght
mi
s
s
.
I
n
f
ields
li
ke
e
duc
a
ti
on,
mac
hine
lea
r
ning
he
lps
pr
e
dict
s
tudent
pe
r
f
or
manc
e
,
d
r
opout
r
a
tes
,
or
the
s
uc
c
e
s
s
of
int
e
r
ve
nti
ons
,
e
na
bli
ng
mo
r
e
e
f
f
e
c
ti
ve
de
c
is
ion
-
making.
M
a
c
hine
lea
r
ning
models
c
onti
n
uous
ly
lea
r
n
f
r
om
da
ta,
a
da
pti
ng
to
c
ha
nge
s
ove
r
ti
me
a
n
d
im
pr
oving
a
c
c
ur
a
c
y.
T
he
ir
a
bil
it
y
to
ha
ndle
nonl
i
ne
a
r
r
e
lations
hips
a
nd
diver
s
e
da
ta
make
s
them
c
r
uc
ial
f
o
r
f
or
e
c
a
s
ti
ng,
a
ll
owing
f
o
r
be
tt
e
r
r
e
s
our
c
e
a
ll
oc
a
ti
on
a
nd
pr
oa
c
ti
ve
s
tr
a
tegie
s
,
ult
im
a
tely
dr
ivi
ng
i
mpr
ove
d
outcome
s
a
nd
e
f
f
icie
nc
y
in
de
c
is
ion
-
mak
ing
pr
oc
e
s
s
e
s
.
As
c
on
s
e
que
nc
e
,
we
ne
e
d
to
identif
y
the
opti
mal
mac
hine
lea
r
ning
model
f
o
r
a
na
lyzing
e
duc
a
ti
ona
l
indi
c
a
tor
s
.
I
n
thi
s
c
ontext
we
c
it
e
the
li
s
t
of
a
lgo
r
it
hms
that
we
will
us
e
in
th
is
pa
pe
r
,
a
s
pr
e
s
e
nted
in
T
a
ble
2.
T
a
ble
2.
M
a
c
hine
lea
r
ning
r
e
gr
e
s
s
ion
a
lgor
it
h
ms
T
it
le
D
e
s
c
r
ip
ti
on
L
in
e
a
r
r
e
gr
e
s
s
io
n
A
s
im
pl
e
r
e
gr
e
s
s
io
n
mode
l
th
a
t
a
s
s
um
e
s
a
li
ne
a
r
r
e
la
ti
ons
hi
p
be
twe
e
n
th
e
in
put
f
e
a
tu
r
e
s
a
nd t
he
t
a
r
ge
t
va
r
ia
bl
e
.
P
ol
ynomi
a
l
r
e
gr
e
s
s
io
n
A
va
r
ia
ti
on
of
li
ne
a
r
r
e
gr
e
s
s
io
n
th
a
t
mode
ls
a
nonl
in
e
a
r
r
e
l
a
ti
ons
hi
p
by
in
c
lu
di
ng
pol
ynomi
a
l
f
e
a
tu
r
e
s
.
R
a
ndom f
or
e
s
t
A
n
e
ns
e
mbl
e
me
th
od
th
a
t
c
ons
tr
uc
ts
mul
ti
pl
e
de
c
is
io
n
tr
e
e
s
a
n
d
me
r
ge
s
th
e
m
to
ge
th
e
r
to
ge
t
mor
e
a
c
c
ur
a
te
a
nd s
ta
bl
e
pr
e
di
c
ti
ons
.
S
uppor
t
ve
c
to
r
r
e
gr
e
s
s
io
n
A
r
e
gr
e
s
s
io
n
a
lg
or
it
hm
th
a
t
tr
ie
s
to
f
it
th
e
be
s
t
hype
r
pl
a
ne
w
it
h
in
a
ma
r
gi
n
of
to
le
r
a
nc
e
w
hi
le
mi
ni
mi
z
in
g t
he
e
r
r
or
out
s
id
e
t
hi
s
ma
r
gi
n.
D
e
c
is
io
n t
r
e
e
A
tr
e
e
-
ba
s
e
d
r
e
gr
e
s
s
io
n
a
lg
or
it
hm
th
a
t
s
pl
it
s
th
e
d
a
ta
in
to
br
a
nc
he
s
to
pr
e
di
c
t
th
e
ta
r
ge
t
va
r
ia
bl
e
by f
ol
lo
w
in
g s
im
pl
e
de
c
is
io
n r
ul
e
s
.
M
ul
ti
-
la
ye
r
pe
r
c
e
pt
r
on (
M
L
P
)
r
e
gr
e
s
s
or
A
ty
pe
of
ne
ur
a
l
ne
twor
k
w
it
h
mul
ti
pl
e
la
ye
r
s
a
nd
node
s
th
a
t
a
ppr
oxi
ma
te
s
c
ompl
e
x
f
unc
ti
ons
f
or
r
e
gr
e
s
s
io
n t
a
s
ks
t
hr
ough ba
c
kpr
opa
ga
ti
on.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
306
3
-
3073
3068
6.
RE
S
UL
T
S
AN
D
DI
S
CU
S
S
I
ON
6
.
1.
Dat
a
vis
u
ali
z
at
ion
an
d
p
r
e
d
i
c
t
ion
6.
1
.
1.
B
e
n
e
f
it
i
n
g
f
r
o
m
s
u
p
p
or
t
p
r
ogr
am
s
in
d
icat
or
s
Us
ing
the
indi
c
a
to
r
s
on
pe
r
c
e
ntage
o
f
p
r
im
a
r
y
s
tu
de
nts
be
ne
f
it
ing
f
r
om
a
c
a
de
mi
c
s
uppor
t
p
r
ogr
a
ms
(
B
e
ne
f
it
ing
f
r
om
s
uppor
t
pr
ogr
a
ms
)
,
we
a
r
e
going
to
a
pply
the
mac
hine
lea
r
ning
a
lgor
it
hms
:
li
ne
a
r
r
e
gr
e
s
s
ion
a
s
s
hown
in
F
igur
e
3
a
nd
r
a
ndom
f
o
r
e
s
t
a
s
s
hown
in
F
igu
r
e
4
.
T
his
is
done
in
or
de
r
to
make
a
c
o
mpar
is
on
be
twe
e
n
them.
T
he
goa
l
is
to
de
ter
mi
ne
the
be
s
t
a
l
gor
it
hm
to
make
the
p
r
e
diction
in
thi
s
f
r
a
mew
or
k.
F
igur
e
3.
P
r
e
diction
us
ing
li
ne
a
r
r
e
gr
e
s
s
ion
F
igur
e
4.
P
r
e
diction
us
ing
r
a
ndom
f
o
r
e
s
t
r
e
gr
e
s
s
ion
T
he
g
r
a
ph
in
F
igur
e
3
is
the
r
e
s
ult
o
f
a
li
ne
a
r
r
e
gr
e
s
s
ion
a
na
lys
is
,
vis
ua
li
z
e
s
t
r
e
nds
f
o
r
two
indi
c
a
tor
s
:
f
e
male
pr
im
a
r
y
s
c
hool
c
ompl
e
ti
on
r
a
te
(
S
E
.
L
P
V.
P
R
I
M
.
B
M
P
.
