I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
2
,
A
p
r
il
20
25
,
p
p
.
1
7
9
3
~
1
8
0
4
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
2
.
pp
1
7
9
3
-
1
8
0
4
1793
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
User beha
v
io
r an
a
ly
sis
f
o
r insider
a
tt
a
ck de
tect
io
n
u
sing
a
co
mbina
tion o
f
m
emo
ry
predict
io
n
mo
del and rec
urs
iv
e f
ea
ture
elimina
tion a
lg
o
rithm
Ya
y
a
Su
da
ry
a
T
ria
na
1
,
M
o
hd
Aza
m
O
s
m
a
n
2
,
Der
is
Stia
wa
n
3
,
Ra
hm
a
t
B
ud
ia
rt
o
4
1
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
S
y
s
t
e
m,
F
a
c
u
l
t
y
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
U
n
i
v
e
r
si
t
a
s
M
e
r
c
u
B
u
a
n
a
,
Ja
k
a
r
t
a
,
I
n
d
o
n
e
s
i
a
2
S
c
h
o
o
l
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
s,
U
n
i
v
e
r
si
t
i
S
a
i
n
s
M
a
l
a
y
s
i
a
,
P
e
n
a
n
g
,
M
a
l
a
y
si
a
3
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
En
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
C
o
mp
u
t
e
r
S
c
i
e
n
c
e
,
U
n
i
v
e
r
si
t
a
s Sr
i
w
i
j
a
y
a
,
I
n
d
r
a
l
a
y
a
,
I
n
d
o
n
e
si
a
4
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
C
o
l
l
e
g
e
o
f
C
o
mp
u
t
i
n
g
a
n
d
I
n
f
o
r
mat
i
o
n
,
A
l
-
B
a
h
a
U
n
i
v
e
r
si
t
y
,
A
l
b
a
h
a
,
S
a
u
d
i
A
r
a
b
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
3
1
,
2
0
2
3
R
ev
is
ed
Ma
r
2
0
,
2
0
2
4
Acc
ep
ted
Dec
2
,
2
0
2
4
Ex
isti
n
g
d
e
fe
n
se
to
o
ls
a
g
a
in
st
th
e
in
sid
e
r
a
tt
a
c
k
s
a
re
ra
re
,
n
o
t
in
re
a
l
ti
m
e
fa
sh
io
n
a
n
d
su
ffe
r
fr
o
m
lo
w
d
e
te
c
ti
o
n
a
c
c
u
ra
c
y
a
s
th
e
a
tt
a
c
k
s
b
e
c
o
m
e
m
o
re
so
p
h
ist
ica
ted
.
T
h
u
s,
a
d
e
tec
ti
o
n
t
o
o
l
with
o
n
li
n
e
lea
rn
in
g
a
b
il
it
y
a
n
d
b
e
tt
e
r
a
c
c
u
ra
c
y
is r
e
q
u
ired
u
r
g
e
n
tl
y
.
Th
i
s stu
d
y
p
ro
p
o
se
s a
n
in
si
d
e
r
a
tt
a
c
k
d
e
tec
ti
o
n
m
o
d
e
l
b
y
lev
e
ra
g
in
g
e
n
ti
ty
b
e
h
a
v
io
r
a
n
a
l
y
sis
tec
h
n
iq
u
e
b
a
se
d
o
n
a
m
e
m
o
ry
p
re
d
ictio
n
m
o
d
e
l
c
o
m
b
in
e
d
wit
h
t
h
e
re
c
u
rsiv
e
fe
a
tu
re
e
li
m
in
a
t
io
n
(RF
E)
fe
a
tu
re
se
lec
ti
o
n
a
l
g
o
r
it
h
m
.
Th
e
m
e
m
o
ry
-
p
re
d
ictio
n
m
o
d
e
l
p
r
o
v
i
d
e
s
a
b
il
it
y
to
p
e
rf
o
rm
o
n
li
n
e
lea
rn
in
g
,
wh
il
e
th
e
RF
E
a
l
g
o
rit
h
m
is
d
e
p
lo
y
e
d
to
re
d
u
c
e
d
a
ta
d
ime
n
sio
n
a
li
ty
.
Da
tas
e
t
fo
r
th
e
e
x
p
e
rime
n
t
wa
s
c
re
a
ted
fr
o
m
a
re
a
l
n
e
two
rk
wit
h
1
5
0
a
c
ti
v
e
u
se
rs,
a
n
d
m
ix
e
d
wi
th
a
tt
a
c
k
s
d
a
ta
fro
m
p
u
b
li
c
ly
a
v
a
il
a
b
le
d
a
tas
e
t.
T
h
e
d
a
tas
e
t
is
s
imu
late
d
o
n
a
tes
tb
e
d
n
e
tw
o
rk
e
n
v
iro
n
m
e
n
t
c
o
n
sistin
g
o
f
a
se
rv
e
r
c
o
n
fi
g
u
re
d
to
ru
n
4
v
irt
u
a
l
se
rv
e
rs
a
n
d
o
th
e
r
tw
o
c
o
m
p
u
ters
a
s
traffic
g
e
n
e
ra
to
r
a
n
d
d
e
tec
ti
o
n
t
o
o
l.
Th
e
e
x
p
e
rime
n
tal
re
su
lt
s
sh
o
w
9
4
.
0
1
%
o
f
d
e
tec
ti
o
n
a
c
c
u
ra
c
y
,
9
5
.
6
4
%
o
f
p
re
c
isi
o
n
,
9
9
.
2
8
%
o
f
se
n
siti
v
it
y
,
a
n
d
9
6
.
0
8
%
o
f
F
1
-
sc
o
re
.
Th
e
p
r
o
p
o
se
d
m
o
d
e
l
is
a
b
le
to
p
e
rfo
rm
on
-
t
h
e
-
fly
lea
rn
in
g
to
a
d
d
re
ss
e
v
o
l
v
in
g
n
a
tu
re
o
f
th
e
a
tt
a
c
k
s.
Co
m
b
in
i
n
g
m
e
m
o
ry
p
re
d
icti
o
n
m
o
d
e
ls
with
th
e
RF
E
f
o
r
u
se
r
b
e
h
a
v
i
o
r
a
n
a
ly
sis
is
a
p
ro
m
isin
g
a
p
p
ro
a
c
h
,
a
n
d
a
c
h
iev
in
g
h
i
g
h
a
c
c
u
ra
c
y
is
d
e
fin
it
e
l
y
a
p
o
siti
v
e
o
u
tco
m
e
.
K
ey
w
o
r
d
s
:
E
n
tity
b
eh
av
i
o
r
an
al
y
s
is
I
n
s
id
er
attac
k
Me
m
o
r
y
p
r
ed
ictio
n
m
o
d
el
R
ea
l tim
e
d
etec
tio
n
R
ec
u
r
s
iv
e
f
ea
tu
r
e
elim
in
atio
n
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Yay
a
Su
d
ar
y
a
T
r
ian
a
Dep
ar
tm
en
t o
f
I
n
f
o
r
m
atio
n
Sy
s
tem
,
Facu
lty
o
f
C
o
m
p
u
te
r
Scien
ce
,
Un
iv
er
s
itas
Me
r
cu
B
u
an
a
J
l.
Me
r
u
y
a
Selatan
No
.
1
Kem
b
an
g
an
,
J
ak
ar
ta
B
ar
at
1
1
6
5
0
,
I
n
d
o
n
esia
E
m
ail: y
ay
a.
s
u
d
ar
y
a@
m
er
cu
b
u
an
a.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
Netwo
r
k
s
ec
u
r
ity
is
g
ettin
g
s
m
ar
ter
all
th
e
tim
e
wh
en
it
c
o
m
es
to
f
in
d
in
g
th
r
ea
ts
in
r
ea
l
-
tim
e.
T
h
e
tr
ick
is
f
in
d
in
g
th
o
s
e
th
r
ea
ts
q
u
ick
ly
with
o
u
t
s
lo
win
g
d
o
w
n
th
e
wh
o
le
s
y
s
tem
.
W
e
n
ee
d
s
y
s
tem
s
th
at
k
ee
p
lear
n
in
g
as
th
ey
g
o
,
ad
j
u
s
tin
g
ev
en
th
e
tin
iest
ch
an
g
es.
T
h
e
b
ig
g
est
ch
allen
g
e
is
m
a
k
in
g
s
u
r
e
th
e
s
y
s
tem
o
n
ly
d
etec
ts
r
ea
l
th
r
ea
ts
,
an
d
d
o
es
n
o
t
g
et
co
n
f
u
s
ed
b
y
n
o
r
m
al
a
ctiv
ity
.
Sin
ce
h
o
w
p
eo
p
le
u
s
e
n
etwo
r
k
s
ch
an
g
es,
s
ec
u
r
ity
s
y
s
tem
s
n
ee
d
to
ch
a
n
g
e
with
th
em
to
f
ig
h
t
o
f
f
n
ew
attac
k
s
.
Ad
a
p
tin
g
o
n
th
e
f
ly
an
d
r
e
-
u
s
in
g
k
n
o
wled
g
e
ac
r
o
s
s
d
if
f
er
e
n
t a
r
ea
s
is
th
e
k
ey
to
s
tay
in
g
s
ec
u
r
e.
Me
m
o
r
y
p
r
e
d
ictio
n
is
g
ain
in
g
s
team
,
b
u
t
it
is
s
til
l
ea
r
ly
d
ay
s
co
m
p
ar
ed
to
n
etwo
r
k
s
ec
u
r
ity
.
As
a
r
esu
lt,
th
e
“b
est
way
”
to
d
o
it
is
s
til
l
b
ein
g
f
ig
u
r
e
d
o
u
t
.
I
n
c
o
n
tr
ast,
u
s
er
b
eh
av
io
r
an
aly
s
is
(
UB
A)
,
wh
ich
also
co
n
s
id
er
s
d
ev
ices,
ap
p
licatio
n
s
,
an
d
ev
en
n
etwo
r
k
tr
af
f
ic,
is
alr
ea
d
y
a
v
ital
to
o
l
in
m
o
d
er
n
cy
b
e
r
s
ec
u
r
ity
.
UB
A
u
s
es
d
if
f
er
en
t
Su
p
er
v
is
e
d
an
d
Un
s
u
p
er
v
is
ed
L
ea
r
n
in
g
alg
o
r
ith
m
s
to
u
n
d
er
s
tan
d
n
o
r
m
al
u
s
er
b
eh
av
i
o
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
7
9
3
-
1
8
0
4
1794
an
d
s
p
o
t
an
y
r
ed
f
la
g
s
.
I
t
als
o
an
aly
ze
s
h
o
w
u
s
er
s
in
ter
ac
t
with
s
y
s
tem
s
,
lik
e
th
eir
ty
p
in
g
p
atter
n
s
,
m
o
u
s
e
m
o
v
em
en
ts
,
an
d
lo
g
in
s
,
to
ca
t
ch
p
o
ten
tial
u
n
a
u
th
o
r
ize
d
ac
ce
s
s
.
Desp
ite
r
ea
l
-
tim
e
attac
k
/an
o
m
aly
d
etec
tio
n
in
n
etwo
r
k
s
ec
u
r
ity
h
as
m
ad
e
b
i
g
s
tr
id
es,
th
er
e
a
r
e
s
till
s
o
m
e
im
p
o
r
tan
t
c
h
allen
g
es
r
esear
ch
er
s
ar
e
wo
r
k
i
n
g
o
n
.
B
y
tack
lin
g
th
ese
ch
allen
g
es,
r
esear
ch
er
s
ca
n
c
r
ea
te
ev
en
m
o
r
e
e
f
f
ec
tiv
e
an
d
r
eliab
le
r
ea
l
-
tim
e
attac
k
/an
o
m
aly
d
etec
tio
n
s
y
s
t
em
s
th
at
m
ak
e
n
etwo
r
k
s
m
o
r
e
s
ec
u
r
e
ag
ain
s
t
th
e
ev
er
-
ch
a
n
g
in
g
wo
r
ld
o
f
cy
b
e
r
th
r
ea
ts
.
