I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
A
I
)
V
ol
. 13, No. 4, D
e
c
e
m
be
r
2024, pp.
4414
~
4421
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v13.i4.pp
4414
-
4421
4414
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
N
e
t
w
o
r
k
i
n
t
r
u
si
on
d
e
t
e
c
t
i
on
i
n
b
i
g
d
a
t
ase
t
s u
si
n
g S
p
ar
k
e
n
vi
r
on
m
e
n
t
an
d
i
n
c
r
e
m
e
n
t
al
l
e
ar
n
i
n
g
A
b
d
e
lwah
e
d
E
lm
ou
t
aou
k
k
il
, M
oh
am
e
d
H
am
li
c
h
,
A
m
in
e
K
h
at
ib
, M
ar
ou
an
e
C
h
r
is
s
L
a
bor
a
t
or
y of
C
om
pl
e
x C
ybe
r
-
P
hys
i
c
a
l
S
ys
t
e
m
s
, N
a
t
i
ona
l
S
c
hool
of
A
r
t
s
a
nd
C
r
a
f
t
s
, C
a
s
a
bl
a
nc
a
, U
ni
ve
r
s
i
t
y H
a
s
s
a
n I
I
,
C
a
s
a
bl
a
n
c
a
, M
or
oc
c
o
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
N
ov
23
,
2023
R
e
vi
s
e
d
M
a
r
24
,
2024
A
c
c
e
pt
e
d
A
pt
17
,
2024
Internet
of
things
(IoT)
systems
have
experienced
significant
growth
i
n
data
traffic
,
resulting
in
security
and
real
-
time
processing
issues.
Int
rusion
detection
systems
(IDS)
are
currently
an
indispensable
tool
for
self
-
protection
against
various
attacks.
However, IoT
systems face
serious challenges
due to
the
functional
diversity
of
attacks,
resulting
in
detection
method
s
with
machine
learning
(ML)
and
limited
static
models
g
enerated
by
the
linear
discriminant
analysis
(LDA)
algorithm.
The
process
entails
adjusti
ng
the
model para
meters in r
eal time
as new da
ta
arrives.
This paper
proposes
a new
method
of
an
IDS
based
on
the
LDA
algorithm
with
the
inc
remental
model.
The
model
framework
is
trained
and
tested
on
the
IoT
intrusion
dataset
(UNSW
-
NB15) using t
he streaming lin
ear discriminant
analysis (SLD
A)
ML
algorit
hm.
Our
approach
increased
model
accuracy
after
each
training
,
resulting
in conti
nuous mod
e
l improvemen
t. The compa
rison reve
als that our
dynamic
model
becomes
more
accurate
after
each
batch
and
can
dete
ct
new
types of a
ttacks.
K
e
y
w
o
r
ds
:
B
ig
da
ta
I
nc
r
e
m
e
nt
a
l
le
a
r
ni
ng
I
nt
e
r
ne
t
of
t
hi
ngs
I
nt
r
us
io
n de
te
c
ti
on s
ys
te
m
M
a
c
hi
ne
l
e
a
r
ni
ng
S
tr
e
a
m
in
g l
in
e
a
r
di
s
c
r
im
in
a
nt
a
na
ly
s
is
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
:
A
bde
lwa
he
d
E
lm
out
a
oukkil
L
a
bor
a
to
r
y of
C
om
pl
e
x C
ybe
r
-
P
hys
ic
a
l
S
ys
te
m
, N
a
ti
ona
l
S
c
hool
of
A
r
ts
a
nd C
r
a
f
ts
C
a
s
a
bl
a
nc
a
U
ni
ve
r
s
it
y H
a
s
s
a
n I
I
C
a
s
a
bl
a
n
c
a
, M
or
oc
c
o
E
m
a
il
:
e
lm
out
a
oukkil
a
bd@
gm
a
il
.c
om
1.
I
N
T
R
O
D
U
C
T
I
O
N
A
n
in
tr
us
io
n
de
te
c
ti
on
s
ys
te
m
(
I
D
S
)
[
1]
,
[
2]
is
a
n
in
s
tr
um
e
nt
or
s
of
twa
r
e
a
ppl
ic
a
ti
on
th
a
t
s
ur
ve
ys
th
e
ne
twor
k
a
nd
th
e
s
y
s
te
m
f
or
c
yni
c
a
l
a
c
ti
vi
ti
e
s
a
nd
w
a
r
ns
th
e
s
ys
te
m
or
ne
twor
k
a
dm
in
is
tr
a
to
r
.
T
he
r
e
a
r
e
two
ki
nds
of
I
D
S
,
ne
twor
k
-
ba
s
e
d
I
D
S
,
a
nd
hos
t
-
ba
s
e
d
I
D
S
.
A
h
os
t
-
ba
s
e
d
I
D
S
ke
e
p
s
tr
a
c
k
of
in
di
vi
dua
l
hos
t
m
a
c
hi
ne
s
a
nd
gi
ve
s
not
if
ic
a
ti
on
to
th
e
us
e
r
if
s
u
s
pi
c
io
us
a
c
ti
v
it
ie
s
li
ke
de
l
e
ti
ng
or
m
odi
f
yi
ng
a
s
ys
te
m
f
il
e
,
unde
s
ir
e
d
c
onf
ig
ur
a
ti
on
c
ha
nge
s
,
or
unne
c
e
s
s
a
r
y
s
e
qu
e
nc
e
o
f
s
ys
te
m
c
a
ll
s
a
r
e
de
te
c
te
d
[
3]
.
G
e
ne
r
a
ll
y,
a
ne
twor
k
-
ba
s
e
d
in
tr
us
io
n
de
te
c
ti
on
s
ys
te
m
(
N
I
D
S
)
[
4]
is
ke
pt
a
t
ne
twor
k
poi
nt
s
li
ke
a
ga
te
w
a
y
or
r
out
e
r
s
t
o
de
te
c
t
in
tr
us
io
ns
in
th
e
ne
twor
k
t
r
a
f
f
ic
.
L
a
te
ly
,
a
r
ti
f
ic
ia
l
in
t
e
ll
ig
e
nc
e
(
A
I
)
ha
s
be
e
n
us
e
d
in
th
e
f
ie
ld
of
c
ybe
r
s
e
c
ur
it
y,
a
nd
to
a
c
hi
e
ve
hi
gh
-
pe
r
f
or
m
a
nc
e
I
D
S
,
m
a
c
hi
ne
le
a
r
ni
ng
(
M
L
)
te
c
hni
que
s
c
a
n
be
us
e
d
f
or
a
ll
ty
pe
s
of
de
te
c
ti
on t
e
c
hni
que
s
.
ML
i
s
a
s
ubc
la
s
s
of
A
I
us
e
d
i
n c
o
m
put
e
r
s
w
it
h t
he
a
bi
li
ty
t
o
l
e
a
r
n w
it
hout
be
in
g
c
a
lc
ul
a
te
d.
ML
a
lg
or
it
hm
s
c
a
n
be
s
upe
r
vi
s
e
d
or
uns
upe
r
vi
s
e
d,
th
e
f
ir
s
t
c
a
t
e
gor
y
bui
ld
s
a
m
a
th
e
m
a
ti
c
a
l
m
ode
l
of
a
da
ta
s
e
t
th
a
t
c
ont
a
in
s
bot
h
th
e
de
s
ir
e
d
in
put
s
a
nd
out
put
s
.
T
he
d
a
ta
is
known
a
s
tr
a
in
in
g
da
ta
a
nd c
ons
is
t
s
of
a
s
e
t
of
t
r
a
in
in
g e
xa
m
pl
e
s
. E
a
c
h t
r
a
in
in
g e
xa
m
pl
e
ha
s
one
or
m
or
e
i
nput
s
a
nd output
s
.
T
he
s
e
c
ond
c
a
te
gor
y
ta
ke
s
a
s
e
t
of
da
ta
c
ont
a
in
in
g
onl
y
in
put
s
a
nd
f
in
ds
a
pa
tt
e
r
n
in
th
e
da
ta
,
s
uc
h
a
s
gr
oupi
ng
or
c
lu
s
te
r
in
g
da
ta
poi
nt
s
. T
he
a
lg
or
it
hm
s
, t
he
r
e
f
or
e
, l
e
a
r
n f
r
om
t
e
s
t
da
ta
t
ha
t
ha
s
n
ot
be
e
n l
a
be
le
d,
c
la
s
s
if
ie
d, or
c
a
te
gor
iz
e
d
[
5]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
N
e
tw
or
k
i
nt
r
us
io
n de
te
c
ti
on i
n bi
g datas
e
ts
us
in
g Spa
r
k
e
n
v
ir
on
m
e
nt
and …
(
A
bde
lw
ahe
d
E
lmoutaoukk
il
)
4415
I
n
our
r
e
s
e
a
r
c
h
w
or
k,
w
e
ha
ve
pr
opos
e
d
a
f
r
a
m
e
w
or
k
in
w
hi
c
h
a
f
e
a
tu
r
e
r
e
duc
ti
on
a
lg
or
it
h
m
is
us
e
d
f
or
e
li
m
in
a
ti
ng
th
e
le
s
s
im
por
ta
nt
f
e
a
tu
r
e
s
.
T
he
n
w
e
a
ppl
ie
d
s
upe
r
vi
s
e
d
da
ta
m
in
in
g
te
c
hni
que
s
on
th
e
U
N
S
W
-
N
B
15 ne
twor
k da
ta
s
e
t
f
or
f
a
s
t,
e
f
f
ic
ie
nt
, a
nd a
c
c
ur
a
te
d
e
te
c
ti
on of
i
nt
r
us
io
n i
n t
he
N
e
tf
lo
w
r
e
c
or
ds
by
le
ve
r
a
gi
ng
th
e
pow
e
r
of
S
pa
r
k.
I
n
th
is
pa
pe
r
,
w
e
ha
ve
us
e
d
pr
in
c
ip
a
l
c
om
pone
nt
a
na
ly
s
is
(
P
C
A
)
a
nd
s
tr
e
a
m
in
g
li
ne
a
r
di
s
c
r
im
in
a
nt
a
na
ly
s
is
(
S
L
D
A
)
a
lg
or
it
hm
s
to
a
na
ly
z
e
th
e
pe
r
f
or
m
a
nc
e
of
th
e
pr
opos
e
d
f
r
a
m
e
w
or
k.
W
e
f
oc
us
e
d on thr
e
e
pe
r
f
or
m
a
nc
e
i
ndi
c
a
to
r
s
:
a
c
c
ur
a
c
y,
tr
a
in
in
g t
im
e
, a
nd dur
a
ti
on
of
pr
e
di
c
ti
on.
2.