F
E
)
a
nd
male
pr
im
a
r
y
s
c
hool
c
ompl
e
ti
on
r
a
te
(
S
E
.
L
P
V.
P
R
I
M
.
B
M
P
.
M
A)
r
e
late
d
to
s
uppor
t
pr
og
r
a
ms
.
His
tor
ica
l
da
ta
(
s
oli
d
li
ne
s
)
i
s
s
hown
a
longs
ide
model
f
it
s
(
da
s
he
d
li
ne
s
)
,
a
nd
tes
t
da
ta
point
s
a
r
e
include
d.
P
r
e
dictions
f
or
f
utur
e
va
lue
s
e
xtend
f
r
om
2017
to
2
020,
r
e
pr
e
s
e
nted
by
r
e
d
c
r
os
s
e
s
.
T
h
e
gr
a
ph
indi
c
a
tes
de
c
li
ning
tr
e
nds
f
or
both
male
a
nd
f
e
male
c
ompl
e
ti
on
r
a
tes
a
f
ter
2015
,
a
s
s
hown
by
the
down
wa
r
d
-
s
lopi
ng
da
s
he
d
a
nd
s
oli
d
li
ne
s
,
s
ugge
s
ti
ng
a
potential
dr
op
in
e
duc
a
ti
ona
l
outcome
s
without
f
u
r
ther
int
e
r
ve
nti
on.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
A
r
ti
fi
c
ial
int
e
ll
igenc
e
pr
e
dictive
mode
li
ng
for
e
duc
ati
onal
indi
c
ator
s
us
ing
data
…
(
Souk
aina
N
ai
)
3069
T
he
gr
a
ph
in
F
igu
r
e
4
il
lus
tr
a
tes
the
r
e
s
ult
s
of
a
r
a
ndom
f
o
r
e
s
t
r
e
gr
e
s
s
ion
model
with
a
t
r
a
in
-
tes
t
s
pli
t
a
nd
pr
e
dictions
f
or
two
indi
c
a
tor
s
:
f
e
male
pr
i
mar
y
s
c
hool
c
ompl
e
ti
on
r
a
te
(
S
E
.
L
P
V.
P
R
I
M
.
B
M
P
.
F
E
)
a
nd
male
pr
im
a
r
y
s
c
hool
c
ompl
e
ti
on
r
a
te
(
S
E
.
L
P
V.
P
R
I
M
.
B
M
P
.
M
A)
.
His
tor
ica
l
da
ta
is
plot
ted
f
r
o
m
a
r
ound
2002
to
2015,
with
tr
a
in
a
nd
tes
t
pr
e
dictions
ove
r
l
a
ye
d.
F
utur
e
pr
e
dictions
f
or
both
indi
c
a
tor
s
e
xtend
be
yond
2017
to
2020.
T
he
o
r
a
nge
a
nd
blue
li
ne
s
r
e
pr
e
s
e
nt
the
his
tor
ica
l
a
nd
pr
e
dicte
d
tr
e
nds
f
or
f
e
male
a
nd
male
c
ompl
e
ti
on
r
a
tes
,
r
e
s
pe
c
ti
ve
ly.
T
he
r
e
d
c
r
os
s
e
s
r
e
p
r
e
s
e
nt
f
utur
e
pr
e
dictions
,
indi
c
a
ti
ng
a
s
tea
dy
o
r
de
c
r
e
a
s
ing
tr
e
nd
a
f
ter
2017
f
or
both
indi
c
a
tor
s
.
6.
1
.
2.
I
n
t
e
r
n
at
io
n
al
as
s
e
s
s
m
e
n
t
s
(
s
u
p
p
or
t
)
i
n
d
ic
at
or
s
Us
ing
the
pe
r
f
o
r
manc
e
indi
c
a
tor
s
f
r
om
int
e
r
na
ti
on
a
l
a
s
s
e
s
s
ments
s
uc
h
a
s
T
I
M
S
S
a
nd
P
I
R
L
S
,
whic
h
indi
r
e
c
tl
y
r
e
f
lec
t
a
c
a
de
mi
c
s
uppor
t,
we
a
ppli
e
d
mac
hine
lea
r
ning
a
lgor
it
hms
.
S
pe
c
if
ica
ll
y
,
we
us
e
d
li
ne
a
r
r
e
gr
e
s
s
ion
a
s
s
hown
in
F
igur
e
5
a
nd
r
a
ndom
f
or
e
s
t
a
s
s
hown
in
F
igur
e
6
.
T
his
wa
s
done
to
c
om
pa
r
e
the
e
f
f
e
c
ti
ve
ne
s
s
of
both
a
lgor
it
hms
in
pr
e
dicting
o
utcome
s
withi
n
thi
s
f
r
a
mew
or
k
a
nd
to
identi
f
y
the
mos
t
s
uit
a
ble
model.
F
igur
e
5.
P
r
e
diction
us
ing
li
ne
a
r
r
e
gr
e
s
s
ion
F
igur
e
6.
P
r
e
diction
us
ing
r
a
ndom
f
o
r
e
s
t
r
e
gr
e
s
s
ion
T
he
gr
a
ph
in
F
igur
e
5
de
r
ived
f
r
om
a
li
ne
a
r
r
e
gr
e
s
s
ion
a
na
lys
is
,
f
o
r
e
c
a
s
ts
tr
e
nds
f
or
two
int
e
r
na
ti
ona
l
a
s
s
e
s
s
ment
s
uppor
t
indi
c
a
tor
s
:
L
O.
P
I
R
L
S
.
R
E
A.
I
NT
a
nd
L
O.
T
I
M
S
S
.
M
AT
8
.
I
NT
.
T
he
h
is
tor
ica
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
306
3
-
3073
3070
da
ta
(
s
oli
d
l
ines
)
a
nd
tes
t
da
ta
poin
ts
r
e
ve
a
l
th
e
pe
r
f
or
manc
e
tr
e
nds
ove
r
ti
me,
while
the
da
s
he
d
li
ne
s
r
e
pr
e
s
e
nt
the
model
f
it
s
f
or
thes
e
indi
c
a
tor
s
.
T
he
gr
a
ph
s
pa
ns
f
r
om
a
r
ound
2000
to
2015
f
or
obs
e
r
ve
d
da
ta,
with
f
utu
r
e
p
r
e
dictions
e
xtending
to
2020
,
r
e
pr
e
s
e
nted
by
r
e
d
c
r
os
s
e
s
.
T
he
P
I
R
L
S
s
c
or
e
s
e
xhi
bit
e
a
r
ly
f
luctua
ti
ons
,
but
the
ge
ne
r
a
l
tr
e
nd
f
or
both
ind
i
c
a
tor
s
is
upwa
r
d,
s
ugge
s
ti
ng
a
n
ove
r
a
ll
i
mpr
ove
ment
in
r
e
a
ding
a
nd
mathe
matics
pe
r
f
or
manc
e
.
T
he
f
utur
e
pr
e
dictions
a
ls
o
indi
c
a
te
a
s
tea
dy
r
is
e
in
both
a
s
s
e
s
s
ment
s
c
or
e
s
,
s
howing
potential
pr
ogr
e
s
s
in
thes
e
int
e
r
na
t
ional
e
va
lu
a
ti
ons
.