T
h
ese
c
h
allen
g
es
in
cl
u
d
e:
i)
B
alan
cin
g
h
o
w
ac
cu
r
a
te
d
etec
tio
n
s
a
r
e
with
h
o
w
ef
f
icien
tly
th
e
s
y
s
tem
r
u
n
s
;
ii)
Ma
k
in
g
s
u
r
e
th
e
s
y
s
t
em
ca
n
ad
ap
t
to
n
ew
th
r
ea
ts
an
d
ch
an
g
es
in
h
o
w
p
eo
p
le
u
s
e
th
e
n
etwo
r
k
;
an
d
iii)
Un
d
er
s
tan
d
in
g
h
o
w
th
e
s
y
s
tem
m
ak
es
its
d
ec
is
io
n
s
,
esp
ec
ially
wh
en
u
s
in
g
ar
tific
i
al
in
tellig
en
ce
(
AI
)
b
ased
m
o
d
els.
Fen
d
in
g
o
f
f
attac
k
s
f
r
o
m
o
u
t
s
id
e
th
e
in
s
titu
tio
n
al/o
r
g
an
iza
tio
n
al
n
etwo
r
k
is
r
elativ
ely
e
asier
to
d
o
th
an
f
en
d
in
g
o
f
f
cy
b
er
cr
im
e
attac
k
s
f
r
o
m
with
in
th
e
n
etwo
r
k
.
E
x
ter
n
al
attac
k
s
ca
n
b
e
p
r
ev
en
ted
b
y
u
s
in
g
f
ir
ewa
lls
,
an
ti
-
v
ir
u
s
an
d
s
p
ec
ial
s
o
f
twar
e
f
o
r
in
tr
u
d
e
r
/m
alwa
r
e
d
etec
tio
n
.
Ho
wev
er
,
d
ef
en
s
e
to
o
ls
ag
ain
s
t
th
e
in
s
id
er
attac
k
s
ar
e
r
ar
e
an
d
s
u
f
f
er
f
r
o
m
lo
w
d
etec
tio
n
ac
c
u
r
ac
y
as
th
e
attac
k
s
b
ec
o
m
e
m
o
r
e
s
o
p
h
is
ticated
.
Ma
n
y
b
ig
cy
b
er
-
attac
k
s
in
v
o
lv
in
g
in
s
id
er
s
h
av
e
h
ap
p
e
n
ed
,
in
clu
d
in
g
m
illi
o
n
s
o
f
Yah
o
o
em
ail
ac
co
u
n
ts
,
illeg
al
d
o
wn
lo
ad
in
g
o
f
d
ig
ital
m
o
v
ie
f
iles
f
r
o
m
So
n
y
Fil
m
C
o
m
p
an
y
an
d
r
an
s
o
m
war
e
i
n
h
o
s
p
itals
in
th
e
UK.
T
h
er
e
ar
e
m
an
y
attac
k
d
etec
tio
n
/p
r
ev
en
tio
n
s
y
s
tem
s
av
ail
ab
le
in
th
e
m
ar
k
et,
ap
ar
t
f
r
o
m
b
ein
g
ex
p
en
s
iv
e;
th
ese
s
y
s
tem
s
s
til
l
h
av
e
s
ev
er
al
wea
k
n
ess
es,
s
u
ch
as:
lo
w
d
etec
tio
n
ac
cu
r
a
cy
,
to
o
m
an
y
f
alse
alar
m
s
,
an
d
th
e
in
ab
ilit
y
to
ca
r
r
y
o
u
t
r
ea
l
-
tim
e
lear
n
i
n
g
f
o
r
n
ew
v
ar
ian
ts
o
f
attac
k
s
.
T
h
e
ab
ilit
y
o
f
th
e
s
y
s
tem
to
ca
r
r
y
o
u
t
r
ea
l
-
tim
e
lear
n
in
g
is
n
ee
d
ed
to
d
ea
l
with
r
ap
id
l
y
ev
o
lv
in
g
attac
k
s
/v
ir
u
s
es/ma
lwar
e.
Dete
ctin
g
an
attac
k
in
v
o
l
v
in
g
th
e
in
s
id
er
s
is
m
o
r
e
d
if
f
icu
lt
b
e
ca
u
s
e
th
e
d
ef
en
s
e
s
y
s
tem
u
s
ed
m
ay
th
in
k
th
at
th
e
attac
k
is
th
e
n
o
r
m
al
ac
tiv
ity
o
f
an
e
n
tity
with
in
th
e
s
y
s
tem
/n
etwo
r
k
.
B
esid
es,
th
e
attac
k
s
m
ay
b
e
ab
le
to
lear
n
to
ac
t
as
leg
itima
te
u
s
er
s
.
R
ef
er
r
in
g
to
th
e
b
ac
k
g
r
o
u
n
d
ab
o
v
e,
th
is
s
tu
d
y
attem
p
ts
to
b
u
ild
an
in
tellig
en
t
s
y
s
tem
f
o
r
d
etec
tin
g
in
s
id
er
attac
k
s
u
s
in
g
en
tity
b
eh
av
io
r
an
aly
s
is
.
I
n
s
tead
o
f
f
o
llo
win
g
th
e
tr
ad
itio
n
a
l
way
o
f
d
etec
tin
g
attac
k
s
/an
o
m
alies
u
s
in
g
r
u
le
-
b
ased
o
r
k
n
o
wled
g
e
-
b
ased
s
y
s
tem
s
,
th
is
s
tu
d
y
p
r
ef
e
r
s
to
u
s
e
en
tity
b
eh
a
v
io
r
an
aly
s
is
b
y
u
tili
zin
g
h
u
m
a
n
m
em
o
r
y
m
o
d
elin
g
to
p
r
ed
ict
e
n
t
ity
b
eh
av
i
o
r
b
ased
o
n
th
e
en
tit
y
'
s
tr
af
f
ic
d
ata
[
1
]
.
I
n
th
is
ca
s
e,
th
e
s
y
s
tem
i
s
f
ir
s
tly
tr
ain
ed
to
b
u
ild
a
p
r
o
f
ile
o
f
en
titi
es
in
th
e
n
etwo
r
k
,
an
d
th
en
ex
am
in
es
th
e
n
o
r
m
ality
o
f
th
at
b
eh
av
i
o
r
.
I
n
ad
d
itio
n
,
to
ac
h
iev
e
f
ast
d
etec
t
io
n
,
th
e
p
r
o
p
o
s
ed
d
etec
tio
n
s
y
s
tem
co
n
s
id
er
s
th
e
r
ec
u
r
s
iv
e
f
ea
t
u
r
e
elim
in
atio
n
(
R
FE)
f
o
r
r
e
d
u
cin
g
th
e
u
n
r
elate
d
f
ea
tu
r
es
o
f
th
e
tr
a
f
f
ic
th
at
d
e
g
r
ad
i
n
g
t
h
e
d
etec
tio
n
ac
c
u
r
ac
y
lev
el.
E
x
p
er
im
en
ts
ar
e
ca
r
r
ied
o
u
t
in
a
t
estb
ed
en
v
ir
o
n
m
e
n
t
with
s
ev
er
al
en
titi
es
an
d
th
e
s
y
s
tem
will
p
r
ed
ict
wh
eth
er
a
p
ar
ticu
lar
en
tity
is
ca
r
r
y
in
g
o
u
t
illeg
al
ac
tiv
itie
s
,
wh
ich
th
en
en
d
s
with
m
ak
in
g
a
d
ec
is
io
n
to
d
eter
m
in
e
wh
eth
er
th
e
en
tity
is
n
o
r
m
al
o
r
an
at
tack
/an
o
m
aly
.
Z
h
an
g
et
a
l.
[
2
]
em
p
lo
y
ed
o
p
t
im
izatio
n
th
eo
r
y
to
e
x
am
in
e
wh
ich
u
s
er
s
co
n
n
ec
ted
with
th
e
s
tr
o
n
g
est
s
h
o
r
t
-
an
d
lo
n
g
-
ter
m
e
f
f
ec
ts
f
o
r
th
eir
r
esp
ec
tiv
e
tar
g
et
u
s
er
s
b
ased
o
n
th
e
m
o
b
ile
s
o
cial
en
v
ir
o
n
m
e
n
t
o
f
u
s
er
s
.
T
h
e
g
o
al
o
f
i
n
teg
r
atin
g
th
ese
u
s
er
b
eh
av
i
o
r
s
am
p
les
in
to
a
t
ar
g
et
u
s
er
s
am
p
le
d
atab
ase
is
to
cr
ea
te
a
s
am
p
lin
g
p
r
o
ce
s
s
th
at
will
g
r
ea
tly
in
cr
ea
s
e
th
e
ac
cu
r
ac
y
o
f
u
s
er
b
eh
a
v
io
r
p
r
e
d
ictio
n
s
.
Nex
t,
two
o
p
tim
izatio
n
m
o
d
els
ar
e
d
ev
elo
p
e
d
b
ased
o
n
t
h
e
d
eg
r
ee
o
f
in
ter
ac
tio
n
an
d
s
im
ilar
ity
,
r
esp
ec
tiv
ely
,
to
ch
o
o
s
e
th
e
b
est
ass
o
ciate
d
u
s
er
s
f
o
r
e
x
am
in
in
g
th
e
two
p
r
im
ar
y
c
o
m
p
o
n
en
ts
o
f
tar
g
et
u
s
er
b
eh
av
i
o
r
;
Fu
r
th
e
r
m
o
r
e
,
a
n
ad
ap
tiv
e
u
p
d
atin
g
s
tr
ateg
y
b
ased
o
n
f
u
zz
y
th
eo
r
y
is
p
r
o
p
o
s
ed
to
d
escr
ib
e
t
h
e
im
p
o
r
tan
ce
o
f
two
f
ac
to
r
s
in
r
ea
l
tim
e
an
d
q
u
an
titativ
e
m
an
n
e
r
.
Nex
t,
Ap
r
io
r
i
th
eo
r
y
is
in
tr
o
d
u
ce
d
to
p
r
ed
ict
th
e
u
s
er
'
s
n
ex
t
s
er
v
ice
b
eh
av
io
r
ac
cu
r
ately
;
in
p
ar
ticu
lar
,
th
e
Ap
r
io
r
i
s
am
p
le
d
atab
ase
u
p
d
ate
m
ec
h
an
is
m
was
b
u
ilt
to
ef
f
ec
tiv
ely
in
t
eg
r
ate
th
e
o
p
tim
al
s
am
p
le
o
f
co
r
r
elate
d
u
s
er
s
.
Fin
ally
,
ex
ten
s
iv
e
s
im
u
latio
n
r
esu
lts
s
h
o
w
th
at
th
e
p
r
o
p
o
s
ed
al
g
o
r
ith
m
o
u
tp
er
f
o
r
m
s
s
ev
er
al
r
elate
d
a
lg
o
r
ith
m
s
in
ter
m
s
o
f
ac
cu
r
ac
y
,
p
r
ed
ictab
ilit
y
an
d
o
p
er
atin
g
ef
f
icien
cy
.
Mo
r
e
r
esear
ch
es r
elate
d
to
th
e
a
n
aly
s
is
o
f
en
tity
b
eh
av
io
r
in
th
e
f
ie
ld
o
f
cy
b
er
s
ec
u
r
ity
ca
n
b
e
s
ee
n
in
[
3
]
–
[
6
]
.
Me
an
wh
ile,
Sh
ar
ip
u
d
d
in
et
a
l.
[
7
]
b
u
ilt
an
in
tr
u
s
io
n
d
et
ec
tio
n
s
y
s
tem
(
I
DS)
an
d
s
u
cc
ee
d
ed
in
im
p
r
o
v
in
g
d
etec
tio
n
ac
c
u
r
ac
y
an
d
p
r
ec
is
io
n
b
y
u
s
in
g
r
ec
u
r
s
iv
e
f
ea
tu
r
e
elim
in
atio
n
(
R
FE)
a
lg
o
r
ith
m
as f
ea
tu
r
e
ex
tr
ac
tio
n
.