M
E
T
H
O
D
A
gr
a
w
a
l
a
nd
A
gr
a
w
a
l
[
6]
ha
ve
s
ur
ve
ye
d
a
bnor
m
a
li
ty
de
te
c
ti
on
w
it
h
da
ta
m
in
in
g
te
c
hni
qu
e
s
to
de
te
c
t
in
tr
us
io
ns
.
T
he
y
ha
ve
c
la
s
s
if
ie
d
th
e
a
nom
a
ly
de
te
c
ti
on
t
e
c
h
ni
que
s
w
it
h
th
r
e
e
f
e
a
tu
r
e
s
:
c
lu
s
te
r
in
g
-
ba
s
e
d
te
c
hni
que
s
,
c
l
a
s
s
if
ic
a
ti
on
-
ba
s
e
d
te
c
hni
qu
e
s
,
a
nd
hybr
id
te
c
hni
q
ue
s
.
T
he
ir
c
om
pr
e
he
ns
iv
e
a
na
ly
s
is
s
he
ds
li
ght
on t
he
di
ve
r
s
e
a
ppr
oa
c
he
s
ut
il
iz
e
d i
n t
he
ongoing pur
s
ui
t
of
e
nha
nc
in
g c
ybe
r
s
e
c
ur
it
y m
e
a
s
ur
e
s
.
B
uc
z
a
k
a
nd
G
uve
n
[
7]
m
a
de
a
s
ur
ve
y
th
a
t
de
s
c
r
ib
e
s
th
e
a
p
pl
ic
a
ti
on
of
da
ta
m
in
in
g
a
nd
ML
te
c
hni
que
s
to
de
te
c
t
known
a
nd
unknown
a
tt
a
c
ks
.
T
h
e
y
e
s
ta
bl
i
s
he
d
a
c
le
a
r
di
f
f
e
r
e
nc
e
be
twe
e
n
M
L
a
nd
d
a
ta
m
in
in
g.
T
he
ir
s
tu
dy
unde
r
s
c
or
e
s
th
e
gr
ow
in
g
s
ig
ni
f
ic
a
nc
e
of
th
e
s
e
a
ppr
oa
c
h
e
s
in
pr
e
v
e
nt
in
g
a
nd
de
te
c
ti
ng
th
r
e
a
ts
w
it
hi
n t
he
c
om
pl
e
x l
a
nds
c
a
pe
of
m
ode
r
n c
ybe
r
s
e
c
ur
it
y.
H
a
ij
a
e
t
al
.
[
8]
ha
ve
done
a
ne
w
in
c
lu
s
iv
e
di
s
c
ove
r
y
s
c
he
m
e
th
a
t
e
va
lu
a
te
s
f
iv
e
s
upe
r
vi
s
e
d
ML
c
la
s
s
if
ie
r
s
:
lo
gi
s
ti
c
r
e
gr
e
s
s
io
n,
de
c
is
io
n
tr
e
e
s
,
li
ne
a
r
/q
ua
dr
a
ti
c
di
s
c
r
im
in
a
nt
,
na
ïv
e
B
a
y
e
s
,
a
nd
e
ns
e
m
bl
e
boo
s
te
d
tr
e
e
s
. P
or
t
s
c
a
nni
ng a
tt
a
c
k
s
i
nvol
ve
a
tt
a
c
ke
r
s
s
e
ndi
ng pa
c
k
e
ts
w
it
h va
r
io
us
por
t
num
be
r
s
t
o s
c
a
n f
or
a
c
c
e
s
s
ib
le
s
e
r
vi
c
e
s
a
nd
id
e
nt
if
y
ope
n
o
r
w
e
a
k
por
ts
in
a
ne
twor
k,
ne
c
e
s
s
it
a
ti
ng
th
e
de
ve
lo
pm
e
nt
of
m
ul
t
ip
le
de
te
c
ti
on
a
nd
pr
e
ve
nt
io
n
te
c
hni
que
s
.
T
hr
ough
pe
r
f
or
m
a
nc
e
c
om
pa
r
is
on
us
in
g
th
e
P
S
A
-
2017
da
ta
s
e
t,
th
e
lo
gi
s
ti
c
r
e
gr
e
s
s
io
n
m
ode
l
de
m
ons
tr
a
te
d s
up
e
r
io
r
r
e
s
ul
ts
w
it
h 99.4%
a
c
c
ur
a
c
y, 99.9
%
pr
e
c
is
io
n, 99.4%
r
e
c
a
ll
, 99.7%
F
-
s
c
or
e
, a
nd
a
de
te
c
ti
on
ove
r
he
a
d
of
0.454
µ
S
e
c
,
hi
ghl
ig
ht
in
g
it
s
e
f
f
e
c
ti
ve
ne
s
s
a
nd
e
nha
nc
e
d
a
tt
a
c
k
di
s
c
ove
r
y
s
pe
e
d
c
om
pa
r
e
d t
o e
xi
s
ti
ng mode
ls
.
H
a
ij
a
e
t
al
.
[
9]
ha
ve
in
tr
oduc
e
d
a
n
a
dv
a
nc
e
d
s
e
lf
-
r
e
li
a
nt
s
y
s
te
m
de
s
ig
ne
d
to
de
te
c
t
m
ut
a
ti
ons
of
in
te
r
ne
t
of
th
in
gs
(
I
oT
)
c
yb
e
r
-
a
tt
a
c
ks
u
s
in
g
a
de
e
p
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
(
C
N
N
)
a
nd
C
U
D
A
-
ba
s
e
d
N
vi
di
a
-
Q
ua
d
G
P
U
s
f
or
pa
r
a
ll
e
l
c
om
put
a
ti
on.
T
hi
s
in
nova
ti
ve
a
ppr
oa
c
h
a
c
hi
e
v
e
d
e
xc
e
pt
io
na
l
a
tt
a
c
k
c
la
s
s
if
ic
a
ti
on
a
c
c
ur
a
c
y,
e
x
c
e
e
di
ng
99.3%
f
or
bi
na
r
y
c
la
s
s
if
ie
r
s
a
nd
98.2%
f
o
r
m
ul
ti
-
c
la
s
s
c
la
s
s
if
ie
r
s
.
T
h
e
y
hi
ghl
ig
ht
th
e
s
ubs
t
a
nt
ia
l
gr
ow
th
a
nd
im
pa
c
t
of
th
e
I
o
T
te
c
hnol
ogy,
w
hi
le
a
l
s
o
a
ddr
e
s
s
in
g
it
s
vul
ne
r
a
bi
li
ty
to
c
ybe
r
-
a
tt
a
c
ks
d
ue
t
o t
he
l
im
it
a
ti
ons
i
n
c
om
put
a
ti
on, s
to
r
a
ge
, a
n
d c
om
m
uni
c
a
ti
on c
a
pa
c
it
y of
e
ndpoint
de
vi
c
e
s
s
uc
h a
s
t
he
r
m
os
t
a
ts
a
nd home
a
ppl
ia
nc
e
s
.
D
ha
nya
e
t
al
.
[
10]
ha
ve
s
ur
ve
ye
d
va
r
io
us
ML
-
ba
s
e
d
te
c
hni
que
s
a
ppl
ie
d
to
U
N
S
W
-
N
B
15
da
ta
s
e
t.
T
he
y c
om
p
a
r
e
d t
he
n
a
ïv
e
B
a
ye
s
a
lg
or
it
hm
w
it
h pr
opos
e
d
pr
oba
bi
li
ty
-
ba
s
e
d s
upe
r
vi
s
e
d
ML
a
lg
or
it
hm
s
us
in
g
a
r
e
duc
e
d
U
N
S
W
N
B
15
da
ta
s
e
t.
T
he
ir
e
xa
m
in
a
ti
on
of
f
e
r
s
va
lu
a
bl
e
in
s
ig
ht
s
in
to
th
e
e
f
f
ic
a
c
y
of
di
f
f
e
r
e
nt
ML
a
ppr
oa
c
he
s
w
he
n
a
ppl
ie
d
to
th
is
da
ta
s
e
t,
c
ont
r
ib
ut
in
g
to
th
e
ongoing
di
s
c
our
s
e
on
c
ybe
r
s
e
c
ur
it
y
m
e
th
odol
ogi
e
s
.
A
ls
ul
a
m
i
e
t
al
.
[
11]
ha
ve
pr
opos
e
d a
pr
e
di
c
ti
ve
ML
m
ode
l
to
de
te
c
t
a
nd c
la
s
s
if
y ne
twor
k
a
c
ti
vi
ty
i
n
a
n
I
oT
s
ys
te
m
.
S
pe
c
if
ic
a
ll
y,
th
e
m
ode
l
di
s
ti
ngui
s
he
s
be
tw
e
e
n
nor
m
a
l
a
nd
a
nom
a
ly
ne
twor
k
a
c
ti
vi
ty
.
F
ur
th
e
r
m
or
e
,
it
c
la
s
s
if
ie
s
ne
twor
k
tr
a
f
f
ic
in
to
f
iv
e
c
a
te
gor
ie
s
:
nor
m
a
l,
M
ir
a
i
a
tt
a
c
k,
de
ni
a
l
of
s
e
r
vi
c
e
(
D
o
S
)
a
tt
a
c
k,
S
c
a
n
a
tt
a
c
k,
a
nd
m
a
n
-
in
-
th
e
-
m
id
dl
e
(
M
I
T
M
)
a
tt
a
c
k.
F
iv
e
s
upe
r
vi
s
e
d
le
a
r
ni
ng
m
ode
l
s
w
e
r
e
im
pl
e
m
e
nt
e
d
to
c
ha
r
a
c
te
r
iz
e
t
he
ir
pe
r
f
or
m
a
nc
e
i
n de
te
c
ti
ng a
nd c
la
s
s
if
yi
ng n
e
twor
k a
c
ti
vi
ti
e
s
f
or
I
oT
s
ys
te
m
s
.
H
a
ij
a
[
12]
in
tr
oduc
e
s
a
n
in
nova
ti
ve
a
nd
ve
r
s
a
ti
le
to
p
-
dow
n
f
r
a
m
e
w
or
k
f
or
e
nha
nc
in
g
in
tr
us
io
n
de
te
c
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
w
it
hi
n
I
oT
ne
twor
ks
th
r
ough
th
e
u
ti
li
z
a
ti
on
of
non
-
c
onve
nt
io
na
l
ML
te
c
hni
que
s
.
T
he
a
r
ti
c
le
put
s
f
or
th
a
nove
l
a
r
c
hi
te
c
tu
r
e
th
a
t
o
f
f
e
r
s
th
e
f
le
xi
b
il
it
y
to
a
da
pt
a
nd
a
ppl
y
to
in
tr
us
io
n
de
te
c
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
ta
s
ks
in
vol
vi
ng
v
a
r
io
us
I
oT
c
ybe
r
-
a
tt
a
c
k
d
a
ta
s
e
ts
,
s
uc
h
a
s
th
e
C
I
C
I
D
S
d
a
ta
s
e
t
a
nd
M
Q
T
T
da
ta
s
e
t.