T
he
gr
a
ph
in
F
igur
e
6
il
lus
tr
a
tes
the
r
e
s
ult
s
of
a
r
a
ndom
f
or
e
s
t
r
e
gr
e
s
s
ion
model
with
a
t
r
a
in
-
tes
t
s
pli
t
a
nd
pr
e
dictions
f
or
two
indi
c
a
tor
s
:
L
O.
P
I
R
L
S
.
R
E
A.
I
NT
a
nd
L
O.
T
I
M
S
S
.
M
AT
8
.
I
NT
.
His
tor
ica
l
da
ta
is
de
picte
d
f
r
om
a
ppr
oxim
a
tely
2002
to
2015
,
wi
th
tr
a
in
a
nd
tes
t
pr
e
dictions
ove
r
laid
.
F
utur
e
p
r
e
dictions
f
or
both
indi
c
a
tor
s
e
xtend
f
r
om
2017
to
2020
.
T
he
or
a
nge
a
nd
blue
li
ne
s
c
or
r
e
s
pond
to
the
his
tor
ica
l
a
nd
pr
e
dicte
d
tr
e
nds
f
or
f
e
male
a
nd
male
c
ompl
e
t
ion
r
a
tes
,
r
e
s
pe
c
ti
ve
ly.
T
he
r
e
d
c
r
os
s
e
s
de
not
e
f
ut
ur
e
pr
e
dictions
,
s
ugge
s
ti
ng
e
it
he
r
a
s
table
o
r
de
c
li
ning
tr
e
nd
be
yond
2017
f
or
both
ind
ica
tor
s
.
R
e
s
ult
:
thr
ough
thi
s
a
na
lys
is
,
we
c
onc
lude
that
it
is
ins
uf
f
icie
nt
to
de
ter
mi
ne
the
be
s
t
mac
hine
lea
r
ning
model
ba
s
e
d
on
jus
t
two
models
.
I
n
c
ons
e
que
nc
e
,
to
id
e
nti
f
y
the
mos
t
e
f
f
e
c
ti
ve
model
f
or
pr
e
dicting
e
duc
a
ti
ona
l
indi
c
a
tor
s
,
we
will
a
pply
va
r
ious
mac
hine
lea
r
ning
a
lgor
it
hms
to
the
e
nti
r
e
da
tas
e
t
obtaine
d
f
r
om
the
W
or
ld
B
a
nk
we
bs
it
e
.
T
he
n,
we
will
c
ompar
e
thes
e
models
ba
s
e
d
on
a
c
c
ur
a
c
y
mea
s
ur
e
s
to
de
ter
mi
ne
whic
h
one
pe
r
f
or
ms
the
be
s
t.
6
.
2.
M
ac
h
in
e
lear
n
in
g
c
om
p
ar
is
on
T
o
identif
y
the
be
s
t
mac
hine
lea
r
ning
model
f
or
pr
e
dicting
e
duc
a
ti
ona
l
indi
c
a
tor
s
,
we
will
a
pply
mul
ti
ple
a
lgor
it
hms
to
the
f
ull
da
tas
e
t
obtaine
d.
T
his
a
ppr
oa
c
h
wil
l
a
ll
ow
us
to
e
va
luate
e
a
c
h
model's
pe
r
f
or
manc
e
a
nd
de
ter
mi
ne
whic
h
is
mos
t
e
f
f
e
c
ti
ve
f
or
making
a
c
c
ur
a
te
pr
e
dictions
ba
s
e
d
on
the
da
ta.
Our
pr
ogr
a
m,
wr
it
ten
in
P
ython
langua
ge
a
nd
us
ing
the
s
a
me
mac
hine,
pr
oduc
e
s
the
T
a
ble
3
.
Ana
lys
is
:
‒
M
e
a
n
a
bs
olut
e
e
r
r
or
(
M
AE
)
:
mea
s
ur
e
s
the
a
v
e
r
a
g
e
magnitude
o
f
e
r
r
o
r
s
in
pr
e
dictions
with
out
c
ons
ider
ing
their
di
r
e
c
ti
on.
L
owe
r
va
lues
indi
c
a
te
be
tt
e
r
pe
r
f
o
r
manc
e
.
B
e
s
t:
r
a
ndom
f
or
e
s
t
(
3
.
14)
W
or
s
t:
M
L
P
r
e
gr
e
s
s
or
(
21.
10
)
‒
M
e
a
n
s
qua
r
e
d
e
r
r
or
(
M
S
E
)
:
mea
s
ur
e
s
the
a
ve
r
a
ge
of
the
s
qua
r
e
s
of
the
e
r
r
o
r
s
,
givi
ng
mor
e
we
ig
ht
to
lar
ge
r
e
r
r
o
r
s
.
L
owe
r
va
lues
indi
c
a
te
be
tt
e
r
pe
r
f
or
m
a
nc
e
.
B
e
s
t:
r
a
ndom
f
or
e
s
t
(
87
.
62)
W
or
s
t:
s
uppor
t
ve
c
tor
r
e
gr
e
s
s
ion
(
9159.
70
)
‒
R
oot
mea
n
s
qua
r
e
d
e
r
r
or
(
R
M
S
E
)
:
s
qua
r
e
r
oot
of
M
S
E
,
pr
ovidi
ng
a
n
e
r
r
o
r
metr
ic
in
the
s
a
me
unit
a
s
the
tar
g
e
t
va
r
iable
.
L
owe
r
va
lues
indi
c
a
te
be
tt
e
r
pe
r
f
or
manc
e
.
B
e
s
t:
r
a
ndom
f
or
e
s
t
(
3
.
91)
W
or
s
t:
M
L
P
r
e
gr
e
s
s
or
(
23.
94
)
‒
R
²
s
c
or
e
:
r
e
pr
e
s
e
nts
the
pr
opor
ti
on
of
va
r
ianc
e
in
the
de
pe
nde
nt
va
r
iable
that
is
pr
e
dicta
ble
f
r
om
the
indepe
nde
nt
va
r
iable
.
0
.
92
indi
c
a
tes
that
the
model
r
a
ndom
f
or
e
s
t
ha
s
c
onve
r
ge
d
s
uc
c
e
s
s
f
ull
y.
R
e
s
ult
:
r
a
ndom
f
o
r
e
s
t
is
the
be
s
t
-
pe
r
f
or
mi
ng
mode
l
a
c
r
os
s
a
ll
pr
ov
ided
metr
ics
,
with
the
lowe
s
t
M
A
E
,
M
S
E
,
a
nd
R
M
S
E
.
T
his
s
ugge
s
ts
that
it
ha
s
the
mos
t
a
c
c
ur
a
te
a
nd
c
ons
is
tent
pr
e
dictions
a
mong
the
models
t
e
s
ted.
T
a
ble
3.
C
ompar
is
on
of
mac
hine
lea
r
ning
f
or
pr
e
di
c
ti
ng
A
lg
or
it
hm
M
A
E
M
S
E
R
M
S
E
R²
L
in
e
a
r
r
e
gr
e
s
s
io
n
5.14
539.95
5.88
0.75
P
ol
ynomi
a
l
r
e
gr
e
s
s
io
n
4.43
140.25
5.04
0.85
R
a
ndom
f
or
e
s
t
3.14
87.62
3.91
0.92
S
uppor
t
ve
c
to
r
r
e
gr
e
s
s
io
n
14.12
9159.70
17.45
0.10
D
e
c
is
io
n
tr
e
e
3.43
136.60
4.12
0.84
M
L
P
r
e
gr
e
s
s
or
21.10
8479.91
23.94
0.15
7.