E
x
p
e
r
im
en
ts
o
n
th
e
f
ea
tu
r
e
ex
tr
ac
tio
n
d
ataset
f
r
o
m
an
in
ter
n
et
o
f
th
in
g
s
(
I
o
T
)
test
b
ed
n
etwo
r
k
h
av
e
b
ee
n
ca
r
r
ied
o
u
t
to
in
v
e
s
tig
ate
th
e
ef
f
ec
t
o
f
th
e
ex
tr
a
ctio
n
p
r
o
ce
s
s
o
n
attac
k
d
etec
t
io
n
an
d
th
e
r
esu
lts
s
h
o
w
p
er
f
ec
t
o
f
1
0
0
%
d
etec
t
io
n
ac
cu
r
ac
y
.
M
o
r
e
r
esear
ch
wo
r
k
s
o
n
attac
k
/an
o
m
aly
d
e
tectio
n
u
s
in
g
R
FE
alg
o
r
ith
m
ar
e
p
r
esen
ted
in
[
8
]
–
[
1
1
]
A
n
ew
f
r
am
ewo
r
k
f
o
r
ex
p
r
ess
in
g
th
e
f
u
n
ctio
n
o
f
th
e
h
u
m
an
b
r
ain
'
s
n
eo
co
r
tex
was
p
r
o
p
o
s
ed
b
y
Haw
k
in
s
an
d
Ah
m
ad
[
1
2
]
;
H
awk
in
s
et
a
l.
[
1
3
]
;
E
ich
en
b
a
u
m
[
1
]
;
T
r
ia
n
a
et
a
l.
[
1
4
]
;
an
d
L
iu
an
d
L
am
[
1
5
]
.
Fu
r
th
er
m
o
r
e,
o
th
e
r
r
esear
ch
h
as
r
ev
ea
led
th
at
g
r
id
ce
lls
—
wh
ich
r
esem
b
le
n
eu
r
o
n
s
—
m
ay
p
o
te
n
tially
ex
i
s
t
in
th
e
n
eo
co
r
te
x
[
1
2
]
,
[
1
6
]
.
Ap
p
licatio
n
s
f
o
r
m
em
o
r
y
p
r
e
d
ictio
n
f
r
am
ewo
r
k
s
ca
n
b
e
f
o
u
n
d
in
a
v
a
r
iety
o
f
d
o
m
ain
s
,
in
clu
d
in
g
r
ea
l
-
tim
e
n
etwo
r
k
in
g
[
1
7
]
,
[
1
8
]
,
o
n
lin
e
e
d
u
ca
tio
n
[
1
9
]
,
[
2
0
]
,
o
b
ject
id
e
n
tific
atio
n
[
2
1
]
,
[
2
2
]
,
an
d
m
e
d
icin
e
[
2
3
]
.
Mo
h
am
e
d
[
2
4
]
ca
r
r
ied
o
u
t
a
n
e
x
tr
em
el
y
th
o
r
o
u
g
h
r
ev
iew
o
f
th
e
liter
atu
r
e
o
n
AI
m
et
h
o
d
s
f
o
r
attac
k
an
d
an
o
m
al
y
d
etec
tio
n
.
I
n
o
r
d
er
to
d
etec
t in
tr
u
s
io
n
in
h
eter
o
g
en
e
o
u
s
n
etwo
r
k
s
,
Sh
ar
ip
u
d
d
in
et
a
l.
[
2
5
]
s
u
g
g
est
co
m
b
in
in
g
a
d
e
ep
lear
n
in
g
s
tr
ateg
y
with
an
R
FE
-
b
ased
f
ea
tu
r
e
ex
tr
ac
tio
n
m
eth
o
d
.
T
h
e
ex
p
e
r
im
en
tal
r
esu
lts
o
n
cr
ea
ted
d
atas
et
f
r
o
m
a
test
b
ed
n
etwo
r
k
s
h
o
w
th
at
th
e
ac
cu
r
ac
y
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
r
ea
ch
es
ac
cu
r
ac
y
lev
el
ab
o
v
e
9
9
%.
Ap
r
u
zz
ese
et
a
l.
[
2
6
]
an
aly
z
e
m
ac
h
in
e
lear
n
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
User
b
eh
a
vio
r
a
n
a
lysi
s
fo
r
in
s
id
er a
tta
ck
d
etec
tio
n
u
s
in
g
a
c
o
mb
in
a
tio
n
o
f
…
(
Ya
ya
S
u
d
a
r
ya
Tr
ia
n
a
)
1795
tech
n
iq
u
es
ap
p
lied
to
th
e
d
e
tectio
n
o
f
in
tr
u
s
io
n
,
m
alwa
r
e,
an
d
s
p
am
.
T
h
e
au
th
o
r
s
r
ev
ea
l
th
e
d
etec
tio
n
ac
cu
r
ac
y
o
f
9
8
.
8
8
%.
Op
tim
al
en
s
em
b
le
I
DS
g
en
er
atio
n
is
p
r
esen
ted
b
y
Sti
awa
n
et
a
l.
[
2
7
]
.
in
f
o
r
m
atio
n
g
ain
,
g
ain
r
atio
,
s
y
m
m
etr
ical
u
n
ce
r
tain
ty
,
R
elief
-
F,
On
e
-
R
,
an
d
C
h
i
-
Sq
u
a
r
e
ar
e
th
e
s
ix
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
es
th
at
ar
e
em
p
lo
y
ed
.
T
h
e
f
o
llo
win
g
clas
s
if
icatio
n
tech
n
iq
u
es
wer
e
ap
p
lied
:
n
aïv
e
B
ay
esian
,
B
ay
esia
n
n
etwo
r
k
,
d
ec
is
io
n
tr
ee
:
J
4
8
,
an
d
s
elf
-
o
r
g
an
izin
g
m
ap
(
SOM)
.
Su
b
s
eq
u
en
tly
,
th
e
o
p
tim
al
f
ea
tu
r
es
f
r
o
m
ea
c
h
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
o
l
o
g
y
a
r
e
co
u
p
led
wit
h
ea
ch
class
if
ier
m
eth
o
d
to
c
r
ea
te
en
s
em
b
le
I
DSs
.
On
th
e
I
T
D
-
UT
M
d
ataset,
ex
p
er
im
en
tal
r
esu
lts
d
em
o
n
s
tr
ate
f
iv
e
o
p
tim
al
e
n
s
em
b
le
I
DSs
:
s
y
m
m
etr
ical
u
n
ce
r
tain
ty
+
B
ay
esian
n
etwo
r
k
;
ch
i
-
s
q
u
ar
e+
B
ay
esian
n
etwo
r
k
,
C
h
i
Sq
u
ar
e+
SOM,
in
f
o
r
m
at
io
n
g
ain
a
n
d
n
aï
v
e
B
ay
es;
an
d
On
e
-
R
+Bay
esian
n
etwo
r
k
,
with
r
esp
ec
tiv
e
ten
,
f
o
u
r
an
d
s
ev
en
b
est
s
elec
ted
f
ea
tu
r
es
ac
h
iev
e
8
1
.
0
3
1
6
%,
8
5
.
2
5
9
3
%,
an
d
8
0
.
8
6
2
5
%
o
f
ac
cu
r
ac
y
,
r
esp
ec
tiv
ely
.
Fu
r
th
er
m
o
r
e,
th
e
b
est
F
-
m
ea
s
u
r
e
v
alu
es,
i.e
.
,
0
.
8
5
3
a
n
d
0
.
8
3
0
,
r
esp
ec
tiv
ely
,
ar
e
ac
h
ie
v
ed
b
y
en
s
em
b
le
I
DSs
u
tili
zin
g
s
y
m
m
etr
ical
u
n
ce
r
tain
ty
+Bay
esian
n
etwo
r
k
an
d
o
n
e
-
R
+J4
8
with
th
e
ten
a
n
d
s
ix
b
est
s
elec
ted
f
ea
tu
r
es,
r
esp
ec
tiv
ely
.
Ad
d
itio
n
ally
,
th
e
lo
n
g
s
h
o
r
t
ter
m
m
em
o
r
y
(
L
STM
)
m
o
d
el,
wh
ich
is
f
u
r
n
is
h
ed
with
th
e
R
FE
f
ea
tu
r
e
s
e
lectio
n
m
eth
o
d
,
was
em
p
lo
y
e
d
b
y
Sti
awa
n
et
a
l.
[
2
8
]
.
Me
a
n
wh
ile,
Ku
r
n
iab
u
d
i
et
a
l
.
[
2
9
]
c
o
m
b
in
e
d
PS
O
-
s
ea
r
ch
an
d
r
a
n
d
o
m
f
o
r
est
to
im
p
r
o
v
e
attac
k
d
etec
tio
n
p
er
f
o
r
m
an
ce
o
n
th
e
in
ter
n
et
o
f
th
in
g
s
.
C
o
n
cu
r
r
en
tly
,
a
g
r
ea
t
d
ea
l
o
f
r
esear
ch
h
as
b
ee
n
d
o
n
e
o
n
th
e
s
u
b
ject
o
f
o
n
lin
e/r
ea
l
-
tim
e
attac
k
/an
o
m
a
ly
d
etec
tio
n
.
So
m
e
o
f
th
ese
s
tu
d
ies
ar
e
th
o
s
e
b
y
R
iv
er
a
et
a
l.
[
3
0
]
,
Kan
d
h
r
o
et
a
l.
[
3
1
]
,
Mo
h
am
m
ed
a
n
d
B
o
u
ch
ac
h
ia
[
3
2
]
,
an
d
T
y
ag
i
a
n
d
Ku
m
ar
[
3
3
]
.
B
u
d
iar
to
et
a
l.
[
3
4
]
p
r
o
p
o
s
ed
a
m
em
o
r
y
m
o
d
el
b
y
a
p
p
ly
in
g
a
m
em
o
r
y
p
r
ed
ictio
n
f
r
am
ewo
r
k
,
ca
lle
d
“
s
im
p
lifie
d
s
in
g
le
ce
ll
ass
em
b
led
s
eq
u
en
tial
h
ier
ar
ch
ical
m
e
m
o
r
y
(
s
.
SC
ASHM)
”.
T
h
en
th
i
s
m
o
d
el
is
u
s
ed
as
a
to
o
l
to
p
r
ed
ict
en
tity
b
e
h
av
io
r
an
d
d
etec
t
attac
k
s
in
v
o
lv
i
n
g
in
s
id
er
attac
k
s
/an
o
m
alies.
T
h
e
ex
p
er
im
e
n
tal
r
esu
lts
s
h
o
w
th
at
th
e
p
r
o
p
o
s
ed
m
em
o
r
y
m
o
d
el
s
u
cc
ess
f
u
l
ly
p
r
e
d
icts
th
e
tr
a
f
f
ic
b
eh
av
i
o
r
o
f
e
n
titi
es
with
v
ar
y
in
g
d
eg
r
ee
s
o
f
ac
cu
r
ac
y
f
r
o
m
7
2
%
to
8
3
%
an
d
is
ca
p
a
b
le
o
f
lear
n
in
g
o
n
-
th
e
-
f
l
y
,
wh
en
n
ew
p
atter
n
s
o
f
attac
k
s
co
m
e.
T
h
e
r
esear
ch
in
th
is
p
ap
er
a
d
o
p
ts
th
e
m
o
d
el
p
r
o
p
o
s
ed
b
y
B
u
d
iar
to
et
a
l.
[
3
4
]
an
d
co
m
b
in
ed
it
with
th
e
R
FE
m
eth
o
d
as
f
ea
tu
r
e
s
elec
tio
n
.
T
h
u
s
,
th
e
p
r
o
p
o
s
ed
m
o
d
el
is
ca
lled
m
em
o
r
y
p
r
ed
ictio
n
m
o
d
el
with
r
ec
u
r
s
iv
e
f
ea
tu
r
e
elim
in
ati
o
n
(
MPM
-
R
FE)
.
T
h
is
r
esear
ch
wo
r
k
co
n
tr
ib
u
tes
to
war
d
s
t
h
e
d
ev
elo
p
m
en
t
o
f
in
tellig
en
t r
ea
l
-
tim
e
an
o
m
aly
/in
s
id
er
s
’
attac
k
d
etec
tio
n
.