M
or
e
s
pe
c
if
ic
a
ll
y,
th
is
ne
w
ly
pr
opos
e
d
s
ys
te
m
c
om
pr
is
e
s
th
r
e
e
di
s
ti
nc
t
s
ub
s
ys
te
m
s
:
th
e
f
e
a
tu
r
e
e
ngi
ne
e
r
in
g
(
F
E
)
s
ubs
ys
te
m
,
th
e
f
e
a
tu
r
e
le
a
r
ni
ng
(
F
L
)
s
ubs
ys
t
e
m
,
a
nd
th
e
de
te
c
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
(
D
C
)
s
ubs
ys
te
m
.
H
a
ij
a
e
t
al
.
[
13
]
c
onduc
te
d
a
s
tu
dy
th
a
t
de
ve
lo
ps
a
nd
e
va
lu
a
te
s
m
a
c
hi
ne
-
le
a
r
ni
ng
-
ba
s
e
d
D
a
r
kne
t
tr
a
f
f
ic
de
te
c
ti
on
s
ys
te
m
s
(
D
T
D
S
)
in
I
oT
ne
twor
ks
,
ut
il
iz
in
g
s
ix
s
upe
r
vi
s
e
d
m
a
c
hi
ne
-
le
a
r
ni
ng
te
c
hni
que
s
.
T
he
ir
r
e
s
e
a
r
c
h
hi
ghl
ig
ht
s
th
e
a
ppl
ic
a
ti
on
of
th
e
s
e
te
c
hni
qu
e
s
to
a
ddr
e
s
s
vu
ln
e
r
a
bi
li
ti
e
s
in
I
oT
in
f
r
a
s
tr
uc
tu
r
e
s
due
to
li
m
it
e
d e
ndpoint
de
vi
c
e
c
a
pa
bi
li
ti
e
s
. T
he
s
tu
dy unde
r
s
c
or
e
s
t
he
e
f
f
e
c
ti
ve
ne
s
s
of
ba
ggi
ng e
ns
e
m
bl
e
t
e
c
hni
que
s
(
B
A
G
-
D
T
)
in
a
c
hi
e
vi
ng
s
upe
r
io
r
a
c
c
ur
a
c
y
a
nd
lo
w
e
r
e
r
r
or
r
a
te
s
,
de
m
ons
tr
a
ti
ng
s
ig
ni
f
ic
a
nt
im
pr
ove
m
e
nt
s
ove
r
e
xi
s
ti
ng D
T
D
S
m
ode
ls
by 1.9%
t
o 27%
.
T
he
m
a
jo
r
it
y
of
r
e
s
e
a
r
c
he
r
s
e
m
pl
oyi
ng
ML
in
th
e
c
ybe
r
s
e
c
ur
it
y
dom
a
in
c
onc
e
nt
r
a
te
on
a
s
s
e
s
s
in
g
th
e
pe
r
f
or
m
a
nc
e
of
m
ode
ls
ge
ne
r
a
te
d
us
in
g
di
ve
r
s
e
ML
a
lg
or
it
hm
s
.
H
ow
e
ve
r
,
th
e
s
e
pe
r
f
or
m
a
nc
e
s
,
s
uc
h
a
s
a
c
c
ur
a
c
y
a
nd
pr
e
di
c
ti
on
ti
m
e
,
do
not
e
xhi
bi
t
te
m
por
a
l
va
r
ia
ti
ons
w
it
h
th
e
e
m
e
r
ge
nc
e
of
ne
w
a
tt
a
c
ks
,
ul
ti
m
a
te
ly
r
e
nde
r
in
g
th
e
s
e
m
ode
ls
out
da
te
d
ove
r
ti
m
e
.
T
hi
s
r
e
s
e
a
r
c
h
un
de
r
s
c
or
e
s
two
s
ig
ni
f
ic
a
nt
a
s
pe
c
t
s
:
f
ir
s
tl
y,
th
e
e
f
f
e
c
ti
ve
ne
s
s
of
i
nt
r
us
io
n de
te
c
ti
on me
th
ods
e
m
pl
oyi
ng i
nc
r
e
m
e
nt
a
l
a
lg
or
it
hm
s
, s
uc
h a
s
S
L
D
A
, i
n i
d
e
nt
if
yi
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
, V
ol
. 13, No. 4, D
e
c
e
m
be
r
2024:
4414
-
4421
4416
e
m
e
r
gi
ng
a
tt
a
c
k
pa
tt
e
r
ns
,
a
nd
s
e
c
ondl
y,
th
e
pr
ogr
e
s
s
iv
e
e
nh
a
nc
e
m
e
nt
of
th
e
ir
a
c
c
ur
a
c
y
ove
r
ti
m
e
.
T
he
U
N
S
W
-
N
B
15 da
ta
s
e
t
h
a
s
be
e
n
c
ur
a
te
d f
or
e
xpe
r
im
e
nt
a
ti
on w
it
hi
n t
he
S
pa
r
k t
ool
e
nvi
r
onm
e
nt
.
2.1.
D
e
s
c
r
ip
t
io
n
of
U
N
S
W
-
N
B
15 d
at
as
e
t
T
he
U
N
S
W
N
B
15
da
t
a
s
e
t
h
a
s
c
a
r
ve
d
it
s
pl
a
c
e
a
s
a
c
or
n
e
r
s
to
ne
in
th
e
f
i
e
ld
of
ne
t
w
or
k
s
e
c
ur
it
y
r
e
s
e
a
r
c
h
.
N
um
e
r
ou
s
r
e
s
e
a
r
c
he
r
s
ha
ve
h
a
r
ne
s
s
e
d
th
e
pot
e
nt
i
a
l
of
th
is
d
a
ta
s
e
t
t
o
unr
a
ve
l
th
e
c
om
pl
e
xi
ti
e
s
of
ne
twor
k
be
ha
vi
or
,
in
tr
u
s
io
n
de
te
c
t
io
n
[
1
4]
,
a
nd
th
e
e
ve
r
-
e
vol
v
in
g
la
nds
c
a
p
e
of
c
y
be
r
th
r
e
a
t
s
.
U
N
S
W
N
B
15
s
ta
n
ds
a
s
a
te
s
ta
m
e
nt
t
o
th
e
pi
vot
a
l
r
ol
e
t
ha
t
d
a
ta
pl
a
y
s
in
c
om
pr
e
he
ndi
n
g
a
nd
s
a
f
e
gu
a
r
di
ng
a
ga
i
ns
t
c
yb
e
r
-
a
tt
a
c
k
s
.
F
or
th
e
e
va
lu
a
ti
on
of
th
e
pe
r
f
or
m
a
nc
e
a
nd
e
f
f
e
c
ti
ve
ne
s
s
of
our
I
D
S
,
w
e
r
e
qui
r
e
d
a
c
om
pr
e
he
n
s
iv
e
da
ta
s
e
t
th
a
t
c
ont
a
in
s
bot
h
nor
m
a
l
a
nd
a
bnor
m
a
l
b
e
ha
vi
or
s
.
A
lo
t
of
r
e
s
e
a
r
c
h
ha
s
be
e
n
done
us
in
g
ol
d
e
r
be
nc
hm
a
r
k
da
ta
s
e
ts
li
ke
K
D
D
C
U
P
99
a
nd
N
S
L
K
D
D
but
th
e
s
e
d
a
ta
s
e
t
s
do
not
of
f
e
r
r
e
a
li
s
ti
c
out
put
pe
r
f
or
m
a
nc
e
. T
he
r
e
a
s
on i
s
t
ha
t
K
D
D
C
U
P
99 ha
s
l
ot
s
of
r
e
dunda
nt
a
nd mi
s
s
in
g r
e
c
or
ds
i
n t
he
t
r
a
in
in
g s
e
t.
S
o
th
e
s
e
da
ta
s
e
ts
a
r
e
not
c
om
pr
e
he
ns
iv
e
r
e
pr
e
s
e
nt
a
ti
ons
of
m
ode
r
n
lo
w
f
oot
pr
in
t
a
tt
a
c
k
e
nvi
r
onm
e
nt
s
.
U
N
S
W
-
N
B
15
d
a
ta
s
e
t
[
15]
w
a
s
c
r
e
a
t
e
d
by
th
e
I
X
I
A
P
e
r
f
e
c
tS
to
r
m
to
ol
in
th
e
C
yb
e
r
R
a
nge
L
a
b
of
th
e
A
u
s
tr
a
li
a
n
C
e
nt
r
e
f
or
C
ybe
r
S
e
c
ur
it
y (
A
C
C
S
)
. I
t
c
ont
a
in
s
bot
h r
e
a
l
m
ode
r
n nor
m
a
l
a
c
ti
vi
ti
e
s
a
nd s
ynt
he
ti
c
c
ont
e
m
por
a
r
y
a
tt
a
c
k
be
ha
vi
or
s
[
16]
.
U
N
S
W
-
N
B
15
da
t
a
s
e
t
i
s
a
v
a
il
a
bl
e
in
c
om
m
a
-
s
e
pa
r
a
te
d
va
lu
e
s
(
C
S
V
)
f
il
e
f
or
m
a
t.
T
he
r
e
a
r
e
175,341 r
e
c
or
ds
i
n t
he
t
r
a
in
in
g s
e
t
a
nd 82,332 r
e
c
or
ds
i
n t
he
t
e
s
ti
ng s
e
t
w
it
h a
ll
di
f
f
e
r
e
nt
9 t
ype
s
of
a
tt
a
c
ks
a
nd
nor
m
a
l
r
e
c
or
ds
.
T
he
r
e
a
r
e
49
a
tt
r
ib
ut
e
s
or
f
e
a
tu
r
e
s
w
it
h
1
0
c
la
s
s
va
lu
e
s
in
th
is
da
ta
s
e
t.
A
ll
r
e
c
or
ds
a
r
e
di
vi
de
d
in
to
two
m
a
jo
r
c
a
te
gor
ie
s
of
th
e
r
e
c
or
d
s
-
nor
m
a
l
a
nd
a
tt
a
c
k.
T
he
a
tt
a
c
k
c
a
te
gor
ie
s
a
s
s
ho
w
n
in
T
a
bl
e
1
a
r
e
a
ga
in
s
ubdi
vi
de
d
in
to
9
c
a
te
gor
ie
s
.
A
tt
a
c
k
ty
pe
s
a
r
e
f
uz
z
e
r
s
,
a
na
ly
s
is
,
ba
c
kdoor
s
,
D
o
S
,
e
xpl
oi
ts
,
ge
ne
r
ic
, r
e
c
onna
is
s
a
nc
e
,
s
he
ll
c
ode
, a
nd
w
or
m
s
[
17]
.