CONC
L
USI
ON
I
n
s
umm
a
r
y,
our
pa
pe
r
e
xplo
r
e
s
how
mac
hine
lea
r
ning
models
c
a
n
i
mpr
ove
p
r
e
diction
a
c
c
ur
a
c
y
in
M
or
oc
c
o's
e
duc
a
ti
on
s
e
c
tor
.
B
y
c
ompar
ing
va
r
ious
a
lgor
it
hms
a
c
r
os
s
dif
f
e
r
e
nt
e
duc
a
ti
ona
l
indi
c
a
tor
s
,
we
a
im
to
identif
y
the
be
s
t
model
f
or
guidi
ng
da
ta
-
dr
iven
de
c
is
ions
.
T
his
r
e
s
e
a
r
c
h
will
he
lp
us
de
ve
lop
be
tt
e
r
,
f
a
ir
e
r
,
a
nd
mor
e
a
c
c
ur
a
te
e
duc
a
ti
ona
l
poli
c
ies
to
a
ddr
e
s
s
ke
y
c
ha
ll
e
nge
s
in
M
or
oc
c
o's
e
duc
a
ti
on
s
ys
tem.
I
n
our
f
utur
e
wor
k
,
we
will
f
oc
us
on
incor
po
r
a
ti
ng
lar
ge
r
,
r
e
a
l
-
ti
me
da
tas
e
ts
to
e
nha
nc
e
the
pr
e
c
is
ion
of
our
models
.
Additi
ona
ll
y,
we
p
lan
to
de
ve
lop
pe
r
s
ona
li
z
e
d
e
du
c
a
ti
on
s
ys
tem
s
ugge
s
ti
ons
to
tailo
r
r
e
c
omm
e
nda
ti
o
ns
ba
s
e
d
on
indi
vidual
ne
e
ds
a
nd
lea
r
ning
s
tyl
e
s
.
T
his
a
ppr
oa
c
h
a
im
s
to
f
ur
the
r
im
p
r
ove
the
e
f
f
e
c
ti
v
e
ne
s
s
of
e
duc
a
ti
on
a
l
int
e
r
ve
nti
ons
a
nd
poli
c
y
de
c
is
ions
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
A
r
ti
fi
c
ial
int
e
ll
igenc
e
pr
e
dictive
mode
li
ng
for
e
duc
ati
onal
indi
c
ator
s
us
ing
data
…
(
Souk
aina
N
ai
)
3071
F
UN
DI
NG
I
NF
ORM
AT
I
ON
Author
s
s
tate
no
f
unding
invol
ve
d.
AU
T
HO
R
CONT
RI
B
U
T
I
ONS
S
T
AT
E
M
E
N
T
T
his
jour
na
l
us
e
s
the
C
ontr
ibut
o
r
R
oles
T
a
xo
nomy
(
C
R
e
diT
)
to
r
e
c
ognize
indi
vidual
a
uthor
c
ontr
ibut
ions
,
r
e
duc
e
a
utho
r
s
hip
dis
pute
s
,
a
nd
f
a
c
il
it
a
te
c
oll
a
bor
a
ti
on.
Nam
e
of
Au
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
S
ouka
ina
Na
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
B
a
ha
a
E
ddine
E
lbagha
z
a
oui
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
R
if
a
i
Ama
l
✓
✓
✓
✓
✓
✓
✓
✓
Abde
lalim
S
a
diq
✓
✓
✓
✓
✓
C
:
C
onc
e
pt
ua
li
z
a
ti
on
M
:
M
e
th
odol
ogy
So
:
So
f
twa
r
e
Va
:
Va
li
da
ti
on
Fo
:
Fo
r
ma
l
a
na
ly
s
is
I
:
I
nve
s
ti
ga
ti
on
R
:
R
e
s
our
c
e
s
D
:
D
a
ta
C
ur
a
ti
on
O
:
W
r
it
in
g
-
O
r
ig
in
a
l
D
r
a
f
t
E
:
W
r
it
in
g
-
R
e
vi
e
w
&
E
di
ti
ng
Vi
:
Vi
s
ua
li
z
a
ti
on
Su
:
Su
pe
r
vi
s
io
n
P
:
P
r
oj
e
c
t
a
dmi
ni
s
tr
a
ti
on
Fu
:
Fu
ndi
ng a
c
qui
s
it
io
n
CONF
L
I
CT
OF
I
NT
E
RE
S
T
S
T
AT
E
M
E
N
T
Author
s
s
tate
no
c
onf
li
c
t
of
int
e
r
e
s
t.
DA
T
A
AV
AI
L
A
B
I
L
I
T
Y
T
he
da
ta
that
s
uppor
t
the
f
indi
ngs
of
thi
s
s
tud
y
a
r
e
ope
nly
a
va
il
a
ble
in
[
T
he
W
or
ld
B
a
nk]
a
t
htt
p:/
/doi
.
or
g
/10.
26599/B
DM
A.
2019.
9020007
,
r
e
f
e
r
e
nc
e
number
[
18
]
.
RE
F
E
RE
NC
E
S
[
1]
R
.
E
jj
a
mi
,
“
R
e
vol
ut
io
ni
z
in
g
M
or
oc
c
a
n
e
duc
a
ti
on
w
it
h
A
I
:
a
pa
th
to
c
us
to
mi
z
e
d
le
a
r
ni
ng,”
I
nt
e
r
nat
io
nal
J
our
nal
F
or
M
ul
ti
di
s
c
ip
li
nar
y
R
e
s
e
ar
c
h
, vol
. 6, no. 3, pp. 1
–
32, 2024, doi:
1
0.36948/i
jf
mr
.2024.v06i03.19462
.
[
2]
A
. Z
in
e
a
nd A
. K
a
a
oua
c
hi
, “
K
e
y de
te
r
mi
na
nt
s
of
l
e
a
r
ni
ng a
na
l
yt
ic
s
a
dopt
io
n i
n M
or
oc
c
a
n unive
r
s
it
ie
s
,”
J
our
nal
of
E
c
ohumanis
m
,
vol
. 3, no. 5, pp. 1
–
12, 2024, doi:
10.62754/j
oe
.v3i
5.3634.
[
3]
A
.
Q
a
z
da
r
,
O
.
H
a
s
id
i,
S
.
Q
a
s
s
im
i,
a
nd
E
.
H
.
A
bde
lwa
h
e
d,
“
N
e
w
ly
pr
opos
e
d
s
tu
de
nt
p
e
r
f
or
ma
nc
e
in
di
c
a
to
r
s
b
a
s
e
d
on
le
a
r
ni
ng
a
na
ly
ti
c
s
f
or
c
ont
in
uous
moni
to
r
in
g
in
le
a
r
ni
ng
ma
na
ge
m
e
nt
s
ys
te
ms
,
”
I
nt
e
r
nat
io
nal
J
ou
r
nal
of
O
nl
in
e
and
B
io
m
e
di
c
al
E
ngi
ne
e
r
in
g
, vol
. 19, no. 11, pp. 19
–
30, 2023, doi:
10.3991/i
jo
e
.v19i11.39471.