T
h
e
r
est
o
f
th
e
p
ap
e
r
is
s
tr
u
ct
u
r
ed
as
f
o
llo
ws
:
Sectio
n
2
d
is
cu
s
s
es
th
e
m
eth
o
d
u
s
ed
in
th
i
s
r
esear
ch
.
S
ec
tio
n
3
p
r
esen
ts
th
e
r
esear
c
h
r
esu
lts
an
d
d
is
cu
s
s
io
n
.
I
n
clo
s
in
g
,
s
ec
tio
n
4
co
n
clu
d
es th
e
e
n
tire
r
esear
ch
.
2.
M
E
T
H
O
D
T
h
e
r
esear
ch
m
eth
o
d
is
d
iv
id
ed
in
to
th
r
ee
s
tag
es.
Stag
e
1
is
th
e
d
ataset
cr
ea
tio
n
f
o
r
ex
p
er
im
en
ts
.
Stag
e
2
is
th
e
f
ea
tu
r
e
ex
tr
ac
ti
o
n
u
s
in
g
FR
E
alg
o
r
ith
m
.
Stag
e
3
is
th
e
d
ev
el
o
p
m
en
t
o
f
a
n
e
n
g
in
e
to
p
r
ed
ict
th
e
u
s
er
b
eh
a
v
io
r
tr
af
f
ic
b
ased
o
n
a
m
em
o
r
y
p
r
ed
ictio
n
m
o
d
e
l.
Fig
u
r
e
1
s
h
o
ws
th
e
w
o
r
k
f
l
o
w
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
an
d
is
ex
p
lain
ed
i
n
d
eta
il in
th
e
f
o
llo
win
g
s
ec
tio
n
s
.
Fig
u
r
e
1
.
Ov
e
r
all
ar
ch
itectu
r
e
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
(
ad
o
p
te
d
f
r
o
m
[
3
4
]
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
7
9
3
-
1
8
0
4
1796
Hier
ar
ch
ical
tem
p
o
r
al
m
em
o
r
y
p
r
o
v
id
es
a
f
r
am
ewo
r
k
th
at
m
o
d
els
s
ev
er
al
co
m
p
u
tatio
n
al
p
r
in
cip
les
th
at
o
cc
u
r
in
th
e
n
eo
c
o
r
tex
(
a
p
ar
t
o
f
th
e
h
u
m
an
b
r
ain
)
.
T
h
e
s
p
atial
p
o
o
ler
m
o
d
els
h
o
w
n
e
u
r
o
n
s
lear
n
f
ee
d
f
o
r
war
d
an
d
f
o
r
m
r
ep
r
es
en
tatio
n
s
o
f
in
p
u
t
d
ata
e
f
f
icien
tly
(
f
ast
an
d
ac
cu
r
ately
)
.
T
h
is
s
p
atial
p
o
o
ler
co
n
v
er
ts
b
in
ar
y
in
p
u
t
p
atter
n
s
(
0
,
1
)
i
n
to
s
p
ar
s
e
d
is
tr
ib
u
ted
r
ep
r
esen
tatio
n
s
(
SDR
s
)
u
s
in
g
a
co
m
b
in
atio
n
o
f
Heb
b
ian
co
m
p
etitiv
e
lear
n
in
g
r
u
les.
T
h
is
SD
R
d
ata
is
u
s
ed
to
m
o
d
el
th
e
m
em
o
r
y
f
o
r
ce
r
tain
ty
p
es
o
f
tr
af
f
ic
(
em
ail,
HT
T
P,
an
d
ap
p
licatio
n
s
)
.
Fo
r
th
e
p
r
ed
ictio
n
p
r
o
ce
s
s
:
in
p
u
t
d
ata
in
SDR
f
o
r
m
is
u
s
ed
to
p
r
ed
ict,
ac
tiv
ate
m
em
o
r
y
a
n
d
also
p
r
o
v
id
e
f
ee
d
b
ac
k
f
o
r
ch
an
g
es
to
th
e
m
em
o
r
y
m
o
d
el
if
th
er
e
ar
e
s
ig
n
if
ican
t
ch
a
n
g
es
in
ex
is
tin
g
p
atter
n
s
in
th
e
m
em
o
r
y
m
o
d
el.
Af
ter
th
at,
th
e
p
r
ed
ictio
n
s
co
r
e
is
ca
lcu
lated
,
an
aly
ze
d
an
d
a
p
atter
n
is
d
eter
m
in
ed
f
r
o
m
th
e
i
n
p
u
t
d
ata.
2
.
1
.
Da
t
a
s
et
cr
ea
t
io
n
Data
f
o
r
th
e
ex
p
er
im
en
t
was
tak
en
f
r
o
m
t
h
e
n
etwo
r
k
in
th
e
b
u
ild
in
g
o
f
th
e
Facu
lty
o
f
C
o
m
p
u
tin
g
an
d
I
n
f
o
r
m
atio
n
,
Alb
ah
a
U
n
iv
er
s
ity
,
Sau
d
i
Ar
ab
ia,
f
o
r
2
we
ek
s
(
1
–
1
4
Ap
r
il
2
0
2
3
)
.
T
h
e
r
atio
n
al
o
f
s
elec
tin
g
th
is
n
etwo
r
k
as
th
e
s
o
u
r
ce
o
f
d
ataset,
is
th
e
n
etwo
r
k
in
t
h
e
b
u
ild
in
g
is
a
f
lat
n
etwo
r
k
co
n
n
ec
ted
t
h
r
o
u
g
h
s
witch
es
th
at
m
ak
es
th
e
tr
af
f
ic
ca
p
tu
r
in
g
p
r
o
ce
s
s
b
ec
o
m
e
ea
s
y
.
T
h
u
s
,
th
e
ca
p
tu
r
ed
tr
a
f
f
ic
r
e
f
lects
a
r
ea
l
-
wo
r
ld
n
etwo
r
k
wh
e
r
e
u
s
er
s
ac
ce
s
s
in
ter
n
et
ap
p
licatio
n
s
s
u
ch
as:
web
b
r
o
wsi
n
g
,
ch
attin
g
,
em
ail,
f
tp
,
s
tr
ea
m
in
g
v
id
eo
,
an
d
g
am
es.
I
t
was
r
ec
o
r
d
ed
th
at
th
er
e
wer
e
m
o
r
e
t
h
an
1
5
0
u
s
er
s
ac
ce
s
s
in
g
th
e
n
etwo
r
k
d
u
r
i
n
g
th
is
p
er
io
d
.
I
n
o
r
d
e
r
f
o
r
th
e
clien
t
ap
p
licatio
n
to
b
e
ab
le
to
co
ll
ec
t
all
tr
af
f
ic,
we
u
s
ed
a
n
etw
o
r
k
tap
to
s
n
if
f
th
e
n
etwo
r
k
tr
a
f
f
ic
b
ef
o
r
e
it
b
ein
g
s
en
t
f
r
o
m
an
d
to
th
e
p
ar
ticu
la
r
s
eg
m
en
t.
I
n
th
is
ex
p
er
i
m
en
t,
th
e
m
ac
h
in
e
wh
er
e
th
e
clien
t
ap
p
licatio
n
r
u
n
s
ar
e
co
n
f
ig
u
r
e
d
to
co
llect
n
etwo
r
k
tr
af
f
ic
in
p
r
o
m
is
cu
o
u
s
m
o
d
e
an
d
s
to
r
ed
in
to
a
d
ataset
in
.
f
o
r
m
at
.
T
h
e
atta
ck
tr
af
f
ic
is
tak
en
f
r
o
m
th
e
Mid
-
Atlan
tic
C
o
lleg
iate
C
y
b
er
Def
en
s
e
C
o
m
p
etitio
n
(
MA
C
C
D
C
)
d
ataset
[
3
5
]
.
T
h
e
r
ef
o
r
e
,
th
e
cr
ea
t
ed
d
ataset
co
n
s
is
ts
o
f
n
o
r
m
al
an
d
attac
k
s
tr
af
f
ics.
B
ec
au
s
e
it
is
n
o
t
p
o
s
s
ib
le
to
i
n
ject
attac
k
/an
o
m
aly
tr
af
f
ic
i
n
to
th
e
p
r
o
d
u
ctio
n
n
etwo
r
k
,
th
e
d
ata
th
at
h
as
b
ee
n
o
b
tain
ed
is
s
im
u
lated
ag
ai
n
i
n
a
test
b
ed
n
etwo
r
k
e
n
v
ir
o
n
m
en
t
co
n
s
is
tin
g
o
f
3
c
o
m
p
u
ter
s
co
n
n
ec
te
d
to
a
s
witch
.
Fig
u
r
e
2
illu
s
tr
ates
th
e
test
b
ed
n
etwo
r
k
f
o
r
th
e
ex
p
er
im
en
ts
.
A
s
er
v
er
is
co
n
f
ig
u
r
ed
to
r
u
n
4
v
ir
tu
al
s
er
v
er
s
.
PC
-
2
co
m
p
u
ter
is
u
s
ed
to
in
ject
tr
af
f
ic
p
ac
k
ets o
f
t
h
e
ca
p
tu
r
ed
d
ataset
in
to
th
e
n
etwo
r
k
.
T
h
e
d
etec
tio
n
m
o
d
u
le
is
in
s
talled
o
n
PC
-
1
,
wh
ich
is
co
n
n
ec
ted
to
th
e
m
ir
r
o
r
ed
p
o
r
t
o
n
th
e
s
witch
s
o
th
at
th
e
d
etec
to
r
is
ab
le
to
s
ee
all
tr
af
f
ic
with
in
th
e
n
etwo
r
k
s
eg
m
en
t
b
ein
g
m
o
n
ito
r
ed
.
T
h
e
s
p
ec
if
icatio
n
s
o
f
PC
-
1
an
d
PC
-
2
ar
e
as
f
o
llo
ws:
I
n
tel
C
o
r
e
i7
C
PU,
8
GB
R
AM
an
d
5
0
0
GB
h
ar
d
d
r
iv
e.
Fo
r
s
er
v
er
,
a
co
m
p
u
ter
with
I
n
tel
C
o
r
e
i7
C
PU,
1
6
GB
R
AM
an
d
1
T
B
h
ar
d
d
is
k
is
u
s
ed
.
All
co
m
p
u
ter
s
r
u
n
th
e
W
in
d
o
ws
1
0
o
p
e
r
atin
g
s
y
s
tem
.
T
h
e
m
o
d
u
les
f
o
r
th
e
m
em
o
r
y
p
r
e
d
ictio
n
m
o
d
el
is
im
p
lem
e
n
ted
in
th
e
J
av
a
p
r
o
g
r
a
m
m
in
g
l
an
g
u
ag
e
,
wh
ile
th
e
p
r
ed
ictio
n
m
o
d
u
le
u
s
in
g
d
ee
p
lear
n
in
g
,
i.e
.
th
e
L
STM
is
im
p
lem
en
ted
in
th
e
Py
th
o
n
p
r
o
g
r
am
m
in
g
lan
g
u
ag
e
an
d
th
e
Scik
it
-
lear
n
lib
r
ar
y
.
Fig
u
r
e
2
.
T
estb
ed
n
etwo
r
k
to
p
o
lo
g
y
2
.
2
.
F
ea
t
ures e
x
t
ra
ct
io
n us
ing
RF
E
T
h
e
n
e
x
t
s
tep
is
ex
tr
ac
tin
g
s
ig
n
if
ican
t
f
ea
tu
r
es
o
f
th
e
tr
a
f
f
ic.
T
h
e
aim
is
to
r
ed
u
ce
th
e
d
im
e
n
s
io
n
ality
o
f
th
e
d
ataset,
wh
ic
h
will
r
e
d
u
ce
th
e
d
etec
tio
n
tim
e
wh
ile
p
r
eser
v
e
a
g
o
o
d
ac
c
u
r
ac
y
lev
el.