T
a
bl
e
1
.
L
is
t
of
a
tt
a
c
ks
U
N
S
W
N
B
-
15 da
ta
s
e
t
C
a
t
e
gor
y
T
r
a
i
ni
ng s
e
t
T
e
s
t
i
ng s
e
t
N
or
m
a
l
56000
37000
G
e
ne
r
i
c
40000
18871
E
xpl
oi
t
s
33393
11132
F
uz
z
e
r
s
18184
6062
D
oS
12264
4089
R
e
c
onna
i
s
s
a
nc
e
10491
3496
A
na
l
ys
i
s
ba
c
kdoor
2000
677
B
a
c
kdoor
1746
583
S
he
l
l
C
ode
1133
378
W
or
m
s
130
44
T
ot
a
l
i
ns
t
a
nc
e
s
175,341
82,332
2.2. S
ys
t
e
m
m
od
e
li
n
g
2.2.1.
I
n
t
r
u
s
io
n
d
e
t
e
c
t
io
n
s
ys
t
e
m
s
a
r
c
h
it
e
c
t
u
r
e
C
r
e
a
ti
ng
a
ge
ne
r
a
l
m
ode
l
f
or
a
ll
I
D
S
w
it
h
a
f
ix
e
d
a
c
c
ur
a
c
y
is
not
a
vi
a
bl
e
s
ol
ut
io
n.
D
if
f
e
r
e
nt
us
e
r
s
ha
ve
di
f
f
e
r
e
nt
m
ode
ls
a
nd
th
e
m
ode
l
m
us
t
be
upda
te
d
on
th
e
f
ly
w
it
h
te
r
r
a
in
da
ta
.
H
ow
e
ve
r
,
th
e
tr
a
in
in
g
m
ode
l
c
ons
um
e
s
a
lo
t
of
r
e
s
our
c
e
s
,
w
hi
c
h
hi
nde
r
s
th
e
pr
om
ot
io
n
a
nd
pr
a
c
ti
c
a
li
ty
of
in
c
r
e
m
e
nt
a
l
m
ode
ls
on
r
e
s
our
c
e
-
c
ons
tr
a
in
e
d
de
vi
c
e
s
.
T
he
pr
opos
e
d
s
ys
t
e
m
a
ll
ow
s
c
ha
ngi
ng
th
e
m
ode
l
pa
r
a
m
e
te
r
s
,
a
s
s
how
n
in
F
ig
ur
e
1.
W
he
n
th
e
da
ta
s
tr
e
a
m
a
r
r
iv
e
s
,
S
L
D
A
c
a
n
in
f
e
r
e
a
c
h
u
pc
om
in
g
s
a
m
pl
e
a
nd
upda
te
it
s
pa
r
a
m
e
te
r
s
by
e
xpl
oi
ti
ng t
he
be
ne
f
it
s
of
onl
in
e
l
e
a
r
ni
ng.
F
ig
ur
e
1. T
he
bui
ld
in
g bl
oc
ks
of
our
I
D
S
s
ys
te
m
2.2.2. M
od
e
l
t
r
ai
n
in
g p
r
oc
e
s
s
I
ni
ti
a
ll
y,
w
e
s
tr
uc
t
ur
e
d
th
e
d
a
ta
b
a
s
e
i
nt
o
s
e
pa
r
a
te
tr
a
in
i
ng
a
nd
te
s
ti
ng
s
e
gm
e
nt
s
.
S
ub
s
e
q
ue
nt
l
y,
w
e
im
pl
e
m
e
nt
e
d
th
e
P
C
A
te
c
hni
q
ue
t
o
r
e
d
uc
e
di
m
e
ns
i
ona
l
it
y.
T
h
e
tr
a
in
in
g
pha
s
e
in
vol
ve
d
f
e
e
di
ng
our
m
od
e
l
s
w
it
h
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
N
e
tw
or
k
i
nt
r
us
io
n de
te
c
ti
on i
n bi
g datas
e
ts
us
in
g Spa
r
k
e
n
v
ir
on
m
e
nt
and …
(
A
bde
lw
ahe
d
E
lmoutaoukk
il
)
4417
da
ta
ba
t
c
he
s
,
m
im
i
c
ki
n
g
a
c
on
s
i
s
te
nt
da
t
a
s
tr
e
a
m
.
T
hi
s
a
ppr
oa
c
h
e
na
bl
e
d
u
s
to
m
oni
to
r
th
e
a
c
c
ur
a
c
y
pr
o
gr
e
s
s
io
n
pos
t
e
a
c
h t
r
a
in
in
g
it
e
r
a
t
io
n.
T
h
e
pr
op
os
e
d f
r
a
m
e
w
or
k
f
or
onl
i
ne
in
tr
us
io
n de
te
c
t
io
n i
s
s
how
n i
n t
he
F
i
gur
e
2
.
F
ig
ur
e
2. P
r
opos
e
d f
r
a
m
e
w
or
k f
or
in
tr
us
io
n de
te
c
ti
on
2.3. S
t
r
e
am
in
g l
in
e
ar
d
is
c
r
im
in
an
t
an
al
ys
is
al
gor
it
h
m
L
in
e
a
r
di
s
c
r
im
in
a
nt
a
na
ly
s
is
(
L
D
A
)
is
a
w
id
e
ly
e
m
br
a
c
e
d
m
e
th
o
d
in
th
e
f
ie
ld
s
of
c
om
put
e
r
vi
s
io
n
a
nd
pa
tt
e
r
n
r
e
c
ogni
ti
on
due
to
it
s
ve
r
s
a
ti
li
ty
in
bot
h
di
m
e
ns
io
na
li
ty
r
e
duc
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
ta
s
ks
.
N
um
e
r
ous
s
tu
di
e
s
[
18]
,
[
19]
ha
ve
de
m
ons
tr
a
te
d i
ts
e
f
f
e
c
ti
ve
ne
s
s
. L
D
A
i
nvol
ve
s
s
e
e
ki
ng a
l
in
e
a
r
da
ta
t
r
a
ns
f
or
m
a
ti
on t
ha
t
opt
im
a
ll
y s
e
pa
r
a
te
s
c
la
s
s
e
s
w
hi
le
r
e
duc
in
g t
he
di
m
e
n
s
io
na
li
ty
o
f
t
he
da
ta
[
20]
.
T
he
ty
pi
c
a
l
im
pl
e
m
e
nt
a
ti
on
of
th
e
L
D
A
te
c
hni
que
r
e
qui
r
e
s
th
a
t
a
ll
s
a
m
pl
e
s
be
a
va
il
a
bl
e
in
a
dv
a
nc
e
[
21]
.
H
ow
e
ve
r
,
th
e
r
e
a
r
e
s
it
ua
ti
on
s
w
he
r
e
th
e
c
om
pl
e
te
d
a
ta
s
e
t
is
not
a
va
il
a
bl
e
a
nd
th
e
in
put
d
a
ta
i
s
obs
e
r
ve
d
a
s
a
f
lo
w
.
I
n
th
is
c
a
s
e
,
th
e
L
D
A
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
s
houl
d
ha
ve
t
he
a
bi
li
ty
to
upda
te
th
e
c
om
put
e
d
L
D
A
f
e
a
tu
r
e
s
by
obs
e
r
vi
ng
th
e
ne
w
s
a
m
pl
e
s
w
it
hout
r
unni
ng
th
e
a
lg
or
it
hm
on
th
e
e
nt
ir
e
da
ta
s
e
t.
F
or
e
xa
m
pl
e
,
in
m
a
ny
r
e
a
l
-
ti
m
e
a
ppl
ic
a
ti
ons
s
uc
h
a
s
m
obi
le
r
obot
ic
s
,
onl
in
e
f
a
c
i
a
l
r
e
c
ogni
ti
on,
a
nd
I
o
T
n
e
twor
ks
,
it
i
s
im
por
ta
nt
to
upda
te
t
he
e
xt
r
a
c
te
d L
D
A
f
e
a
tu
r
e
s
a
s
s
oon
a
s
ne
w
ob
s
e
r
va
ti
ons
a
r
e
a
va
il
a
bl
e
.
S
im
pl
y
obs
e
r
vi
ng
ne
w
s
a
m
pl
e
s
is
a
s
tr
e
a
m
in
g
L
D
A
a
lg
or
it
hm
,
a
nd
th
is
id
e
a
ha
s
be
e
n
w
id
e
ly
s
tu
di
e
d
ove
r
th
e
pa
s
t
two
de
c
a
de
s
.
C
ha
tt
e
r
je
e
a
nd
R
oyc
how
dhur
y
[
22]
pr
opos
e
d
a
n
in
c
r
e
m
e
nt
a
l
s
e
lf
-
or
ga
ni
z
in
g
L
D
A
a
lg
or
it
hm
to
upda
te
L
D
A
f
e
a
tu
r
e
s
.
D
e
m
ir
a
nd
O
z
m
e
hm
e
t
[
19]
pr
opos
e
d
lo
c
a
l
onl
in
e
le
a
r
ni
ng
a
lg
or
it
hm
s
t
o
upda
te
L
D
A
f
e
a
tu
r
e
s
i
nc
r
e
m
e
nt
a
ll
y us
in
g e
r
r
or
c
or
r
e
c
ti
on a
nd
H
e
bbi
a
n l
e
a
r
ni
ng r
ul
e
s
. L
a
te
r
,
G
ha
s
s
a
be
h
et
al
.
[
21]
d
e
r
iv
e
d
f
a
s
t
in
c
r
e
m
e
nt
a
l
a
lg
or
it
hm
s
to
upda
te
L
D
A
f
unc
ti
ona
li
ty
by
obs
e
r
vi
ng
ne
w
s
a
m
pl
e
s
.
L
e
t
be
th
e
f
ol
lo
w
in
g pr
e
di
c
ti
on e
qua
ti
on:
y
t
=
W
.z
t
+b
(
1)
W
he
r
e
z
t
ϵ
I
R
d
is
a
ve
c
to
r
,
W
z
t
ϵ
I
R
k*d
a
nd
b
ϵ
I
R
k
a
r
e
upda
te
d
onl
in
e
pa
r
a
m
e
te
r
s
,
a
nd
k
is
th
e
to
ta
l
num
be
r
of
c
la
s
s
e
s
.
S
L
D
A
s
to
r
e
s
one
m
e
a
n
ve
c
to
r
pe
r
c
la
s
s
μ
k
ϵ
I
R
d
w
it
h
a
n
a
s
s
o
c
ia
te
d
num
be
r
c
k
ϵ
I
R
a
nd
a
s
in
gl
e
m
a
tr
ix
of
s
ha
r
e
d
c
ova
r
ia
nc
e
Ʃ
ϵ
R
d×
d.