[
4]
M
.
B
ougr
oum
a
nd
A
.
I
bour
k,
“
A
c
c
e
s
s
a
nd
e
qu
it
y
in
f
in
a
nc
in
g
hi
ghe
r
e
duc
a
ti
on:
th
e
c
a
s
e
of
M
or
oc
c
o,”
P
r
o
s
pe
c
t
s
,
vol
.
41,
no
.
1,
pp. 115
–
134, 2011, doi:
10.1007/s
11125
-
011
-
9184
-
8.
[
5]
J
. B
r
a
s
s
e
,
M
. F
ör
s
te
r
, P
. H
ühn, J
. K
li
e
r
, M
. K
li
e
r
,
a
nd L
.
M
oe
s
tu
e
, “
P
r
e
pa
r
in
g f
or
t
he
f
ut
ur
e
of
w
or
k:
a
nove
l
da
ta
-
d
r
iv
e
n a
ppr
o
a
c
h
f
or
th
e
id
e
nt
if
ic
a
ti
on
o
f
f
ut
ur
e
s
ki
ll
s
,”
J
our
nal
of
B
us
in
e
s
s
E
c
onomic
s
,
vol
.
94,
no.
3,
pp.
467
–
500,
2024,
doi
:
10.1007/s
11573
-
023
-
01169
-
1.
[
6]
N
.
A
.
D
a
hr
i
e
t
al
.
,
“
I
nve
s
ti
ga
ti
ng
A
I
-
ba
s
e
d
a
c
a
de
mi
c
s
uppor
t
a
c
c
e
pt
a
nc
e
a
nd
it
s
im
pa
c
t
on
s
tu
de
nt
s
’
p
e
r
f
or
ma
nc
e
in
M
a
l
a
y
s
ia
n
a
nd
P
a
ki
s
ta
ni
hi
ghe
r
e
duc
a
ti
on
in
s
ti
tu
ti
ons
,”
E
duc
at
io
n
and
I
n
fo
r
m
at
io
n
T
e
c
hnol
ogi
e
s
,
vol
.
29,
no.
14,
pp.
18695
–
18744,
20
24,
doi
:
10.1007/s
10639
-
024
-
12599
-
x.
[
7]
A
. I
bour
k a
nd S
.
R
a
oui
, “
I
nc
lu
s
iv
e
e
duc
a
ti
on a
nd s
c
hool
dr
opout
of
s
pe
c
ia
l
ne
e
ds
s
tu
de
nt
s
i
n M
or
oc
c
o:
a
s
pa
ti
a
l
a
na
ly
s
is
,”
R
e
v
ie
w
of
E
duc
at
io
n
, vol
. 12, no. 1, 2024, doi
:
10.1002/r
e
v3.3453.
[
8]
K
.
S
a
oudi
,
R
.
C
hr
oqui
,
a
nd
C
.
O
ka
r
,
“
S
tu
de
nt
a
c
hi
e
ve
me
nt
in
M
or
oc
c
a
n
e
duc
a
ti
ona
l
r
e
f
or
ms
:
a
s
ig
ni
f
ic
a
nt
ga
p
be
tw
e
e
n
a
s
p
ir
e
d
o
ut
c
ome
s
a
nd c
ur
r
e
nt
pr
a
c
ti
c
e
s
,”
I
nt
e
r
c
hang
e
, vol
. 51, no. 2, pp. 117
–
136, 2020, doi:
10.1007/s
10780
-
019
-
09380
-
2.
[
9]
N
.
M
or
c
hi
d,
“
I
nve
s
ti
ga
ti
ng
qua
li
ty
e
duc
a
ti
on
in
M
or
oc
c
a
n
e
du
c
a
ti
ona
l
r
e
f
or
ms
f
r
om
1999
to
2019,”
I
O
SR
J
our
nal
of
R
e
s
e
ar
c
h
&
M
e
th
od i
n E
duc
a
ti
on (
I
O
S
R
-
J
R
M
E
)
, vol
. 10, no. 1, pp. 54
–
61,
2020, doi:
10.9790/7388
-
1001015461.
[
10]
O
.
E
lk
ha
lf
i,
R
.
C
ha
a
bi
ta
,
K
.
Z
a
hr
a
oui
,
a
nd
H
.
E
l
A
la
oui
,
“
P
ubl
ic
s
pe
ndi
ng
on
huma
n
c
a
pi
ta
l
a
nd
e
c
onomi
c
gr
ow
th
in
M
or
oc
c
o,”
I
nt
e
r
nat
io
nal
J
our
nal
of
E
c
onomic
s
and F
in
anc
i
al
I
s
s
ue
s
, vol
. 13, no. 4, pp. 102
–
110, 2023, doi:
10.32479/i
je
f
i.
14374.
[
11]
I
.
T
a
mm
ouc
h,
A
.
E
lo
ua
f
i,
S
.
E
dda
r
oui
c
h,
a
nd
R
.
T
oua
hni
,
“
I
de
nt
if
yi
ng
lo
w
-
pe
r
f
or
mi
ng
r
e
gi
ons
in
M
or
oc
c
a
n
e
duc
a
ti
on:
a
de
e
p
le
a
r
ni
ng
a
ppr
oa
c
h
us
in
g
th
e
P
I
S
A
da
ta
s
e
t,
”
I
nt
e
r
nat
io
nal
J
our
nal
of
A
dv
anc
e
d
and
A
ppl
ie
d
Sc
ie
nc
e
s
,
vol
.
10,
n
o.
7,
pp. 138
–
144, 2023, doi:
10.21833/i
ja
a
s
.2023.07.015.
[
12]
H
.
B
oudi
ne
,
H
.
S
a
ji
d,
M
.
B
e
nt
a
l
e
b,
M
.
T
a
ye
bi
,
a
nd
D
.
E
l
K
a
r
f
a
,
“
M
-
le
a
r
ni
ng
a
nd
a
ut
onomous
e
duc
a
ti
on:
th
e
im
p
a
c
t
o
f
th
e
M
or
oc
c
a
n
di
gi
ta
l
c
la
s
s
r
oom
pr
oj
e
c
t
on
s
c
ie
nc
e
s
ubj
e
c
t’
s
le
a
r
ni
ng,”
I
nt
e
r
nat
io
nal
J
our
nal
of
C
iv
il
iz
at
io
ns
St
udi
e
s
&
T
ol
e
r
anc
e
Sc
ie
nc
e
s
, v
ol
. 1, no. 1, pp. 44
–
54, 2024, doi:
10.54878/1qcw
95
64.
[
13]
R
. B
oukbe
c
h a
nd M
. L
io
ua
e
ddi
ne
, “
E
c
onomi
c
a
na
ly
s
is
of
e
duc
a
ti
ona
l
e
f
f
ic
ie
nc
y i
n M
or
oc
c
o:
a
n
a
ppl
ic
a
ti
on
of
t
he
D
E
A
me
th
od,”
I
nt
e
r
nat
io
nal
J
our
nal
of
St
r
at
e
gi
c
M
anage
m
e
nt
and
E
c
onomic
St
udi
e
s
(
I
J
SM
E
S)
,
vol
.
2,
no.
3,
pp.
1534
–
1547,
2023,
doi
:
10.5281/z
e
nodo.8337959.