T
h
is
s
tu
d
y
u
s
es
th
e
R
FE
alg
o
r
ith
m
as
a
f
ea
tu
r
e
ex
tr
ac
tio
n
/s
elec
tio
n
m
eth
o
d
.
T
h
e
R
FE
tr
an
s
f
o
r
m
s
th
e
o
r
ig
in
al
d
ata
lin
ea
r
ly
in
to
a
n
ew
co
o
r
d
in
ate
s
y
s
tem
b
y
m
ax
im
izin
g
th
e
v
ar
ia
n
ce
v
a
lu
e
[
3
6
]
.
I
t
ex
am
i
n
es
th
e
d
ataset
co
n
tain
in
g
o
b
s
er
v
atio
n
s
th
at
h
av
e
i
n
ter
-
c
o
r
r
elate
d
q
u
an
titativ
e
d
ep
en
d
en
t
v
ar
iab
les.
T
h
e
f
ield
s
o
f
tr
af
f
ic
p
ac
k
et
h
ea
d
er
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
User
b
eh
a
vio
r
a
n
a
lysi
s
fo
r
in
s
id
er a
tta
ck
d
etec
tio
n
u
s
in
g
a
c
o
mb
in
a
tio
n
o
f
…
(
Ya
ya
S
u
d
a
r
ya
Tr
ia
n
a
)
1797
s
u
ch
as:
s
o
u
r
ce
,
an
d
d
esti
n
atio
n
I
P
a
d
d
r
ess
es,
s
o
u
r
ce
an
d
d
esti
n
atio
n
p
o
r
ts
,
p
r
o
to
c
o
ls
,
tim
e
s
tam
p
an
d
attac
k
.
ar
e
co
n
s
id
er
ed
as
th
e
in
p
u
t
f
o
r
th
e
R
FE.
T
h
e
R
FE
th
en
ex
tr
ac
ts
th
e
s
ig
n
if
ican
ce
in
f
o
r
m
atio
n
f
r
o
m
t
h
e
d
atasets
with
th
e
aim
is
to
s
ig
n
if
y
th
e
d
ataset
as
a
s
et
o
f
n
ew
o
r
t
h
o
g
o
n
al
v
ar
iab
les,
i.e
.
p
r
in
ci
p
al
co
m
p
o
n
en
ts
an
d
r
ep
r
esen
ts
th
e
p
atter
n
s
im
ilar
ity
o
f
th
e
o
b
s
er
v
atio
n
s
g
r
ap
h
ically
.
T
h
e
n
in
e
e
x
t
r
a
ct
ed
f
e
at
u
r
es
wit
h
t
h
e
h
ig
h
est
p
r
o
d
u
c
ed
b
y
t
h
e
R
F
E
im
p
l
e
m
e
n
ta
ti
o
n
ar
e:
i
p
.
t
tl
,
ip
.
h
d
r
_
len
,
i
p
.
le
n
,
t
cp
.
f
la
g
s
,
w
i
n
d
o
w
.
s
iz
e
,
tc
p
.
h
d
r
_
le
n
,
w
p
a
n
.
s
r
c
_
p
a
n
,
w
p
a
n
.
d
s
t1
6
,
an
d
w
p
a
n
.
c_
md
.
T
h
e
R
FE
alg
o
r
i
th
m
f
o
r
p
er
f
o
r
m
in
g
th
e
tr
af
f
ic
f
ea
tu
r
e
ex
tr
ac
tio
n
is
s
h
o
wn
in
a
lg
o
r
ith
m
1
[
3
7
]
.
Alg
o
r
ith
m
1
.
T
h
e
r
ec
u
r
s
iv
e
f
e
atu
r
e
elim
in
atio
n
Input: Traffic Data
Output: FEATURE (a set of selected attributes)
Import modul decomposition from sklearn
data ← load_dataset
def main()
Y ← read(data)
RFE =
decomposition.RFE(n_components=8),
RFE.fit(Y))
Y=RFE.transform(Y)
FEATURE←Y
End
2
.
3
.
P
re
dict
io
n pro
ce
s
s
T
h
is
p
ap
er
ad
o
p
ts
th
e
m
eth
o
d
in
tr
o
d
u
ce
d
b
y
B
u
d
ia
r
to
et
a
l.
[
3
4
]
f
o
r
p
r
e
d
ictin
g
attac
k
/an
o
m
al
y
tr
af
f
ic.
T
h
e
m
eth
o
d
u
s
es
a
h
u
m
an
n
eo
co
r
te
x
in
s
p
ir
ed
h
y
b
r
i
d
g
en
e
-
co
n
tr
o
lled
m
ac
h
in
e
in
t
ellig
en
ce
ap
p
r
o
ac
h
.
T
h
e
m
eth
o
d
aim
s
to
u
s
e
th
e
h
u
m
an
n
eo
co
r
te
x
m
em
o
r
y
as
a
m
o
d
el
to
g
en
er
ate
h
u
m
an
in
tellig
en
ce
f
o
r
r
ec
o
g
n
itio
n
an
d
t
o
ap
p
l
y
a
n
e
u
r
o
g
en
etics m
eth
o
d
f
o
r
co
m
p
le
x
id
en
tific
atio
n
a
n
d
p
r
ed
ictio
n
.
2
.
3
.
1
.
Sp
a
rse
dis
t
rib
ute
d r
ep
re
s
ent
a
t
io
n (
SDR)
Me
m
o
r
ies
th
at
s
h
ar
e
s
o
m
e
co
m
m
o
n
f
ea
tu
r
es
ten
d
to
clu
s
ter
to
g
eth
er
in
h
u
m
an
m
em
o
r
y
,
e
v
en
if
th
e
y
h
av
e
n
o
c
o
n
n
ec
tio
n
.
T
h
e
SDR
is
a
way
o
f
ex
p
r
ess
in
g
h
u
m
a
n
m
em
o
r
y
with
m
ath
em
atics,
an
d
it
u
s
es
a
s
p
ac
e
with
m
an
y
d
im
en
s
io
n
s
to
ca
p
t
u
r
e
th
e
h
u
g
e
am
o
u
n
t
o
f
m
em
o
r
y
th
at
r
esem
b
les
th
e
h
u
m
a
n
b
r
ain
’
s
n
etwo
r
k
o
f
n
eu
r
o
n
s
.
A
k
ey
f
ea
tu
r
e
o
f
s
u
c
h
s
p
ac
es
with
m
an
y
d
im
e
n
s
io
n
s
is
th
at
two
v
ec
to
r
s
p
ick
ed
at
r
an
d
o
m
ar
e
v
er
y
f
ar
ap
ar
t,
wh
ich
m
ea
n
s
th
ey
h
av
e
n
o
r
elatio
n
.
T
h
e
SDR
s
av
es
a
lo
t
o
f
d
ata
in
a
s
m
all
s
p
ac
e,
u
s
in
g
s
o
m
e
s
p
ec
ial
p
lace
s
ca
lled
h
ar
d
lo
ca
tio
n
s
.
T
h
ese
p
lace
s
ar
e
s
p
r
ea
d
o
u
t e
v
en
ly
in
a
b
ig
g
e
r
s
p
ac
e
th
at
is
n
o
t r
ea
l.
E
ac
h
p
iece
o
f
d
ata
is
s
av
ed
b
y
u
s
in
g
s
o
m
e
o
f
t
h
ese
p
lace
s
,
an
d
ta
k
en
b
ac
k
b
y
c
o
m
b
in
i
n
g
t
h
em
.
Ho
wev
er
,
th
is
m
a
y
n
o
t w
o
r
k
v
er
y
well,
an
d
th
e
q
u
ality
o
f
th
e
d
ata
m
ay
c
h
an
g
e
d
ep
en
d
in
g
o
n
h
o
w
f
u
ll th
e
s
p
ac
e
is
.
T
h
e
tr
af
f
ic
d
ata
o
b
tain
e
d
f
r
o
m
th
e
f
ea
t
u
r
es
ex
tr
ac
tio
n
p
h
ase
is
co
n
v
e
r
ted
in
to
a
s
er
ies
o
f
in
d
iv
id
u
al
n
etwo
r
k
p
ac
k
ets,
b
y
r
ep
r
esen
tin
g
ea
c
h
b
y
te
o
f
th
e
tr
af
f
ic
d
ata
in
an
ato
m
ic
f
o
r
m
as
SDR
.
T
h
is
b
asic
f
o
r
m
is
in
th
e
f
o
r
m
o
f
a
v
ec
t
o
r
c
o
n
s
is
tin
g
o
f
a
s
eq
u
e
n
ce
o
f
2
0
4
8
b
its
.
Fo
r
ex
am
p
le,
c
o
n
s
id
er
th
e
v
alu
e
o
f
ℎ
_
f
ea
tu
r
e
is
8
0
0
b
y
tes
.
T
h
is
v
alu
e
is
r
ep
r
esen
ted
as
7
ato
m
ic
s
eq
u
en
ce
s
o
f
SDR
.
T
h
ese
b
it
v
ec
to
r
s
ar
e
in
p
u
tted
in
to
th
e
m
em
o
r
y
p
r
e
d
ictio
n
m
o
d
el
m
o
d
u
le
f
o
r
an
aly
s
is
.
I
n
th
e
p
r
ed
ictio
n
p
r
o
ce
s
s
,
th
is
v
al
u
e
is
co
m
p
ar
ed
with
th
e
th
r
esh
o
l
d
s
in
s
id
e
th
e
m
em
o
r
y
m
o
d
el
an
d
will
b
e
d
ec
id
ed
wh
eth
er
it
is
a
n
ew
f
ea
tu
r
e
o
r
n
o
t.
I
f
s
o
,
th
e
n
th
e
m
em
o
r
y
m
o
d
el
is
u
p
d
ated
.
2
.
3
.
2
.
M
em
o
ry
predict
io
n mo
del
I
n
h
u
m
an
n
eo
co
r
te
x
m
em
o
r
y
m
o
d
el,
am
o
n
g
th
e
b
r
ain
’
s
n
eu
r
al
cir
cu
its
,
th
e
ce
r
eb
ellu
m
’
s
co
r
tex
is
th
e
m
o
s
t
s
im
ilar
to
th
e
s
p
ar
s
e
d
is
tr
ib
u
ted
m
em
o
r
y
.
An
ass
o
ciativ
e
m
em
o
r
y
k
ee
p
s
a
wo
r
ld
m
o
d
el
th
at
co
n
n
ec
ts
s
en
s
o
r
y
in
p
u
t
to
ac
tio
n
.
T
h
e
m
em
o
r
y
r
ec
eiv
es
th
e
wo
r
ld
’
s
ev
en
ts
as
a
s
er
ies
o
f
lar
g
e
p
att
er
n
s
.
T
h
ese
p
atter
n
s
r
ep
r
esen
t
s
en
s
o
r
d
ata,
in
ter
n
a
l
-
s
tate
v
ar
iab
les,
an
d
co
m
m
a
n
d
s
to
th
e
ac
tu
at
o
r
s
.
T
h
e
m
e
m
o
r
y
’
s
ca
p
ac
ity
t
o
s
to
r
e
an
d
r
etr
iev
e
t
h
ese
s
er
ies
u
n
d
er
s
im
ilar
s
itu
atio
n
s
en
a
b
les
its
u
s
e
f
o
r
p
r
ed
ictio
n
[
3
8
]
.
T
h
u
s
,
a
h
u
m
a
n
n
eo
co
r
tex
in
s
p
ir
ed
h
y
b
r
id
g
e
n
e
-
co
n
tr
o
lled
m
ac
h
i
n
e
in
tellig
e
n
ce
ap
p
r
o
ac
h
is
co
n
s
id
er
ed
.
T
h
e
h
y
b
r
id
ap
p
r
o
ac
h
aim
s
to
m
o
d
el
h
u
m
an
n
eo
c
o
r
t
ex
m
em
o
r
y
to
p
r
o
d
u
ce
h
u
m
an
in
tellig
en
ce
f
o
r
r
ec
o
g
n
itio
n
an
d
to
im
p
lem
en
t
a
n
eu
r
o
g
en
etics
ap
p
r
o
ac
h
f
o
r
co
m
p
lex
id
en
tific
atio
n
an
d
p
r
ed
ictio
n
.