W
he
n
a
ne
w
da
ta
poi
nt
(
z
t,
y)
a
r
r
iv
e
s
,
th
e
a
ve
r
a
ge
ve
c
to
r
a
nd
th
e
a
s
s
oc
ia
t
e
d c
ount
e
r
a
r
e
upda
te
d a
s
:
(
=
,
+
1
)
=
(
=
,
)
+
c
(
=
,
)
+
1
(
2)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
, V
ol
. 13, No. 4, D
e
c
e
m
be
r
2024:
4414
-
4421
4418
(
=
,
+
1
)
=
c
(
=
,
)
+
1
(
3)
w
it
h
a
s
th
e
m
e
a
n
of
th
e
c
la
s
s
y
h
a
s
th
e
m
om
e
nt
t
a
nd
c
(
=
,
)
is
th
e
a
s
s
oc
ia
te
d
c
ount
e
r
.
F
or
S
L
D
A
w
it
h
a
n
onl
in
e
va
r
ia
bl
e
va
r
ia
nc
e
, w
e
u
s
e
t
he
f
ol
lo
w
in
g upda
te
:
Ʃ
+
1
=
Ʃ
+
∆
+
1
(
4)
∆
t
is
c
a
lc
ul
a
t
e
d a
s
:
∆
=
t
(
−
(
=
,
)
)
(
−
(
=
,
)
)
+
1
(
5)
W
e
us
e
(
1)
a
nd W
k
to
c
a
lc
ul
a
te
t
he
pr
e
di
c
ti
on.
T
he
c
ol
um
n
s
of
W
i
s
gi
ve
n a
s
:
W
k
=
Λ µ
k
(
6)
w
it
h Λ a
r
e
de
te
r
m
in
e
d by:
Λ=
[
(
1
-
ε)
Ʃ
+
ε
.
I
]
-
1
(
7)
b
k
is
upda
te
d a
s
:
=
−
1
2
(
µ
.
Λ
µ
)
(
8)
2.4.
B
ig
d
at
a p
r
oc
e
s
s
in
g t
ool
s
:
A
p
ac
h
e
S
p
ar
k
T
he
A
pa
c
h
e
S
pa
r
k
e
nvi
r
onm
e
nt
is
a
r
obus
t
to
ol
in
th
e
r
e
a
lm
o
f
bi
g
da
ta
,
of
f
e
r
in
g
a
n
ope
n
-
s
our
c
e
di
s
tr
ib
ut
e
d c
lu
s
te
r
-
c
om
put
in
g f
r
a
m
e
w
or
k. I
t
boa
s
ts
a
c
om
pr
e
he
ns
iv
e
ML
li
br
a
r
y known a
s
M
L
li
b, s
pe
c
if
ic
a
ll
y
de
s
ig
ne
d
f
or
M
L
c
la
s
s
if
ie
r
s
.
A
pa
c
he
S
pa
r
k e
xc
e
ls
in
in
-
m
e
m
or
y
pr
oc
e
s
s
in
g
a
nd
pr
ovi
de
s
s
uppor
t
f
or
m
ul
ti
pl
e
pr
ogr
a
m
m
in
g
la
ngua
ge
s
,
in
c
lu
di
ng
J
a
va
,
S
c
a
la
,
P
yt
hon,
S
Q
L
,
a
nd
R
,
w
it
h
pa
r
ti
c
ul
a
r
ly
s
tr
ong
c
om
pa
ti
bi
li
ty
w
it
h P
yt
hon.
2.
4.
1. C
om
p
on
e
n
t
s
of
t
h
e
S
p
ar
k
e
c
os
y
s
t
e
m
:
T
he
S
pa
r
k
e
c
os
ys
t
e
m
c
om
pr
is
e
s
s
e
v
e
r
a
l
ke
y
c
om
pone
nt
s
,
in
c
l
udi
ng
S
pa
r
k
c
or
e
c
om
pone
nt
,
S
pa
r
k
S
Q
L
,
S
pa
r
k
s
tr
e
a
m
in
g
,
S
pa
r
k
M
L
li
b,
S
pa
r
k
G
r
a
phX
,
a
nd
S
pa
r
kR
.
E
a
c
h
c
om
pone
nt
s
e
r
ve
s
a
di
s
ti
nc
t
pur
pos
e
,
f
a
c
il
it
a
ti
ng
di
f
f
e
r
e
nt
a
s
pe
c
ts
of
bi
g
da
ta
pr
oc
e
s
s
in
g
a
nd
a
na
ly
ti
c
s
.
I
n
th
is
s
pe
c
if
ic
w
or
k,
w
e
f
oc
us
on
us
in
g
S
pa
r
k
M
L
li
b,
a
s
c
a
la
bl
e
ML
li
br
a
r
y
th
a
t
e
nc
om
pa
s
s
e
s
va
r
io
us
M
L
a
lg
or
it
hm
s
.
S
pa
r
k
M
L
li
b
'
s
c
a
pa
bi
li
ti
e
s
e
na
bl
e
e
f
f
ic
ie
nt
im
pl
e
m
e
nt
a
ti
on
a
nd
e
xe
c
ut
io
n
of
ML
ta
s
k
s
on
la
r
ge
d
a
ta
s
e
ts
,
hi
ghl
ig
ht
in
g
it
s
s
ig
ni
f
ic
a
nc
e
w
it
hi
n
th
e
S
pa
r
k e
c
os
ys
te
m
f
or
a
dva
nc
e
d d
a
ta
a
na
ly
s
is
a
nd pr
e
di
c
ti
ve
m
od
e
li
ng
[
23]
, [
24]
.
2.
4
.
2.
F
e
at
u
r
e
s
of
A
p
ac
h
e
S
p
ar
k
:
S
om
e
not
a
bl
e
f
e
a
tu
r
e
s
of
A
pa
c
he
S
pa
r
k i
nc
lu
de
[
25]
:
−
R
a
pi
d
pr
oc
e
s
s
in
g
:
A
pa
c
he
S
pa
r
k
boa
s
ts
hi
gh
-
s
pe
e
d
d
a
ta
pr
oc
e
s
s
in
g
c
a
p
a
bi
li
ti
e
s
,
pe
r
f
or
m
in
g
a
ppr
oxi
m
a
te
ly
100 ti
m
e
s
f
a
s
te
r
i
n m
e
m
or
y a
nd 10 ti
m
e
s
f
a
s
te
r
on dis
k c
om
pa
r
e
d t
o t
r
a
di
ti
ona
l
m
e
th
ods
.
−
D
yna
m
ic
:
it
f
a
c
il
it
a
te
s
th
e
de
ve
lo
pm
e
nt
of
pa
r
a
ll
e
l
a
ppl
ic
a
ti
ons
w
it
hi
n
S
pa
r
k,
th
a
nks
to
th
e
a
va
il
a
bi
li
ty
o
f
a
r
ound 80 high
-
le
ve
l
ope
r
a
to
r
s
.
−
In
-
m
e
m
or
y
c
om
put
a
ti
on
:
A
pa
c
he
S
pa
r
k
s
uppor
ts
in
-
m
e
m
or
y
c
o
m
put
a
ti
on,
e
na
bl
in
g
e
nha
nc
e
d
pr
oc
e
s
s
in
g
s
pe
e
ds
by k
e
e
pi
ng da
ta
i
n m
e
m
or
y f
or
qui
c
ke
r
a
c
c
e
s
s
a
nd a
n
a
ly
s
is
.
O
ve
r
a
ll
,
A
pa
c
he
S
pa
r
k
of
f
e
r
s
a
pow
e
r
f
ul
a
nd
ve
r
s
a
ti
le
e
nvi
r
onm
e
nt
f
or
pr
oc
e
s
s
in
g
la
r
ge
-
s
c
a
le
da
ta
a
nd
pe
r
f
or
m
in
g
ML
ta
s
ks
e
f
f
ic
ie
nt
ly
.
2.5
. P
e
r
f
or
m
an
c
e
o
f
m
a
c
h
in
e
l
e
ar
n
in
g
m
od
e
l
T
he
pe
r
f
or
m
a
nc
e
of
a
M
L
m
od
e
l
is
a
c
r
it
ic
a
l
a
s
pe
c
t
th
a
t
de
t
e
r
m
in
e
s
i
ts
e
f
f
e
c
t
iv
e
n
e
s
s
in
m
a
ki
n
g
a
c
c
ur
a
te
pr
e
di
c
ti
on
s
.
E
va
lu
a
ti
ng
th
e
pe
r
f
or
m
a
nc
e
in
v
ol
ve
s
a
s
e
t
of
m
e
tr
i
c
s
th
a
t
pr
ovi
d
e
in
s
ig
ht
s
in
to
how
w
e
ll
th
e
m
o
de
l
c
a
n
c
l
a
s
s
if
y
or
pr
e
di
c
t
da
t
a
po
in
ts
. K
e
y
m
e
a
s
ur
e
s
i
nc
l
ude
a
c
c
ur
a
c
y, w
hi
c
h
i
ndi
c
a
t
e
s
t
he
pr
opor
ti
on of
c
or
r
e
c
t
pr
e
di
c
ti
on
s
;
s
e
n
s
it
iv
it
y
(
or
r
e
c
a
ll
)
,
w
h
ic
h
m
e
a
s
ur
e
s
th
e
a
bi
li
ty
to
id
e
nt
if
y
tr
ue
pos
it
iv
e
s
;
s
pe
c
if
ic
it
y,
w
hi
c
h
a
s
s
e
s
s
e
s
th
e
id
e
nt
if
i
c
a
ti
on
of
tr
ue
ne
g
a
ti
v
e
s
;
pr
e
c
i
s
io
n
,
w
hi
c
h
e
v
a
lu
a
te
s
th
e
a
c
c
ur
a
c
y
of
p
os
it
i
ve
pr
e
di
c
t
io
n
s
;
a
nd
th
e
F
1
s
c
or
e
,
w
hi
c
h
ba
l
a
nc
e
s
pr
e
c
i
s
io
n
a
nd
r
e
c
a
ll
.
T
he
s
e
m
e
tr
ic
s
c
ol
l
e
c
ti
ve
ly
of
f
e
r
a
c
om
pr
e
h
e
ns
iv
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
N
e
tw
or
k
i
nt
r
us
io
n de
te
c
ti
on i
n bi
g datas
e
ts
us
in
g Spa
r
k
e
n
v
ir
on
m
e
nt
and …
(
A
bde
lw
ahe
d
E
lmoutaoukk
il
)
4419
unde
r
s
ta
nd
in
g
of
t
he
m
od
e
l’
s
s
tr
e
ngt
hs
a
n
d
w
e
a
kne
s
s
e
s
,
gui
di
ng
im
pr
ove
m
e
nt
s
a
nd
e
ns
ur
i
ng
th
e
m
od
e
l'
s
r
e
li
a
bi
li
t
y i
n
pr
a
c
t
ic
a
l
a
p
pl
ic
a
ti
on
s
.