[
14]
H
.
B
e
r
ba
r
,
S
.
L
ot
f
i,
a
nd
M
.
T
a
lb
i,
“
V
a
li
da
ti
on
a
nd
de
ve
lo
pm
e
nt
of
a
M
or
oc
c
a
n
s
c
hool
qua
li
ty
e
va
lu
a
ti
on
s
ys
te
m,”
E
duc
at
io
n
R
e
s
e
ar
c
h I
nt
e
r
nat
io
nal
, vol
. 2021, no. 1, 2021, doi
:
10.115
5/
20
21/
1829259.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
306
3
-
3073
3072
[
15]
K
.
G
ha
noua
ne
a
nd
T
.
B
e
nka
r
a
a
c
he
,
“
T
he
r
ol
e
of
a
na
ly
ti
c
a
l
s
k
il
ls
in
bi
g
da
ta
-
dr
iv
e
n
de
c
is
io
n
-
ma
ki
ng
in
a
f
r
ic
a
n
f
ir
ms
:
e
vi
de
nc
e
f
r
om M
or
oc
c
o,”
J
our
nal
of
E
c
onomic
s
and M
anage
m
e
nt
, vol
. 2
0, pp. 75
–
97, 2024.
[
16]
M
.
L
io
ua
e
ddi
ne
,
M
.
E
la
tr
a
c
hi
,
a
nd
E
.
M
.
K
a
r
a
m,
“
T
he
a
na
ly
s
i
s
of
th
e
e
f
f
ic
ie
nc
y
of
p
r
im
a
r
y
s
c
hool
s
in
M
or
oc
c
o:
mode
ll
in
g
us
in
g
T
I
M
S
S
da
ta
ba
s
e
(
2011)
,”
J
our
nal
of
N
or
th
A
fr
i
c
an
St
udi
e
s
,
vol
.
23,
no.
4,
pp.
624
–
647,
2
018,
doi
:
10.1080/13629387.2
017.1422978.
[
17]
Y
.
M
our
di
,
M
.
S
a
dga
l,
W
.
B
.
F
a
th
i,
a
nd
H
.
E
l
K
a
bt
a
ne
,
“
A
ma
c
hi
ne
le
a
r
ni
ng
ba
s
e
d
a
ppr
oa
c
h
to
e
nha
n
c
e
M
O
O
C
u
s
e
r
s
’
c
la
s
s
if
ic
a
ti
on,”
T
ur
k
is
h O
nl
in
e
J
ou
r
nal
of
D
is
ta
nc
e
E
duc
at
io
n
-
T
O
J
D
E
, vol
. 21, no. 2, pp. 47
–
68, 2020.
[
18]
T
he
W
or
ld
B
a
nk, “
M
or
oc
c
o
-
e
duc
a
ti
on,”
O
C
H
A
Se
r
v
ic
e
s
. 202
4
, doi
:
10.26599/B
D
M
A
.2019.9020007.
[
19]
H
.
B
a
r
a
ka
t,
“
P
ubl
ic
-
s
e
c
to
r
e
duc
a
ti
on
in
M
or
oc
c
o:
th
e
pe
r
s
pe
c
ti
ve
of
th
e
hi
ghe
r
c
ounc
il
f
or
e
duc
a
ti
on,
tr
a
in
in
g
a
nd
s
c
ie
nt
if
ic
r
e
s
e
a
r
c
h
(
C
S
E
F
R
S
)
,”
R
e
v
ue
M
ar
oc
ai
ne
d’
é
v
al
uat
io
n
e
t
de
la
R
e
c
he
r
c
he
E
duc
at
iv
e
(
R
M
E
R
E
)
,
pp.
77
–
96,
2021,
doi
:
10.48423/I
M
I
S
T
.P
R
S
M
/r
me
r
e
-
v0i
0.32908.
[
20]
Z
. B
ous
s
ouf
, H
. A
mr
a
ni
, M
. Z
. K
ha
l,
a
nd F
. D
a
id
a
i,
“
A
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
i
n e
duc
a
ti
on:
a
s
ys
te
ma
ti
c
l
it
e
r
a
tu
r
e
r
e
vi
e
w
,”
D
at
a
and
M
e
ta
dat
a
, vol
. 3, 2024, doi:
10.56294/dm
2024288.
[
21]
A
.
I
bou
r
k,
K
.
H
ni
ni
,
a
nd
I
.
O
ua
a
di
,
“
A
na
ly
s
is
of
th
e
pe
da
g
ogi
c
a
l
e
f
f
e
c
ti
ve
ne
s
s
of
te
a
c
he
r
qua
li
f
ic
a
ti
on
c
yc
le
in
M
or
oc
c
o:
a
ma
c
hi
ne
le
a
r
ni
ng
mode
l
a
ppr
oa
c
h,
”
in
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
on
A
dv
anc
e
d
I
nt
e
ll
ig
e
nt
Sy
s
te
m
s
fo
r
Sus
ta
in
abl
e
D
e
v
e
lo
pm
e
nt
,
2023, pp. 344
–
35
3
, doi
:
10.1007/978
-
3
-
031
-
26384
-
2_30.
[
22]
A
.
O
ua
jd
ouni
,
K
.
C
ha
f
ik
,
a
nd
O
.
B
oubke
r
,
“
M
e
a
s
ur
in
g
e
-
le
a
r
ni
ng
s
ys
te
ms
s
uc
c
e
s
s
:
da
ta
f
r
om
s
tu
de
nt
s
of
hi
ghe
r
e
duc
a
ti
on
in
s
ti
tu
ti
ons
i
n M
or
oc
c
o,”
D
at
a i
n B
r
ie
f
, vol
. 35, 2021, doi:
10.1016/j
.di
b.2021.106807.
[
23]
A
.
S
a
dqui
,
M
.
E
r
te
l,
H
.
S
a
di
ki
,
a
nd
S
.
A
ma
li
,
“
E
va
lu
a
ti
ng
ma
c
hi
ne
le
a
r
ni
ng
mode
l
s
f
or
pr
e
di
c
ti
ng
gr
a
du
a
ti
on
ti
me
li
ne
s
in
M
or
oc
c
a
n
uni
ve
r
s
it
ie
s
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
A
dv
anc
e
d
C
om
put
e
r
Sc
ie
nc
e
and
A
ppl
ic
at
io
ns
,
vol
.
14,
no.
7,
pp.
304
–
310,
2023, doi:
10.
14569/I
J
A
C
S
A
.2023.0140734.
[
24]
K
.
H
a
mda
ni
a
nd
S
.
K
ouba
a
,
“
T
he
s
tr
a
te
gi
c
pl
a
nni
ng
of
uni
ve
r
s
it
y
tr
a
ns
f
or
ma
ti
on:
th
e
c
a
s
e
of
M
or
oc
c
a
n
publ
ic
uni
ve
r
s
it
ie
s
,”
P
r
oj
e
c
ti
c
s
/
P
r
oy
é
c
ti
c
a /
P
r
oj
e
c
ti
que
, vol
. 28, no. 1, pp. 51
–
68,
2021, doi:
10.3917/pr
oj
.028.0051.
[
25]
A
.
C
ha
tr
i,
O
.
C
ha
hbi
,
a
nd
M
.