T
h
e
co
r
te
x
o
f
th
e
ce
r
eb
ellu
m
is
th
e
n
eu
r
o
n
al
cir
cu
it
i
n
th
e
b
r
ain
th
at
m
o
s
t
clo
s
ely
m
im
ics
s
p
a
r
s
e
d
is
tr
ib
u
ted
m
em
o
r
y
.
An
a
s
s
o
ciativ
e
m
em
o
r
y
k
ee
p
s
tr
ac
k
o
f
a
wo
r
ld
m
o
d
el
th
at
co
n
n
ec
ts
p
er
ce
p
tio
n
to
b
eh
av
io
r
.
T
h
e
wo
r
ld
'
s
ev
en
ts
a
r
e
p
r
esen
ted
to
th
e
m
em
o
r
y
as
a
s
er
ies
o
f
ex
p
an
s
i
v
e
p
atter
n
s
.
T
h
ese
p
atter
n
s
r
e
p
r
esen
t
co
m
m
an
d
s
to
th
e
ac
tu
ato
r
s
,
in
ter
n
al
s
tate
v
ar
iab
les,
an
d
s
en
s
o
r
d
ata.
T
h
e
m
em
o
r
y
ca
n
b
e
u
s
ed
f
o
r
p
r
ed
ictio
n
s
in
ce
it
h
as
th
e
ca
p
ac
ity
to
s
to
r
e
th
ese
s
eq
u
en
ce
s
an
d
r
et
r
iev
e
th
em
i
n
s
itu
atio
n
s
s
im
ilar
to
th
e
p
ast.
I
n
s
p
ir
ed
b
y
th
e
b
io
lo
g
ical
c
o
n
ce
p
t
o
f
ce
ll
ass
em
b
ly
,
a
s
in
g
le
n
eu
r
o
n
al
ce
ll
m
o
d
el
is
in
t
r
o
d
u
ce
d
to
f
o
r
m
a
r
tific
ial
ce
ll
ass
em
b
ly
th
at
s
to
r
es
th
e
d
ata.
C
o
m
p
a
r
e
d
to
n
e
u
r
o
n
s
,
wh
ich
ar
e
ty
p
i
ca
lly
th
o
u
g
h
t
o
f
in
ter
m
s
o
f
ar
tific
ial
ce
ll
ass
em
b
ly
m
o
d
els
o
r
n
eu
r
al
n
etwo
r
k
s
,
th
e
n
atu
r
e
o
f
a
s
in
g
le
n
e
u
r
o
n
al
ce
ll
m
o
d
el
is
ess
en
tially
d
if
f
er
en
t.
I
f
th
e
v
al
u
e
th
at
th
e
n
eu
r
al
n
etwo
r
k
s
to
r
es
is
alter
ed
,
th
e
s
y
n
ap
s
e
—
wh
ich
d
o
es
n
o
t
k
ee
p
th
e
d
ata
—
m
u
s
t
b
e
s
tr
en
g
th
en
e
d
o
r
tr
ain
ed
a
g
ain
[
3
8
]
.
W
ith
in
th
e
m
em
o
r
y
m
o
d
el,
th
e
s
in
g
le
n
eu
r
al
ce
ll
m
o
d
el
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
7
9
3
-
1
8
0
4
1798
s
eq
u
en
tially
ass
em
b
les
ce
lls
with
o
u
t
th
e
n
ee
d
f
o
r
co
m
p
licated
tr
ain
in
g
an
d
lear
n
in
g
co
m
p
u
tatio
n
s
.
T
h
e
p
u
r
p
o
s
e
o
f
th
e
co
n
n
ec
tio
n
b
et
wee
n
ce
lls
is
to
d
eter
m
in
e
wh
ich
ce
ll,
in
a
s
p
ec
if
ic
s
eq
u
en
c
ed
co
n
s
tr
u
cted
ce
ll,
m
ay
ac
tiv
ate
n
e
x
t
wh
e
n
an
o
th
er
ce
ll
d
o
es.
T
h
e
ce
ll
is
a
d
ata
co
n
tain
er
b
y
its
elf
.
B
ec
au
s
e
n
o
weig
h
ts
n
ee
d
to
b
e
r
etr
ain
ed
,
th
is
g
r
ea
tly
r
ed
u
c
es th
e
am
o
u
n
t
o
f
tr
ain
in
g
tim
e
n
ee
d
ed
wh
e
n
th
e
d
ata
n
ee
d
s
to
b
e
u
p
d
ated
.
I
n
ad
d
itio
n
,
a
p
latf
o
r
m
ca
lle
d
s
eq
u
en
tial
h
ier
ar
ch
ical
s
u
p
er
s
et
is
in
tr
o
d
u
ce
d
f
o
r
m
ate
r
ializin
g
th
e
r
ep
licatio
n
o
f
t
h
e
ac
tu
al
h
u
m
a
n
n
e
o
co
r
tex
m
em
o
r
y
.
T
h
e
id
e
a
is
to
m
im
ic
th
e
6
lay
er
h
ier
a
r
ch
ical
s
tr
u
ctu
r
e
o
f
h
u
m
an
n
eo
c
o
r
tex
b
y
h
av
in
g
a
h
ier
ar
ch
ical
s
u
p
er
s
et
im
p
lem
en
tatio
n
[
3
8
]
.
I
n
s
id
e
th
e
p
latf
o
r
m
,
a
n
ass
em
b
le
d
s
in
g
le
n
eu
r
o
n
al
ce
ll
m
o
d
el
is
p
lace
d
as
a
s
et
an
d
a
s
eq
u
e
n
ce
o
f
th
ese
s
ets
f
o
r
m
a
s
u
p
e
r
s
et
in
a
h
ier
a
r
ch
y
,
s
tar
tin
g
f
r
o
m
th
e
lo
west
lay
er
6
,
wh
ich
co
n
tain
s
s
et
o
f
ce
lls
with
a
s
p
ec
if
ic
an
d
d
etailed
d
ata,
u
n
til
th
e
h
ig
h
est
lay
er
1
wh
ich
co
n
tain
s
ce
lls
with
d
ata
th
at
ca
n
b
e
co
n
s
id
er
ed
as
o
b
ject
in
its
ab
s
tr
ac
ted
f
o
r
m
.
T
h
e
o
b
ject
in
ab
s
tr
ac
ted
f
o
r
m
is
to
r
ea
lize
o
n
e
o
f
th
e
k
ey
ch
a
r
ac
ter
is
tics
o
f
h
u
m
an
n
eo
c
o
r
tex
as
ex
p
lain
ed
in
th
e
m
em
o
r
y
-
p
r
ed
ictio
n
f
r
a
m
ewo
r
k
,
wh
ich
is
th
e
n
eo
co
r
tex
s
to
r
es
d
ata
th
at
is
in
v
ar
ian
t
r
ep
r
esen
tatio
n
[
3
8
]
.
T
h
e
o
u
tp
u
t
o
f
th
e
m
em
o
r
y
m
o
d
el
m
o
d
u
le
is
th
en
u
s
ed
f
o
r
th
e
er
r
o
r
p
r
ed
ic
tio
n
p
r
o
ce
s
s
an
d
th
e
p
r
o
b
a
b
ilit
y
th
at
th
e
ev
en
t
is
in
clu
d
ed
in
th
e
attac
k
/an
o
m
aly
ca
teg
o
r
y
.
T
h
e
p
s
eu
d
o
c
o
d
e
o
f
th
e
m
em
o
r
y
m
o
d
el
m
o
d
u
le
i
s
s
h
o
wn
in
Fig
u
r
e
3
.
Fig
u
r
e
3
.
Ps
eu
d
o
c
o
d
e
o
f
m
em
o
r
y
m
o
d
el
co
n
s
tr
u
ctio
n
2
.
4
.
E
x
perim
ent
s
ce
na
rio
Af
ter
th
e
test
b
ed
n
etwo
r
k
is
p
r
ep
ar
ed
,
th
e
s
im
u
latio
n
o
f
t
r
af
f
i
c
d
ata
t
h
at
h
as
b
ee
n
o
b
tain
ed
d
u
r
in
g
th
e
2
wee
k
s
o
f
o
b
s
er
v
atio
n
is
r
u
n
with
th
e
f
o
llo
win
g
s
ce
n
a
r
io
.
−
T
r
af
f
ic
d
ata
s
im
u
latio
n
d
esig
n
an
d
p
lan
,
.
.
tim
es
to
in
ject
an
d
d
eter
m
in
e
ty
p
e
o
f
attac
k
/an
o
m
aly
.
(
W
h
en
an
attac
k
/an
o
m
aly
o
cc
u
r
s
?
an
d
wh
at
attac
k
/an
o
m
al
y
o
cc
u
r
s
?)
.
−
Def
in
e
twelv
e
u
s
er
s
/n
o
d
es (
f
o
u
r
s
er
v
er
n
o
d
es a
n
d
eig
h
t
m
o
s
t a
ctiv
e
u
s
er
s
)
b
ased
o
n
th
e
tr
af
f
ic
v
o
lu
m
e.
−
C
r
ea
te
r
eq
u
ir
ed
n
o
r
m
al
tr
af
f
ic
p
ac
k
a
g
es
(
f
r
o
m
th
e
ca
p
tu
r
e
d
d
ataset)
as
well
as
attac
k
/an
o
m
aly
p
ac
k
ag
es
(
f
r
o
m
MA
C
C
DC
d
ataset)
to
b
e
in
jecte
d
in
to
th
e
n
etwo
r
k
.
−
Star
t in
jectin
g
tr
af
f
ic
d
ata
in
t
o
th
e
n
etwo
r
k
an
d
at
t
h
e
s
am
e
tim
e
lo
g
g
in
g
tr
af
f
ic
v
ia
p
o
r
t m
ir
r
o
r
in
g
.
−
L
ab
elin
g
s
im
u
lated
an
o
m
alies
an
d
s
p
ec
if
ic
ap
p
licatio
n
tr
af
f
ic
m
an
u
ally
is
n
ec
ess
ar
y
f
o
r
ex
p
er
im
en
ts
to
v
alid
ate
r
esu
lts
.
−
Sav
e
th
e
o
b
tain
ed
tr
af
f
ic
to
a
f
ile
in
.
f
o
r
m
at
as
a
d
ata
s
et
f
o
r
en
tity
/u
s
er
b
eh
av
io
r
an
aly
s
is
ex
p
er
im
en
ts
.
T
h
e
ex
p
e
r
im
en
ts
in
v
o
l
v
ed
in
p
u
ttin
g
r
aw
tr
af
f
ic
d
ata
f
o
r
o
n
g
o
in
g
lear
n
i
n
g
a
n
d
d
etec
tio
n
p
u
r
p
o
s
es.
I
n
th
e
L
STM
lear
n
in
g
tr
ials
,
th
e
in
itial
7
d
ay
s
o
f
tr
af
f
ic
d
ata
s
er
v
e
as
th
e
tr
ain
in
g
d
ataset,
wh
ile
th
e
f
in
al
7
d
ay
s
s
er
v
e
as th
e
test
in
g
d
ataset.
T
h
e
f
in
al
r
esu
lt is
d
eter
m
in
ed
b
y
av
er
ag
in
g
th
e
s
co
r
es o
b
tain
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
User
b
eh
a
vio
r
a
n
a
lysi
s
fo
r
in
s
id
er a
tta
ck
d
etec
tio
n
u
s
in
g
a
c
o
mb
in
a
tio
n
o
f
…
(
Ya
ya
S
u
d
a
r
ya
Tr
ia
n
a
)
1799
2
.
5
.