‒
T
o
e
va
lu
a
te
th
e
a
c
c
ur
a
c
y
or
s
upe
r
io
r
it
y
of
our
c
la
s
s
if
ie
r
s
in
pr
e
di
c
ti
ng
th
e
c
la
s
s
la
b
e
ls
of
tu
pl
e
s
,
w
e
us
e
s
e
ve
r
a
l
ke
y
m
e
tr
ic
s
.
T
he
s
e
c
la
s
s
if
ie
r
e
va
lu
a
ti
on
m
e
a
s
ur
e
s
in
c
lu
de
a
c
c
ur
a
c
y,
s
e
ns
it
iv
it
y
(
or
r
e
c
a
ll
)
,
s
pe
c
if
ic
it
y,
pr
e
c
is
io
n,
a
nd
F
1
s
c
or
e
.
T
he
a
c
c
ur
a
c
y
of
a
c
la
s
s
if
ie
r
on
a
gi
ve
n
te
s
t
s
e
t
is
de
te
r
m
in
e
d
by
th
e
pe
r
c
e
nt
a
ge
of
t
r
ue
pos
it
iv
e
s
a
nd t
r
ue
ne
g
a
ti
ve
s
out
of
a
ll
c
or
r
e
c
t
ly
c
la
s
s
if
ie
d i
ns
ta
n
c
e
s
.
=
+
+
+
+
(
9)
H
e
r
e
, T
P
is
t
r
ue
pos
it
iv
e
,
T
N
is
tr
ue
ne
ga
ti
ve
, F
P
is
f
a
ls
e
pos
it
iv
e
,
a
nd
FN
is
f
a
ls
e
ne
g
a
ti
ve
.
‒
T
r
a
in
in
g
time
:
m
e
a
s
ur
e
m
e
nt
of
t
he
t
im
e
t
a
ke
n by the
a
lg
or
it
hm
to
t
r
a
in
a
ne
w
ba
tc
h
.
‒
P
r
e
di
c
ti
on
time
:
m
e
a
s
ur
e
m
e
nt
of
t
im
e
t
a
ke
n by a
c
la
s
s
if
ie
r
to
pr
e
di
c
t
ne
w
da
ta
.
3.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
T
a
bl
e
2
a
nd F
ig
ur
e
3 i
ll
u
s
tr
a
t
e
s
t
he
e
vol
ut
io
n of
t
h
e
a
c
c
ur
a
c
y of
a
m
a
c
hi
ne
-
l
e
a
r
ni
ng mod
e
l
a
c
r
os
s
12
tr
a
in
in
g
ba
t
c
h
e
s
.
T
he
in
it
i
a
l
a
c
c
ur
a
c
y
s
t
a
nd
s
a
t
9
.85
%
a
f
te
r
th
e
f
ir
s
t
b
a
tc
h.
S
u
bs
e
qu
e
nt
ly
,
a
s
th
e
m
od
e
l
i
s
e
xpo
s
e
d
to
m
or
e
tr
a
in
in
g
da
ta
a
nd
l
e
a
r
ni
n
g
it
e
r
a
ti
o
ns
,
it
s
a
c
c
ur
a
c
y
gr
a
dua
ll
y
im
pr
ov
e
s
th
r
ou
ghout
th
e
s
e
ba
tc
h
e
s
.
I
t
be
c
om
e
s
e
vi
de
nt
t
ha
t
t
he
m
ode
l
und
e
r
goe
s
s
ig
n
if
ic
a
nt
i
m
pr
o
ve
m
e
nt
.
T
he
m
o
s
t
r
e
m
a
r
k
a
bl
e
a
s
p
e
c
t
of
th
i
s
tr
a
in
i
ng
pr
o
c
e
s
s
i
s
t
he
s
ub
s
t
a
nt
ia
l
le
a
p
in
a
c
c
ur
a
c
y,
c
ul
m
in
a
ti
ng
in
a
n
im
pr
e
s
s
iv
e
r
a
t
e
of
9
6.37
%
a
f
te
r
t
he
l
a
s
t
(
th
e
12
th
)
b
a
tc
h
.
T
h
is
in
di
c
a
t
e
s
th
a
t
th
e
m
ode
l
ha
s
m
a
de
s
ig
ni
f
ic
a
nt
pr
ogr
e
s
s
in
it
s
a
bi
li
t
y
to
m
a
k
e
a
c
c
ur
a
t
e
pr
e
di
c
ti
on
s
a
s
it
le
a
r
n
s
f
r
om
th
e
da
t
a
.
W
he
n
w
e
c
om
pa
r
e
th
e
r
e
s
ul
ts
of
our
a
ppr
oa
c
he
to
pr
e
vi
ous
s
tu
di
e
s
ba
s
e
d
on
tr
a
di
ti
ona
l
le
a
r
ni
ng
te
c
hni
que
s
,
I
D
S
ba
s
e
d
on
th
e
pr
ogr
e
s
s
iv
e
le
a
r
ni
ng
m
ode
l
s
how
e
d
s
upe
r
io
r
a
da
pt
a
bi
li
ty
a
nd
e
f
f
ic
ie
nc
y.
N
ot
a
bl
y,
th
e
s
te
p
-
by
-
s
te
p
a
ppr
oa
c
h
a
c
hi
e
ve
d
s
im
il
a
r
or
hi
ghe
r
a
c
c
ur
a
c
y l
e
ve
ls
w
hi
le
s
ig
ni
f
ic
a
nt
ly
r
e
duc
in
g c
om
put
a
ti
ona
l
r
e
s
our
c
e
s
a
nd t
r
a
in
in
g t
im
e
.
T
a
bl
e
2
.
P
e
r
f
or
m
a
nc
e
e
va
lu
a
ti
on us
in
g
S
L
D
A
on da
ta
s
e
t
–
1 a
nd
da
ta
s
e
t
–
2
of
(
U
N
S
W
N
B
-
15 da
ta
s
e
t)
D
ur
a
t
i
on of
t
r
a
i
ni
ng (
S
e
c
)
D
ur
a
t
i
on of
pr
e
di
c
t
i
on (
S
e
c
)
A
c
c
ur
a
c
y (
%
)
B
a
t
c
h 1
0.0779
0.0229
9.85
B
a
t
c
h 2
0.0929
0.0109
18.88
B
a
t
c
h 3
0.0829
0.0099
28.18
B
a
t
c
h 4
0.1159
0.0089
37.01
B
a
t
c
h 5
0.0849
0.0079
49.39
B
a
t
c
h 6
0.0829
0.0069
52.60
B
a
t
c
h 7
0.0849
0.0059
56.63
B
a
t
c
h 8
0.0829
0.0049
62.40
B
a
t
c
h 10
0.0969
0.0059
69.48
B
a
t
c
h 11
0.0889
0.00299
7
76.28
B
a
t
c
h 12
0.0879
0.002998
96.37
F
ig
ur
e
3. I
D
S
pe
r
f
o
r
m
a
nc
e
on UN
S
W
-
N
B
15 t
e
s
t
a
f
te
r
e
a
c
h t
r
a
i
ni
ng s
e
s
s
io
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
, V
ol
. 13, No. 4, D
e
c
e
m
be
r
2024:
4414
-
4421
4420
4.
C
O
N
C
L
U
S
I
O
N
T
he
pa
pe
r
pr
opos
e
d
a
f
r
a
m
e
w
or
k
th
a
t
w
a
s
f
a
s
t
a
nd
e
f
f
e
c
ti
ve
f
o
r
in
tr
us
io
n
de
te
c
ti
on.
W
e
ha
ve
us
e
d
th
e
U
N
S
W
N
B
-
15
da
ta
s
e
t
f
or
pe
r
f
or
m
a
nc
e
e
va
lu
a
ti
on
of
th
e
pr
opos
e
d
f
r
a
m
e
w
or
k
by
a
ppl
yi
ng
f
e
a
tu
r
e
r
e
duc
ti
on
us
in
g
P
C
A
a
nd
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
S
L
D
A
.
I
t
is
f
ound
th
a
t
us
in
g
S
L
D
A
is
m
or
e
e
f
f
ic
ie
nt
c
om
pa
r
e
d
to
a
ut
hor
s
ta
ti
c
a
lg
or
it
hm
s
e
s
pe
c
ia
ll
y
w
he
n
th
e
in
put
da
ta
is
obs
e
r
ve
d
a
s
a
f
lo
w
a
s
is
th
e
c
a
s
e
in
our
s
ubj
e
c
t
w
he
r
e
t
he
pr
e
c
is
io
n
of
th
e
m
ode
l
im
pr
ove
s
dur
in
g
th
e
a
r
r
iv
a
l
of
th
e
n
e
w
ba
tc
h
of
da
ta
.
I
t
c
a
n
be
c
onc
lu
de
d
th
a
t
th
is
a
ppr
oa
c
h i
s
be
tt
e
r
, f
a
s
te
r
, a
nd mor
e
e
f
f
ic
ie
nt
w
he
n us
e
d on Apa
c
he
S
pa
r
k.
5.
R
E
C
O
M
M
E
N
D
A
T
I
O
N
S
F
O
R
F
U
T
U
R
E
WORK
S
e
ve
r
a
l
s
ugge
s
ti
ons
f
or
f
ut
ur
e
r
e
s
e
a
r
c
h
e
nde
a
vor
s
c
oul
d
be
e
xpl
or
e
d
to
e
xpa
nd
upon
th
is
s
tu
dy.
T
he
s
e
a
ddi
ti
ona
l
r
e
c
om
m
e
nda
ti
ons
a
r
e
out
li
ne
d a
s
f
ol
lo
w
s
:
th
e
pr
opos
e
d a
lg
or
it
hm
ha
s
t
he
pot
e
nt
ia
l
to
be
f
in
e
-
tu
ne
d
a
nd
ut
il
iz
e
d
f
or
va
r
io
us
r
e
a
l
-
w
or
ld
a
ppl
ic
a
ti
ons
th
a
t
ne
c
e
s
s
it
a
te
im
a
ge
r
e
c
ogni
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on.
T
hi
s
in
c
lu
de
s
dom
a
in
s
s
uc
h
a
s
m
e
di
c
a
l
im
a
gi
ng,
bi
om
e
di
c
a
l
a
na
ly
s
is
,
a
nd
ha
ndw
r
it
in
g
r
e
c
ogni
ti
on
a
ppl
ic
a
ti
ons
.