S
ni
hj
i,
“
T
he
mul
ti
le
ve
l
a
na
ly
s
is
of
s
tu
de
nt
s
’
a
c
hi
e
ve
me
nt
:
e
vi
de
nc
e
f
r
om
M
or
oc
c
o,”
A
fr
ic
an
D
e
v
e
lo
pm
e
nt
R
e
v
ie
w
, vol
. 33, no. 1, pp. 117
–
129, 2021, doi:
10
.1111/1467
-
8268.12497.
[
26]
R
.
E
l
H
a
ti
mi
,
C
.
F
.
C
houkha
n,
a
nd
M
.
E
s
ghi
r
,
“
A
n
im
pr
ove
d
k
-
me
a
ns
c
lu
s
t
e
r
in
g
a
lg
or
it
hm
to
w
a
r
ds
a
n
e
f
f
ic
ie
nt
e
duc
a
ti
ona
l
a
nd
e
c
onomi
c
a
l
da
ta
mode
li
ng,”
I
nt
e
r
nat
io
nal
J
our
nal
of
A
d
v
anc
e
d
C
om
put
e
r
Sc
ie
n
c
e
and
A
ppl
ic
at
io
ns
,
vol
.
15,
no
.
1,
pp. 1104
–
1114, 2024
, doi
:
10.14569/I
J
A
C
S
A
.2024.01501109.
[
27]
S
.
M
it
ta
l,
S
.
M
a
h
e
ndr
a
,
V
.
S
a
n
a
p,
a
nd
P
.
C
hur
i,
“
H
ow
c
a
n
ma
c
hi
ne
le
a
r
ni
ng
be
u
s
e
d
in
s
tr
e
s
s
ma
na
g
e
me
nt
:
a
s
ys
te
m
a
ti
c
li
te
r
a
t
ur
e
r
e
vi
e
w
of
a
ppl
ic
a
ti
ons
in
w
or
kpl
a
c
e
s
a
nd e
duc
a
ti
on,”
I
nt
e
r
nat
i
onal
J
our
nal
of
I
nf
or
m
a
ti
on
M
anage
m
e
nt
D
at
a
I
ns
ig
ht
s
,
vol
.
2,
no.
2, 2022, doi:
10.1016/j
.j
ji
me
i.
2022.100110.
[
28]
Z
.
K
houdi,
M
.
N
a
c
ha
oui
,
a
nd
S
.
L
ya
qi
ni
,
“
I
de
nt
if
yi
ng
th
e
c
o
nt
e
xt
ua
l
f
a
c
to
r
s
r
e
la
te
d
to
th
e
r
e
a
di
ng
p
e
r
f
or
ma
nc
e
of
M
or
oc
c
a
n
f
our
th
-
gr
a
de
s
tu
de
nt
s
f
r
om
a
ma
c
hi
ne
le
a
r
ni
ng
-
ba
s
e
d
a
ppr
o
a
c
h,”
E
duc
at
io
n
and
I
nf
or
m
at
io
n
T
e
c
hnol
ogi
e
s
,
vol
.
29,
no
.
3,
pp. 3047
–
3073, 2024, doi:
10.1007/s
10639
-
023
-
11881
-
8.
[
29]
G
.
W
.
D
e
kke
r
,
M
.
P
e
c
he
ni
z
ki
y,
a
nd
J
.
M
.
V
le
e
s
houwe
r
s
,
“
P
r
e
di
c
ti
ng
s
tu
de
nt
s
dr
op
out
:
a
c
a
s
e
s
tu
dy,”
in
2nd
I
nt
e
r
nat
i
on
al
C
onf
e
r
e
nc
e
on E
duc
at
io
nal
D
at
a M
in
in
g
, 2009, pp. 41
–
50.
[
30]
A
.
J
.
B
ow
e
r
s
,
R
.
S
pr
ot
t,
a
nd
S
.
A
.
T
a
f
f
,
“
D
o
w
e
know
w
ho
w
il
l
dr
op
out
?
:
a
r
e
vi
e
w
of
th
e
pr
e
di
c
to
r
s
of
dr
oppi
ng
out
of
hi
gh
s
c
hool
:
pr
e
c
is
io
n,
s
e
ns
it
iv
it
y,
a
nd
s
pe
c
if
ic
it
y,”
T
he
H
i
gh
Sc
hool
J
our
nal
,
vol
.
96,
no.
2,
pp.
77
–
100,
20
13,
doi
:
10.1353/hs
j.
2013.0000.
[
31]
T
.
M
.
A
la
m,
M
.
M
us
ht
a
q,
K
.
S
ha
uka
t,
I
.
A
.
H
a
me
e
d,
M
.
U
.
S
a
r
w
a
r
,
a
nd
S
.
L
uo,
“
A
nove
l
me
th
od
f
or
pe
r
f
or
ma
nc
e
me
a
s
ur
e
me
nt
of
publ
ic
e
duc
a
ti
ona
l
in
s
ti
tu
ti
ons
u
s
in
g
ma
c
hi
ne
le
a
r
ni
ng
mo
de
ls
,”
A
ppl
ie
d
Sc
ie
nc
e
s
,
vol
.
11,
no.
19,
20
21,
doi
:
10.3390/a
pp11199296.
[
32]
V
.
K
ul
e
to
e
t
al
.
,
“
E
xpl
or
in
g
oppor
tu
ni
ti
e
s
a
nd
c
ha
ll
e
nge
s
of
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
a
nd
ma
c
hi
ne
le
a
r
ni
ng
in
h
ig
he
r
e
duc
a
ti
on
in
s
ti
tu
ti
ons
,”
Sus
ta
in
abi
li
ty
, vol
. 13, no. 18, 20
21, doi:
10.3390/s
u131810424.
[
33]
M
.
P
.
I
li
ć
,
D
.
P
ă
un,
N
.
P
.
Š
e
vi
ć
,
A
.
H
a
dž
ić
,
a
nd
A
.
J
ia
nu,
“
N
e
e
ds
a
nd
pe
r
f
or
ma
nc
e
a
na
ly
s
is
f
or
c
ha
nge
s
in
hi
ghe
r
e
duc
a
ti
on
a
n
d
im
pl
e
me
nt
a
ti
on
of
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
,
ma
c
hi
ne
le
a
r
ni
ng,
a
nd
e
xt
e
nde
d
r
e
a
li
ty
,”
E
duc
at
io
n
Sc
ie
n
c
e
s
,
vol
.
11,
no.
10,
2
021,
doi
:
10.3390/e
duc
s
c
i1
1100568.
[
34]
K
.
L
.
M
.
A
ng,
F
.
L
.
G
e
,
a
nd
K
.
P
.
S
e
ng,
“
B
ig
e
duc
a
ti
ona
l
d
a
ta
&
a
na
ly
ti
c
s
:
S
ur
ve
y,
a
r
c
hi
te
c
tu
r
e
a
nd
c
ha
ll
e
nge
s
,”
I
E
E
E
A
c
c
e
s
s
,
vol
. 8, pp. 116392
–
116414, 2020, doi:
10.1109/AC
C
E
S
S
.2020.
299
4561.
[
35]
O
.
B
.
J
os
e
ph
a
nd
N
.
C
.