P
er
f
o
rma
nce
ev
a
lua
t
io
n
I
n
ev
alu
atin
g
th
e
e
x
p
er
im
e
n
tal
r
esu
lts
,
th
e
au
th
o
r
s
u
s
e
s
ev
er
al
m
etr
ics to
m
ea
s
u
r
e
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
h
ese
p
er
f
o
r
m
an
ce
m
etr
ics
in
clu
d
e:
Acc
u
r
ac
y
,
Pre
cisi
o
n
,
Sen
s
iti
v
ity
,
F1
-
s
co
r
e,
an
d
Sp
ec
if
icity
u
s
in
g
th
e
f
o
r
m
u
la
i
n
(
1
)
–
(
5
)
[
3
9
]
.
=
+
+
+
+
(
1
)
=
/
(
+
)
(
2
)
=
/
(
+
)
(
3
)
=
/
(
+
)
(
4
)
1
=
(
2
∗
(
∗
)
)
/
(
+
)
(
5
)
W
h
er
e,
=
T
r
u
e
Po
s
itiv
e,
n
am
ely
p
o
s
itiv
e
d
ata
th
at
is
d
etec
ted
co
r
r
ec
tly
.
=
T
r
u
e
Neg
ativ
e,
n
am
ely
n
eg
ati
v
e
d
ata
th
at
is
co
r
r
ec
tl
y
d
etec
t
ed
.
=
Fals
e
Po
s
i
tiv
e,
n
am
ely
n
e
g
ativ
e
d
ata
b
u
t
d
etec
ted
as p
o
s
itiv
e
d
ata.
=
Fals
e
Neg
ativ
e,
n
am
ely
p
o
s
itiv
e
d
ata
b
u
t d
etec
ted
as n
e
g
ativ
e
d
ata.
,
,
,
ar
e
r
ef
er
r
e
d
to
as
th
e
co
m
p
o
n
en
ts
o
f
th
e
co
n
f
u
s
io
n
m
atr
ix
a
n
d
th
eir
v
alu
es
ar
e
o
b
tain
ed
f
r
o
m
o
b
s
er
v
atio
n
s
d
u
r
in
g
t
h
e
ex
p
er
i
m
en
t.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
Cre
a
t
ed
da
t
a
s
et
a
nd
f
ea
t
ure
ex
t
ra
ct
io
n r
esu
lt
s
T
ab
le
1
s
h
o
ws
th
e
r
ec
ap
itu
latio
n
o
f
tr
af
f
ic
d
ata
th
at
was
s
u
cc
es
s
f
u
lly
r
ec
o
r
d
ed
d
u
r
in
g
th
e
ex
p
er
im
en
t,
an
d
th
en
th
is
tr
af
f
ic
d
ata
was
ex
t
r
ac
ted
a
n
d
s
el
ec
ted
u
s
in
g
th
e
R
FE
m
eth
o
d
.
T
h
e
cr
ea
ted
d
ataset
is
d
o
m
in
ated
b
y
tr
af
f
ic
f
r
o
m
u
s
er
s
b
r
o
wsi
n
g
web
(
7
8
.
9
1
%),
in
clu
d
in
g
attac
k
tr
af
f
ic
o
f
E
d
o
n
k
ey
(
5
.
7
5
%),
an
d
Un
k
n
o
wn
t
r
af
f
ic
(
2
.
9
%).
T
h
e
r
esu
lts
o
f
f
ea
tu
r
e
e
x
tr
ac
tio
n
ar
e
s
h
o
wn
i
n
T
ab
le
2
.
Stre
am
in
g
an
d
Gam
es
tr
af
f
ics
h
av
e
h
ig
h
n
u
m
b
er
o
f
f
ea
tu
r
es,
b
ec
au
s
e
th
e
two
a
p
p
licatio
n
s
h
av
e
m
an
y
ty
p
es
o
f
tr
af
f
ic.
B
itto
r
en
t
tr
af
f
ic
h
as
th
e
lo
west
n
u
m
b
e
r
o
f
f
ea
tu
r
es
am
o
n
g
th
e
tr
af
f
ic
c
lass
es
s
in
ce
it
is
v
er
y
s
p
ec
if
ic
ap
p
licatio
n
th
at
h
as
s
p
ec
if
ic
f
ea
tu
r
es.
E
d
o
n
k
ey
m
a
lwar
e/attac
k
tr
af
f
ic
co
n
s
is
ts
o
f
1
2
f
ea
tu
r
es.
T
ab
le
1
.
C
ap
tu
r
e
d
tr
af
f
ic
Stati
s
tics
C
l
a
s
s
F
l
o
w
s
A
p
p
l
i
c
a
t
i
o
n
p
r
o
t
o
c
o
l
Tr
a
f
f
i
c
A
m
o
u
n
t
(
%)
W
EB
7
0
2
9
0
0
0
7
8
.
9
1
B
r
o
w
si
n
g
:
H
TTP,
H
TTPS
H
TTP
-
S
TR
1
5
3
0
0
0
1
.
7
1
H
TTP St
r
e
a
mi
n
g
ED
O
N
K
EY
5
6
2
5
0
0
5
.
7
5
e
D
o
n
k
e
y
,
e
M
u
l
e
o
b
f
u
sc
a
t
e
d
B
I
TTO
R
R
EN
T
5
1
3
0
0
0
.
5
7
B
i
t
t
o
r
e
n
t
C
H
A
T
4
3
8
3
0
0
1
.
8
7
M
S
N
,
I
R
C
,
Y
a
h
o
o
M
s
n
,
H
T
TP
C
h
a
t
,
Jab
b
e
r
EM
A
I
L
5
3
3
7
0
0
4
.
5
6
S
M
TP,
H
TTP
M
a
i
l
,
P
O
P
3
,
P
O
P
3
s
,
I
M
A
P
,
I
M
A
P
s
F
TP
5
4
0
0
0
.
0
5
F
TP
-
d
a
t
a
,
F
TP c
o
n
t
r
o
l
S
TR
EA
M
I
N
G
2
5
2
0
0
0
.
2
8
M
s.
M
e
d
i
a
S
e
r
v
e
r
,
R
e
a
l
P
l
a
y
e
r
,
i
T
u
n
e
s,
Q
u
i
c
k
T
i
me
G
A
M
ES
4
5
0
0
0
.
0
5
N
F
S
3
,
H
TTP
G
a
me
s,
B
l
i
z
z
a
r
d
B
a
t
t
l
e
n
e
t
,
Q
u
a
k
e
I
I
/
I
I
I
,
C
o
u
n
t
e
r
S
t
r
i
k
e
U
N
K
N
O
W
N
1
9
7
1
0
0
2
.
1
9
N
B
S
,
M
s
-
d
s
,
E
p
ma
p
,
A
t
t
a
c
k
s
T
ab
le
2
.
Featu
r
e
e
x
tr
ac
tio
n
r
esu
lts
C
l
a
s
s
O
p
t
i
mi
z
e
d
f
e
a
t
u
r
e
s
#
C
l
a
s
s
O
p
t
i
mi
z
e
d
f
e
a
t
u
r
e
s
#
W
EB
9
EM
A
I
L
11
H
TTP
-
S
TR
9
F
TP
8
ED
O
N
K
EY
12
S
TR
EA
M
I
N
G
15
B
I
TTO
R
R
EN
T
7
G
A
M
ES
14
C
H
A
T
11
U
N
K
N
O
W
N
13
3
.
2
.
T
o
p us
er
pro
f
ili
ng
I
n
th
is
ex
p
er
im
en
t,
th
e
eig
h
t
to
p
u
s
er
s
wer
e
p
r
o
f
iled
.
T
h
e
p
r
o
f
ilin
g
r
esu
lts
s
h
o
wn
in
Fig
u
r
e
4
ar
e
tak
en
f
r
o
m
t
h
e
d
ata
o
n
th
e
a
v
er
ag
e
n
u
m
b
er
o
f
p
ac
k
ets
f
o
r
1
4
d
ay
s
o
f
o
b
s
er
v
atio
n
.
Fo
u
r
u
s
er
s
r
ep
r
esen
t
a
s
im
ilar
p
atter
n
o
f
tr
a
f
f
ic,
wh
il
e
th
e
o
th
er
f
o
u
r
s
h
o
w
an
o
th
er
s
im
ilar
tr
af
f
ic
p
atter
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
7
9
3
-
1
8
0
4
1800
Fig
u
r
e
4
.
R
esu
lt o
f
p
r
o
f
ilin
g
8
to
p
ac
tiv
e
u
s
er
s
(
d
aily
av
er
ag
e
)
3
.
3
.
Co
m
pa
riso
n wit
h s.SCASH
M
a
nd
L
ST
M
-
RF
E
m
o
dels
T
h
is
ex
p
er
im
e
n
t
aim
s
to
co
m
p
ar
e
th
e
ac
cu
r
ac
y
in
p
r
e
d
ictin
g
tr
af
f
ic
f
lo
w
b
etwe
en
th
e
MPM
-
R
FE
m
o
d
el
a
n
d
t
h
e
s
.
SC
ASHM
[
3
4
]
a
n
d
L
STM
-
R
FE
[
4
0
]
m
o
d
els.
Fo
r
th
is
p
u
r
p
o
s
e,
t
h
e
tr
af
f
ic
r
ec
o
r
d
s
in
th
e
d
ataset
ar
e
lab
ele
d
a
u
to
m
atica
lly
u
s
in
g
Py
th
o
n
p
ac
k
ag
e.
T
h
e
ex
p
e
r
im
en
tal
r
esu
lts
s
h
o
wn
in
Fig
u
r
e
5
ex
h
ib
it
th
e
s
.
SC
ASHM
an
d
M
PM
-
R
FE
m
o
d
els
ar
e
ab
le
to
m
ain
ta
in
co
n
s
is
ten
t
av
er
ag
e
ac
cu
r
ac
y
d
u
r
in
g
p
e
r
io
d
s
o
f
f
lu
ctu
atio
n
f
r
o
m
9
am
to
5
p
.
m
.
p
er
d
ay
,
wh
ile
th
e
L
STM
-
R
FE
m
o
d
el
f
ails
to
m
ain
tain
p
r
ed
ictio
n
ac
cu
r
ac
y
b
ec
au
s
e
th
is
m
o
d
el
d
o
es n
o
t
h
av
e
th
e
a
b
ilit
y
to
d
o
o
n
-
th
e
-
f
ly
lear
n
in
g
an
d
m
ay
b
e
b
ec
a
u
s
e
th
e
s
am
p
le
u
s
ed
f
o
r
tr
ain
in
g
is
n
o
t
e
n
o
u
g
h
.
Nev
er
t
h
eless
,
in
av
er
ag
e,
th
e
p
r
ed
ict
io
n
ac
cu
r
ac
y
o
f
MPM
-
R
FE
m
o
d
el
is
b
etter
th
a
n
th
e
s
.
SC
AS
HM
m
o
d
el.
T
ab
le
3
s
h
o
ws
th
e
r
esu
lts
o
f
p
er
f
o
r
m
a
n
ce
m
ea
s
u
r
em
en
t
ca
lcu
latio
n
s
f
r
o
m
th
e
d
etec
tio
n
o
f
attac
k
s
/an
o
m
alies
th
at
o
cc
u
r
i
n
th
e
s
elec
ted
ap
p
licatio
n
s
b
as
ed
o
n
c
o
n
f
u
s
io
n
m
etr
ics
o
b
s
er
v
atio
n
s
.
I
n
g
en
er
al,
th
e
MPM
-
R
FE
m
o
d
el
p
r
o
d
u
c
es
b
etter
ac
cu
r
ac
y
th
a
n
th
e
L
STM
-
R
FE
m
o
d
el.
T
h
e
ac
cu
r
a
cy
r
ea
ch
e
d
9
4
.
0
1
%
f
o
r
th
e
d
etec
tio
n
o
f
th
e
UNKN
OW
N
clas
s
.
T
h
e
p
r
o
ce
s
s
o
f
f
o
r
m
in
g
a
m
em
o
r
y
m
o
d
el
f
o
r
n
o
r
m
al
an
d
an
o
m
alo
u
s
tr
a
f
f
ic
in
Fig
u
r
e
1
was
ex
p
er
im
en
tally
o
b
s
er
v
ed
th
r
o
u
g
h
s
tatis
tics
o
n
th
e
ac
tiv
ity
o
f
ac
tiv
atin
g
ce
lls
in
m
em
o
r
y
.