T
he
pr
opos
e
d s
ys
te
m
c
oul
d be
i
nt
e
gr
a
te
d i
nt
o a
n I
oT
de
vi
c
e
t
o of
f
e
r
i
nt
r
us
io
n de
te
c
ti
on
s
e
r
vi
c
e
s
f
or
I
oT
ne
twor
ks
.
F
ur
th
e
r
in
ve
s
ti
ga
ti
on
in
to
th
is
pr
opos
e
d
I
D
S
c
oul
d
d
e
lv
e
in
to
a
s
pe
c
ts
s
uc
h
a
s
pow
e
r
c
on
s
um
pt
io
n,
m
e
m
or
y
ut
il
iz
a
ti
on,
c
om
m
uni
c
a
ti
on
pr
ot
oc
ol
s
,
a
nd
c
om
put
a
ti
ona
l
c
om
pl
e
xi
ty
,
pa
r
ti
c
ul
a
r
ly
w
he
n
im
pl
e
m
e
nt
e
d
on
lo
w
-
pow
e
r
I
oT
node
s
w
it
h
s
m
a
ll
-
s
c
a
le
s
y
s
te
m
c
om
pone
nt
s
,
s
uc
h
a
s
ba
tt
e
r
y
-
ope
r
a
te
d
or
e
ne
r
gy
-
e
f
f
ic
ie
nt
de
vi
c
e
s
.
R
E
F
E
R
E
N
C
E
S
[
1]
K
.
A
l
bul
a
yhi
,
Q
.
A
.
A
.
-
H
a
i
j
a
,
S
.
A
.
A
l
s
uhi
ba
ny,
A
.
A
.
J
i
l
l
e
pa
l
l
i
,
M
.
A
s
hr
a
f
uz
z
a
m
a
n,
a
nd
F
.
T
.
S
he
l
don,
“
I
oT
i
nt
r
us
i
on
de
t
e
c
t
i
on
us
i
ng m
a
c
hi
ne
l
e
a
r
ni
ng w
i
t
h a
nove
l
hi
gh pe
r
f
or
m
i
ng f
e
a
t
ur
e
s
e
l
e
c
t
i
on m
e
t
hod,
”
A
ppl
i
e
d Sc
i
e
nc
e
s
, vol
. 12,
no. 10,
M
a
y 2022,
doi
:
10.3390/
a
pp12105015.
[
2]
R
. H
e
a
dy, G
. L
uge
r
, A
. M
a
c
c
a
be
, a
nd M
. S
e
r
vi
l
l
a
, “
T
he
a
r
c
hi
t
e
c
t
ur
e
of
a
ne
t
w
or
k l
e
ve
l
i
nt
r
us
i
on de
t
e
c
t
i
on s
ys
t
e
m
,”
D
e
p
a
r
t
m
e
nt
of
C
om
put
e
r
S
c
i
e
nc
e
,
U
ni
ve
r
s
i
t
y of
N
e
w
M
e
xi
c
o
,
A
l
buque
r
que
, M
e
xi
c
o,
A
ug. 19
90, doi
:
10.2172/
425295.
[
3]
S
. A
xe
l
s
s
on, “
I
nt
r
us
i
on
de
t
e
c
t
i
on s
y
s
t
e
m
s
:
a
s
ur
ve
y a
nd t
a
xonom
y
,”
C
i
t
e
S
e
e
r
X
,
pp. 1
–
27, 2000, doi
:
10.1.1.1.6603.
[
4]
G
.
V
i
gna
a
nd R
.
A
.
K
e
m
m
e
r
e
r
,
“
N
e
t
S
T
A
T
:
A
ne
t
w
or
k
-
ba
s
e
d
i
nt
r
us
i
on
de
t
e
c
t
i
o
n
s
ys
t
e
m
,”
J
our
nal
of
C
om
put
e
r
Se
c
u
r
i
t
y
,
vol
.
7,
no
.
1, pp. 37
–
71, J
a
n. 1999, doi
:
10.3233/
J
C
S
-
1999
-
7103.
[
5]
M
.
A
z
ha
r
i
,
A
.
A
b
a
r
da
,
B
.
E
t
t
a
ki
,
J
.
Z
e
r
oua
oui
,
a
nd
M
.
D
a
kkon,
“
H
i
ggs
bo
s
o
n
di
s
c
ove
r
y
us
i
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
m
e
t
hod
s
w
i
t
h
P
ys
pa
r
k,”
P
r
oc
e
di
a C
om
put
e
r
Sc
i
e
nc
e
, vol
. 170, pp. 1141
–
1146, 2020, doi
:
10.
1016/
j
.pr
oc
s
.2020.03.053.
[
6]
S
. A
gr
a
w
a
l
a
nd
J
. A
g
r
a
w
a
l
, “
S
ur
ve
y on
a
nom
a
l
y de
t
e
c
t
i
on us
i
ng
da
t
a
m
i
ni
ng
t
e
c
hni
que
s
,”
P
r
oc
e
di
a C
om
put
e
r
Sc
i
e
nc
e
, vol
. 60, pp.
708
–
713, 2015, doi
:
10.1016/
j
.pr
oc
s
.2015.08.220.
[
7]
A
. L
. B
uc
z
a
k a
nd E
. G
uve
n, “
A
s
ur
ve
y of
da
t
a
m
i
ni
ng a
nd m
a
c
hi
ne
l
e
a
r
ni
ng m
e
t
hods
f
or
c
ybe
r
s
e
c
ur
i
t
y i
nt
r
us
i
on de
t
e
c
t
i
on
,”
I
E
E
E
C
om
m
uni
c
at
i
ons
Sur
v
e
y
s
&
T
ut
or
i
al
s
, vol
. 18, no. 2, pp. 1153
–
1176, 2016, doi
:
10.1109/
C
O
M
S
T
.2015.2494502.
[
8]
Q
.
A
.
A
.
-
H
a
i
j
a
,
E
.
S
a
l
e
h,
a
nd
M
.
A
l
na
bh
a
n,
“
D
e
t
e
c
t
i
ng
por
t
s
c
a
n
a
t
t
a
c
ks
us
i
ng
l
ogi
s
t
i
c
r
e
gr
e
s
s
i
on
,”
i
n
2021
4t
h
I
nt
e
r
nat
i
onal
Sy
m
pos
i
um
on
A
dv
anc
e
d
E
l
e
c
t
r
i
c
al
and
C
om
m
uni
c
at
i
on
T
e
c
hnol
ogi
e
s
(
I
SA
E
C
T
)
,
D
e
c
.
2021,
pp.
1
–
5,
doi
:
10.1109/
I
S
A
E
C
T
53699.2021.9668562.
[
9]
Q
.
A
.
A
.
-
H
a
i
j
a
,
C
.
D
.
M
c
C
ur
r
y,
a
nd
S
.
Z
.
-
S
a
ba
t
t
o,
“
I
nt
e
l
l
i
ge
nt
s
e
l
f
-
r
e
l
i
a
nt
c
ybe
r
-
a
t
t
a
c
ks
de
t
e
c
t
i
on
a
nd
c
l
a
s
s
i
f
i
c
a
t
i
on
s
ys
t
e
m
f
or
I
oT
c
om
m
uni
c
a
t
i
on
us
i
ng
de
e
p
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
k
,”
Se
l
e
c
t
e
d
P
ape
r
s
f
r
om
t
he
12t
h
I
nt
e
r
nat
i
onal
N
e
t
w
o
r
k
i
ng
C
onf
e
r
e
nc
e
,
2021, pp. 100
–
116, doi
:
10.1007/
978
-
3
-
030
-
64758
-
2_8.
[
10]
K
.
A
.
D
ha
nya
,
S
.
V
a
j
i
pa
ya
j
ul
a
,
K
.
S
r
i
ni
va
s
a
n,
A
.
T
i
br
e
w
a
l
,
T
.
S
.
K
um
a
r
,
a
nd
T
.
G
.
K
um
a
r
,
“
D
e
t
e
c
t
i
on
of
ne
t
w
or
k
a
t
t
a
c
ks
us
i
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
a
nd
de
e
p
l
e
a
r
ni
ng
m
ode
l
s
,”
P
r
oc
e
di
a
C
om
put
e
r
Sc
i
e
nc
e
,
vol
.
218,
pp.
57
–
66,
2023,
doi
:
10.1016/
j
.pr
oc
s
.2022.12.401.
[
11]
A
. A
. A
l
s
ul
a
m
i
, Q
. A
. A
.
-
H
a
i
j
a
, A
. T
a
ye
b, a
nd A
. A
l
qa
ht
a
ni
, “
A
n
i
nt
r
us
i
on de
t
e
c
t
i
on a
nd c
l
a
s
s
i
f
i
c
a
t
i
on s
y
s
t
e
m
f
or
I
oT
t
r
a
f
f
i
c
w
i
t
h
i
m
pr
ove
d da
t
a
e
ngi
ne
e
r
i
ng
,”
A
ppl
i
e
d Sc
i
e
nc
e
s
, vol
. 12, no. 23, D
e
c
. 2022, doi
:
10.3390/
a
pp122312336.
[
12]
Q
.
A
.
A
.
-
H
a
i
j
a
,
“
T
op
-
dow
n
m
a
c
hi
ne
l
e
a
r
ni
ng
-
ba
s
e
d
a
r
c
hi
t
e
c
t
ur
e
f
or
c
yb
e
r
a
t
t
a
c
ks
i
de
nt
i
f
i
c
a
t
i
on
a
nd
c
l
a
s
s
i
f
i
c
a
t
i
on
i
n
I
oT
c
om
m
uni
c
a
t
i
on ne
t
w
or
ks
,”
F
r
ont
i
e
r
s
i
n B
i
g D
at
a
, vol
. 4, J
a
n. 2022, doi
:
10.3389/
f
da
t
a
.2021.782902.
[
13]
Q
.
A
.
A
.
-
H
a
i
j
a
,
M
.
K
r
i
c
he
n,
a
nd
W
.
A
.
E
l
ha
i
j
a
,
“
M
a
c
hi
ne
-
l
e
a
r
ni
ng
-
ba
s
e
d
da
r
kne
t
t
r
a
f
f
i
c
de
t
e
c
t
i
on
s
ys
t
e
m
f
or
I
o
T
a
ppl
i
c
a
t
i
ons
,”
E
l
e
c
t
r
oni
c
s
, vol
. 11, no. 4, F
e
b. 2022, doi
:
10.3390/
e
l
e
c
t
r
oni
c
s
11040556.
[
14]
M
.