U
z
ondu,
“
I
nt
e
gr
a
ti
ng
A
I
a
nd
ma
c
hi
ne
le
a
r
ni
ng
in
S
T
E
M
e
duc
a
ti
on:
c
ha
ll
e
nge
s
a
nd
oppor
tu
ni
ti
e
s
,
”
C
om
put
e
r
Sc
ie
n
c
e
& I
T
R
e
s
e
ar
c
h
J
our
nal
, vol
. 5, no. 8, pp. 173
2
–
1750, 2024, doi:
10.51594/cs
it
r
j.
v5i
8.1379.
[
36]
A
.
J
.
G
a
ni
mi
a
n
a
nd
R
.
J
.
M
ur
na
ne
,
“
I
mpr
ovi
ng
e
duc
a
ti
on
in
de
ve
lo
pi
ng
c
ount
r
ie
s
:
le
s
s
ons
f
r
om
r
ig
or
ous
im
pa
c
t
e
va
lu
a
ti
o
ns
,”
R
e
v
ie
w
of
E
duc
at
io
nal
R
e
s
e
ar
c
h
, vol
. 86, no. 3, pp. 719
–
755, 2
016, doi:
10.3102/003465431
5627499.
[
37]
B
.
I
.
K
a
ngi
w
a
,
O
.
E
.
O
lu
da
r
e
,
H
.
S
.
N
a
s
s
a
r
a
w
a
,
N
.
S
.
A
bu
ba
ka
r
,
E
.
L
.
E
f
e
oma
,
a
nd
H
.
A
.
E
ne
f
ol
a
,
“
L
e
ve
r
a
gi
ng
a
r
ti
f
i
c
ia
l
in
te
ll
ig
e
nc
e
f
or
e
nha
nc
in
g
e
nt
r
e
pr
e
ne
ur
s
hi
p
a
nd
c
r
e
a
ti
vi
ty
in
S
T
E
M
e
duc
a
ti
on,”
J
our
nal
of
E
duc
at
io
nal
R
e
s
e
ar
c
h
and
P
r
ac
ti
c
e
,
vol
. 4, no. 8, pp. 149
–
162, 2024.
[
38]
M
.
S
ki
tt
ou,
M
.
M
e
r
r
ouc
hi
,
a
nd
T
.
G
a
di
,
“
A
mode
l
of
a
n
in
te
gr
a
te
d
e
duc
a
ti
ona
l
ma
na
ge
m
e
nt
in
f
or
ma
ti
on
s
ys
t
e
m
to
s
up
por
t
e
duc
a
ti
ona
l
pl
a
nni
ng
a
nd
de
c
i
s
io
n
ma
ki
ng:
a
M
or
oc
c
a
n c
a
s
e
,
”
i
n
W
I
T
S
2020:
P
r
oc
e
e
di
ngs
of
th
e
6t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
on
W
ir
e
le
s
s
T
e
c
hnol
ogi
e
s
, E
m
be
dde
d, and I
nt
e
ll
ig
e
nt
Sy
s
t
e
m
s
, 202
2, pp. 167
–
177
, doi
:
10.1007/978
-
981
-
33
-
6893
-
4_16.
[
39]
H
.
A
lm
a
ghr
a
bi
,
B
.
S
oh,
a
nd
A
.
L
i,
“
U
s
in
g
M
L
to
pr
e
di
c
t
us
e
r
s
a
ti
s
f
a
c
ti
on
w
it
h
I
C
T
te
c
hnol
ogy
f
or
e
duc
a
ti
ona
l
in
s
ti
tu
ti
on
a
dmi
ni
s
tr
a
ti
on. i
nf
or
ma
ti
on,”
I
nf
or
m
at
io
n
, vol
. 15, no. 4, 2024,
doi
:
10.3390/i
nf
o15040218.
[
40]
A
.
Y
a
s
s
in
e
a
nd
F
.
B
a
k
a
s
s
, “
Y
out
h’
s
pov
e
r
ty
a
nd
in
e
qua
li
ty
of
oppor
tu
ni
ti
e
s
:
e
mpi
r
ic
a
l
e
vi
de
nc
e
f
r
om
M
or
oc
c
o,”
Soc
ia
l
Sc
ie
nc
e
s
,
vol
. 12, no. 1, 2023, doi
:
10.3390/s
oc
s
c
i1
2010028.
[
41]
M
.
B
e
r
r
ia
ne
,
H
.
D
e
H
a
a
s
,
a
nd
K
.
N
a
tt
e
r
,
“
S
oc
ia
l
tr
a
ns
f
or
ma
ti
ons
a
nd
mi
gr
a
ti
ons
in
M
or
oc
c
o,”
I
nt
e
r
nat
io
nal
M
ig
r
a
ti
on
I
ns
ti
tu
te
W
or
k
in
g P
ape
r
Se
r
ie
s
, vol
. 171, pp. 1
–
47, 2021.
[
42]
W
. A
. O
w
in
gs
a
nd L
. S
. K
a
pl
a
n,
E
qui
ty
audit
s
and s
c
hool
r
e
s
our
c
e
al
lo
c
at
io
n:
apply
in
g c
r
it
ic
al
r
e
s
ou
r
c
e
t
he
or
y
t
o i
nc
r
e
as
e
e
qual
oppor
tu
ni
ty
i
n s
c
hool
s
. N
e
w
Y
or
k, U
ni
te
d S
ta
te
s
:
R
out
le
dg
e
,
2
024
, doi
:
10.4324/978100349
3907.
[
43]
Y
.
C
he
n
a
nd
L
.
Z
ha
i,
“
A
c
ompa
r
a
ti
ve
s
tu
dy
on
s
tu
de
nt
p
e
r
f
or
ma
nc
e
pr
e
di
c
ti
on
us
in
g
ma
c
hi
ne
le
a
r
ni
ng,”
E
duc
at
io
n
and
I
nf
or
m
at
io
n T
e
c
hnol
ogi
e
s
, vol
. 28, no. 9, pp. 12039
–
12057, 2023, doi:
10.1007/s
10639
-
023
-
11672
-
1.
[
44
]
M
.
A
kdi
m
e
t
al
.
, “
S
c
a
le
’
s
im
pa
c
t
in
e
duc
a
ti
on
s
ys
t
e
m’
s
pe
r
f
or
ma
nc
e
:
C
a
s
e
s
in
D
r
a
a
-
T
a
f
il
a
le
t,
M
or
oc
c
o,”
I
nt
e
r
nat
io
nal
J
our
n
al
of
E
v
al
uat
io
n and R
e
s
e
ar
c
h i
n E
duc
at
io
n
, vol
. 12, no. 4, pp. 2374
–
2386, 2023, doi:
10.11591/i
je
r
e
.v12i4.25100.
[
45]
U
N
E
S
C
O
,
U
N
I
C
E
F
,
a
nd
W
or
ld
B
a
nk,
T
he
s
ta
te
of
th
e
gl
obal
e
duc
at
io
n
c
r
is
is
:
a
pat
h
to
r
e
c
ov
e
r
y
.
W
a
s
hi
ngt
on,
U
ni
te
d
S
ta
te
s
:
T
he
W
or
ld
B
a
nk, UN
E
S
C
O
a
nd
U
N
I
C
E
F
, 2021
, doi
:
10.54675/J
L
U
G
7649.
Evaluation Warning : The document was created with Spire.PDF for Python.