S
ouha
i
l
e
t
.
al
.,
“
N
e
t
w
or
k
ba
s
e
d
i
nt
r
us
i
on
d
e
t
e
c
t
i
on
u
s
i
ng
t
he
U
N
S
W
-
N
B
15
da
t
a
s
e
t
,
”
I
nt
e
r
nat
i
onal
J
our
nal
of
C
om
put
i
ng
an
d
D
i
gi
t
al
Sy
s
t
e
m
s
, vol
. 8, no. 5, pp. 477
–
487, J
a
n. 2019, doi
:
10.12785/
i
j
c
ds
/
0805
05.
[
15]
M
.
T
a
va
l
l
a
e
e
,
E
.
B
a
ghe
r
i
,
W
. L
u,
a
nd
A
.
A
.
G
hor
ba
ni
,
“
A
d
e
t
a
i
l
e
d
a
na
l
ys
i
s
of
t
he
K
D
D
C
U
P
99
da
t
a
s
e
t
,”
i
n
2009
I
E
E
E
Sy
m
pos
i
um
on C
om
put
at
i
onal
I
nt
e
l
l
i
ge
nc
e
f
or
Se
c
ur
i
t
y
and D
e
f
e
ns
e
A
ppl
i
c
at
i
ons
, J
ul
. 2009, pp. 1
–
6, doi
:
10.1109/
C
I
S
D
A
.2009.5356528.
[
16]
N
.
M
ous
t
a
f
a
a
nd
J
.
S
l
a
y,
“
U
N
S
W
-
N
B
15:
a
c
om
pr
e
he
n
s
i
ve
da
t
a
s
e
t
f
or
ne
t
w
or
k
i
nt
r
us
i
on
de
t
e
c
t
i
on
s
y
s
t
e
m
s
(
U
N
S
W
-
N
B
15
ne
t
w
or
k
da
t
a
s
e
t
)
,”
i
n
2015
M
i
l
i
t
ar
y
C
om
m
uni
c
at
i
ons
and
I
nf
or
m
at
i
on
Sy
s
t
e
m
s
C
onf
e
r
e
nc
e
(
M
i
l
C
I
S)
,
N
ov.
2015,
pp.
1
–
6,
doi
:
10.1109/
M
i
l
C
I
S
.2015.7348942.
[
17]
A
.
R
.
S
onul
e
,
M
.
K
a
l
l
a
,
A
.
J
a
i
n,
a
nd
D
.
S
.
C
houh
a
n,
“
U
N
S
W
-
N
B
15
d
a
t
a
s
e
t
a
nd
m
a
c
hi
ne
l
e
a
r
ni
ng
ba
s
e
d
i
nt
r
us
i
on
de
t
e
c
t
i
o
n
s
ys
t
e
m
s
,
”
I
nt
e
r
nat
i
onal
J
our
nal
of
E
ngi
ne
e
r
i
ng
and
A
dv
anc
e
d
T
e
c
hnol
ogy
,
vol
.
9,
no.
3,
pp.
2638
–
2648,
F
e
b.
2020,
doi
:
10.35940/
i
j
e
a
t
.C
5809.029320.
[
18]
L
. J
i
n, K
. D
i
ng,
a
nd
Z
. H
ua
ng,
“
I
nc
r
e
m
e
nt
a
l
l
e
a
r
ni
ng of
L
D
A
m
od
e
l
f
or
C
hi
ne
s
e
w
r
i
t
e
r
a
da
pt
a
t
i
on,”
N
e
u
r
oc
om
put
i
ng
, vol
.
73, no
.
10
–
12, pp. 1614
–
1623, J
un. 2010, doi
:
10.1016/
j
.ne
uc
om
.2009.11.039.
[
19]
G
.
K
.
D
e
m
i
r
a
nd
K
.
O
z
m
e
hm
e
t
,
“
O
nl
i
ne
l
oc
a
l
l
e
a
r
ni
ng
a
l
gor
i
t
hm
s
f
or
l
i
ne
a
r
d
i
s
c
r
i
m
i
na
nt
a
na
l
ys
i
s
,”
P
at
t
e
r
n
R
e
c
ogni
t
i
on
L
e
t
t
e
r
s
,
vol
. 26, no. 4, pp. 421
–
431, M
a
r
. 2005, doi
:
10.1016/
j
.pa
t
r
e
c
.2004.08.005.
[
20]
H
. A
. M
ogh
a
dda
m
a
nd
K
. A
.
Z
a
de
h, “
F
a
s
t
a
da
pt
i
ve
a
l
gor
i
t
hm
s
a
nd
ne
t
w
or
ks
f
or
c
l
a
s
s
-
s
e
pa
r
a
bi
l
i
t
y f
e
a
t
ur
e
s
,”
P
at
t
e
r
n
R
e
c
ogni
t
i
on
,
vol
. 36, no. 8, pp. 1695
–
1702, A
ug. 2003, doi
:
10.1016/
S
0031
-
3203(
03
)
00006
-
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
N
e
tw
or
k
i
nt
r
us
io
n de
te
c
ti
on i
n bi
g datas
e
ts
us
in
g Spa
r
k
e
n
v
ir
on
m
e
nt
and …
(
A
bde
lw
ahe
d
E
lmoutaoukk
il
)
4421
[
21]
Y
. A
.
G
ha
s
s
a
be
h, F
. R
udz
i
c
z
, a
nd H
. A
. M
ogha
dda
m
, “
F
a
s
t
i
nc
r
e
m
e
nt
a
l
L
D
A
f
e
a
t
ur
e
e
xt
r
a
c
t
i
on,”
P
at
t
e
r
n
R
e
c
ogni
t
i
on
, vol
. 48,
no.
6, pp. 1999
–
2012, J
un. 2015, doi
:
10.1016/
j
.pa
t
c
og.2014.12.012.
[
22]
C
.
C
ha
t
t
e
r
j
e
e
a
nd
V
.
P
.
R
oyc
how
dhur
y,
“
O
n
s
e
l
f
-
or
ga
ni
z
i
ng
a
l
go
r
i
t
hm
s
a
n
d
ne
t
w
or
ks
f
or
c
l
a
s
s
-
s
e
pa
r
a
bi
l
i
t
y
f
e
a
t
ur
e
s
,”
I
E
E
E
T
r
ans
ac
t
i
ons
on N
e
ur
al
N
e
t
w
or
k
s
, vol
. 8, no. 3, pp. 663
–
678, M
a
y 1997, doi
:
1
0.1109/
72.572105.
[
23]
X
. M
e
ng
e
t
al
.
, “
M
l
l
i
b:
m
a
c
hi
ne
l
e
a
r
ni
ng i
n a
pa
c
he
s
p
a
r
k,”
J
our
nal
of
M
ac
hi
ne
L
e
ar
ni
ng R
e
s
e
ar
c
h
,
vol
. 17, no. 34, pp.
1
-
7
,
201
6.
[
24]
M
.
A
s
s
e
f
i
,
E
.
B
e
hr
a
ve
s
h,
G
.
L
i
u,
a
nd
A
.
P
.
T
a
f
t
i
,
“
B
i
g
da
t
a
m
a
c
hi
ne
l
e
a
r
ni
ng
us
i
n
g
a
pa
c
he
s
pa
r
k
M
L
l
i
b,”
i
n
2017
I
E
E
E
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on B
i
g D
at
a (
B
i
g D
at
a)
, D
e
c
. 2017, pp. 3492
–
3498, doi
:
10.1109/
B
i
gD
a
t
a
.2017.8258338.
[
25]
M
.
A
r
m
br
us
t
e
t
al
.
,
“
S
pa
r
k
S
Q
L
:
R
e
l
a
t
i
ona
l
da
t
a
pr
oc
e
s
s
i
ng
i
n
s
p
a
r
k,”
P
r
oc
e
e
di
ngs
of
t
he
A
C
M
SI
G
M
O
D
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on M
anage
m
e
nt
of
D
at
a
, vol
. 2015, pp. 1383
–
1394, 2015, doi
:
10.1145/
2723372.2742797.
B
I
O
G
R
A
P
H
I
E
S
O
F
A
U
T
H
O
R
S
Abdelwahed
Elmoutaoukkil
is
a
lecturer
in
electrical
engineering
at
BTS
Elkhawarizmi Casablanca
, Morocco. He
obtained his
Master'
s
degree
in Big Data and
the
Internet
of
Things
at
Hassan
2
University
,
Casablanca,
Morocco.
In
2
021,
h
e
has
been
a
computer
science
teacher
in
high
school
since
2008.
He
is
currently
a
w
riter
and
researcher
on
artificial
intelligence
applied
in
industry.
His
research
interes
ts
include
tinyML
technologie
s,
continuous
learning,
industrial
electroni
cs,
roboti
cs,
home
automation,
predictive
maintenance,
and
STEAM
models
in
the
field
of
educ
ation.
He
can
be
contacted
at email
:
elmoutao
ukkilab
d@
gmail.co
m.
Mohamed
Hamlich
is
the
Director
of
the
“Complex
Cyber
-
Physica
l
Systems”
Resear
ch
Labora
tory
at
ENSAM
Casablan
ca
(UH2C).
He
obtained
hi
s
doctora
l
thesis
in
Computer Sc
ience
from Ha
ssan II
Univer
sity in Casa
blanca
. His ar
eas
of res
earch
intere
st
include
robotics,
artificial
intell
igence
,
IoT,
and
big
data
.
He
is
th
e
president
of
the
Association
of
Connected
Objects
and
Intelligent
Systems
and
the
chair
of
The
International
Conference
“SADASC”.
He
is
the
editor
of
two
SPRIN
GER
books
and
the
author
of
several
papers
published
in
indexed
journals.
He
can
be
contacted
at
email:
moha.hamlich
@
gmail.com.
Amine
Khatib
is
a
scientific
researcher
at
Littoral
Côte
d'
Opale
U
niversity,
Franc
e.
He
obtaine
d
his
bache
lor
of
engine
ering
degre
e
in
auto
matic
at
Hassa
n
2
University,
Casablanca,
Morocco.
His
research
interests
include
roboti
cs
and
data
science
.
He can be contacted at email:
aminekhat
ib04@
gmail.co
m.
Marouan
e Chriss
is a teac
her of e
lectri
cal engine
ering at
BTS Elkh
awariz
mi
Casablan
ca, Mor
occo. I
n 2002 he
obtaine
d a
Diplôme d
'
Etudes Supéri
eures Approfondies
(
DESA)
in
Electrical
Engineering
from
ENSEM,
Hassan
2
Univ
ersity,
Casablanca,
Morocco.
He
is
currently
a
Ph
.
D
.
student
at
LCCPS
Laboratory
,
ENSAM,
Hassan
2
University.
His
research
interests
include
artificial
intelligence,
rob
otics
,
and
systems
control.
He can be contacted at email:
chriss_m
arouane@
yahoo.fr.
Evaluation Warning : The document was created with Spire.PDF for Python.