T
ab
le
4
s
h
o
ws
th
e
r
esu
lts
o
f
s
tati
s
tical
o
b
s
er
v
ati
o
n
s
o
f
th
e
ac
tiv
atio
n
o
f
m
e
m
o
r
y
ce
lls
th
at
o
cc
u
r
d
u
r
i
n
g
th
e
tr
ain
in
g
o
f
th
e
p
r
o
p
o
s
ed
MPM
+RF
E
m
o
d
el.
Me
asu
r
em
en
t o
n
d
etec
tio
n
an
d
p
r
o
f
ilin
g
tim
e
is
s
h
o
wn
in
T
ab
le
5
.
Fig
u
r
e
5
.
Acc
u
r
ac
y
c
o
m
p
ar
is
o
n
: s.SC
AS
HM
v
s
L
STM
-
R
FE
v
s
MPM
-
R
FE
T
ab
le
3
.
Per
f
o
r
m
an
ce
m
etr
ics r
esu
lts
f
o
r
d
if
f
er
en
t a
p
p
licatio
n
s
(
L
STM
-
R
FE
v
s
MPM
-
R
FE)
A
p
p
l
i
c
a
t
i
o
n
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
S
e
n
s
i
t
i
v
i
t
y
(
%)
F1
-
sc
o
r
e
(
%)
S
p
e
c
i
f
i
c
i
t
y
M
P
M
-
R
F
E
LSTM
-
R
F
E
M
P
M
-
R
F
E
LSTM
-
R
F
E
M
P
M
-
R
F
E
LSTM
-
R
F
E
M
P
M
-
R
F
E
LSTM
-
R
F
E
M
P
M
-
R
F
E
LSTM
-
R
F
E
W
EB
9
0
.
1
5
7
9
.
5
4
8
8
.
5
5
7
6
.
9
6
9
8
.
6
5
9
6
.
3
2
9
1
.
8
0
8
4
.
0
0
5
7
.
9
3
4
6
.
1
8
C
H
A
T
9
0
.
1
8
7
2
.
9
9
9
0
.
3
3
7
5
.
8
8
9
8
.
5
9
9
2
.
4
3
9
3
.
1
3
8
2
.
8
7
3
0
.
0
0
2
0
.
0
7
EM
A
I
L
9
2
.
7
8
7
4
.
1
0
9
5
.
6
4
7
5
.
2
5
9
8
.
7
7
9
3
.
0
2
9
5
.
8
7
8
2
.
1
1
5
6
.
6
7
2
8
.
6
0
S
TR
EA
M
I
N
G
9
3
.
6
6
8
0
.
0
5
9
5
.
1
1
8
2
.
0
1
9
9
.
0
1
9
5
.
9
9
9
6
.
0
8
8
6
.
5
6
5
5
.
6
7
2
8
.
5
6
G
A
M
ES
9
3
.
2
2
8
1
.
5
2
9
4
.
8
8
8
2
.
1
7
9
9
.
1
4
9
6
.
2
3
9
5
.
6
5
8
7
.
3
2
5
9
.
2
7
3
8
.
9
0
U
N
K
N
O
W
N
9
4
.
0
1
9
1
.
7
6
9
3
.
7
5
7
9
.
9
8
9
9
.
2
8
9
6
.
7
8
9
5
.
4
5
8
6
.
7
7
6
0
.
3
9
4
0
.
7
9
0
10
20
30
40
50
60
70
80
90
100
00:00
01:30
03:00
04:30
06:00
07:30
09:00
10:30
12:00
13:30
15:00
16:30
18:00
19:30
21:00
22:30
Pre
d
i
c
ti
o
n
A
c
c
u
racy
(
%
)
O
b
s
e
rvati
o
n
T
i
me
s
.S
CA
S
H
M
L
S
T
M-R
F
E
MPM-R
F
E
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
User
b
eh
a
vio
r
a
n
a
lysi
s
fo
r
in
s
id
er a
tta
ck
d
etec
tio
n
u
s
in
g
a
c
o
mb
in
a
tio
n
o
f
…
(
Ya
ya
S
u
d
a
r
ya
Tr
ia
n
a
)
1801
T
ab
le
4
.
Statis
tics
o
f
m
em
o
r
y
c
el
l
ac
tiv
atio
n
d
u
r
i
n
g
th
e
e
x
p
e
r
im
en
t
Ex
p
e
r
i
m
e
n
t
n
u
mb
e
r
1
2
3
4
5
A
c
t
i
v
e
c
e
l
l
s
8
6
%
5
1
%
8
0
%
8
5
%
4
9
%
N
o
n
a
c
t
i
v
e
c
e
l
l
s
1
2
%
3
7
%
1
4
%
9%
4
1
%
D
i
e
d
c
e
l
l
s
2%
1
2
%
6%
7%
1
0
%
T
ab
le
5
.
Dete
ctio
n
p
r
o
ce
s
s
in
g
tim
e
v
s
p
r
o
f
ilin
g
tim
e
N
o
d
e
I
D
1
0
.
1
0
.
1
.
1
6
1
0
.
1
0
.
1
.
4
4
1
0
.
1
0
.
1
.
4
6
1
0
.
1
0
.
1
.
6
1
0
.
1
0
.
1
.
8
7
1
0
.
1
0
.
1
.
1
2
0
1
0
.
1
0
.
1
.
1
0
1
0
.
1
0
.
1
.
1
1
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
D
e
t
.
P
r
o
f
.
s.SC
A
S
H
M
0
.
1
2
0
.
0
1
0
.
1
4
0
.
0
1
0
.
1
4
0
.
0
1
0
.
1
5
0
.
0
1
0
.
2
7
0
.
0
1
0
.
4
4
0
.
0
3
0
.
5
1
0
.
0
6
0
.
4
9
0
.
0
5
LSTM
-
R
F
E
0
.
0
4
0
.
1
4
0
.
0
5
0
.
1
4
0
.
0
5
0
.
1
5
0
.
0
6
0
.
1
6
0
.
0
7
0
.
1
6
0
.
0
9
0
.
1
0
0
.
1
6
0
.
1
7
0
.
1
2
0
.
1
9
M
P
M
-
R
F
E
0
.
0
2
0
.
0
1
0
.
0
2
0
.
0
1
0
.
0
2
0
.
0
1
0
.
0
2
0
.
0
2
0
.
0
2
0
.
0
2
0
.
0
4
0
.
0
3
0
.
0
7
0
.
0
6
0
.
0
6
0
.
0
5
D
e
t
.
:
D
e
t
e
c
t
i
o
n
,
P
r
o
f
.
:
P
r
o
f
i
l
i
n
g
3
.
4
.
Dis
cu
s
s
io
n
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el,
MPM
-
R
FE
is
m
o
r
e
co
n
s
is
ten
t
th
an
th
e
L
STM
-
R
FE
m
o
d
el
as
s
h
o
wn
in
Fig
u
r
e
5
.
T
h
e
c
o
n
s
is
ten
t
p
er
f
o
r
m
an
ce
is
ac
h
iev
ed
b
ec
au
s
e
th
e
MPM
-
R
FE
m
o
d
el
u
s
es
s
tr
ea
m
in
g
d
ata
wh
ile
th
e
L
STM
-
R
FE
m
o
d
el
u
s
es
s
tatic
d
ata.
T
h
e
in
a
b
ilit
y
to
a
d
ap
t
to
ch
a
n
g
es
in
in
co
m
in
g
d
ata
ca
u
s
es
a
d
ec
r
ea
s
e
in
ac
c
u
r
ac
y
lev
el.
O
n
th
e
o
th
er
h
an
d
,
MPM
-
R
FE
ex
ce
ls
in
attac
k
/an
o
m
aly
d
ete
ctio
n
f
o
r
s
tr
ea
m
in
g
d
ata,
d
u
e
to
its
p
r
o
v
is
io
n
o
f
co
m
p
u
tatio
n
al
n
eu
r
o
g
e
n
etic
tr
aits
lik
e
ad
a
p
tab
ilit
y
,
s
im
p
licity
,
co
n
tin
u
o
u
s
lear
n
in
g
,
an
d
r
o
b
u
s
t
co
m
p
u
tatio
n
f
o
r
f
ac
ilit
atin
g
r
ea
l
-
tim
e
u
s
er
b
eh
av
io
r
an
aly
s
is
,
th
e
m
o
d
el
s
tan
d
s
o
u
t.
C
o
m
p
r
is
in
g
th
r
ee
p
r
im
ar
y
c
o
m
p
o
n
en
ts
,
o
n
e
o
f
wh
ich
is
th
e
m
em
o
r
y
c
o
m
p
o
n
en
t.,
th
e
co
m
p
u
tatio
n
al
co
m
p
o
n
en
t
(
p
r
e
d
ictio
n
p
r
o
ce
s
s
)
an
d
th
e
co
n
tr
o
ller
c
o
m
p
o
n
e
n
t
(
p
r
ed
ictio
n
a
n
aly
s
is
)
.
T
h
e
m
ain
s
tr
en
g
th
o
f
th
e
p
r
o
p
o
s
ed
MPM
+RF
E
m
o
d
el
lies
in
th
e
in
teg
r
atio
n
am
o
n
g
th
ese
th
r
ee
co
m
p
o
n
en
ts
.
W
e
co
m
p
ar
e
th
e
p
r
o
p
o
s
ed
MPM
-
R
FE
with
th
e
s
.
SC
AS
HM
in
ter
m
o
f
d
etec
tio
n
tim
e
an
d
p
r
o
f
ilin
g
tim
e
a
n
d
th
e
r
esu
lts
ar
e
s
h
o
wn
in
T
ab
le
5
.
T
h
e
MPM
-
R
FE
d
etec
ts
th
e
attac
k
/an
o
m
aly
an
d
p
r
o
f
iles
th
e
u
s
er
s
f
aster
th
an
th
e
s
.
SC
AS
HM
.
T
h
e
wo
r
s
t
ac
cu
r
ac
y
p
er
f
o
r
m
an
ce
o
f
th
e
MPM
-
R
FE
m
o
d
el
th
at
was
o
b
s
er
v
ed
was
9
0
.
1
5
%.
T
h
e
MPM
-
R
FE
m
o
d
el
ac
h
iev
es
f
a
s
ter
lear
n
in
g
co
n
v
er
g
en
c
e
th
a
n
th
e
L
STM
-
R
FE
m
o
d
el
b
ec
a
u
s
e
th
er
e
is
n
o
n
ee
d
to
u
p
d
ate
th
e
cu
r
r
en
t
ac
tiv
e
m
em
o
r
y
ce
lls
,
as
lo
n
g
as
a
s
ig
n
if
ican
t
ch
an
g
e
i
n
th
e
d
ata
s
tr
ea
m
d
o
es
n
o
t
o
cc
u
r
.
As
ca
n
b
e
s
ee
n
in
T
ab
le
4
,
m
e
m
o
r
y
ce
lls
ar
e
f
o
r
m
ed
wh
e
n
n
ew
in
f
o
r
m
atio
n
is
o
b
tain
ed
f
r
o
m
s
tr
ea
m
in
g
d
ata
.
W
h
ile
MPM
+
R
FE
an
d
L
ST
M+
R
FE
ar
e
p
r
o
m
is
in
g
ap
p
r
o
ac
h
es
f
o
r
p
r
e
d
ictin
g
f
u
tu
r
e
m
em
o
r
y
ac
ce
s
s
es,
th
ey
ea
ch
h
av
e
d
is
tin
ct
s
tr
en
g
th
s
an
d
wea
k
n
ess
es.
T
ab
le
6
s
u
m
m
a
r
izes th
e
an
aly
s
is
o
f
b
o
th
m
o
d
els.
T
ab
le
6
.
Su
m
m
a
r
y
o
f
an
aly
s
is
o
f
MPM
+RF
E
v
s
L
STM
+R
FE
M
P
M
+
R
F
E
LSTM
+
R
F
E
A
c
c
u
r
a
c
y
P
o
t
e
n
t
i
a
l
f
o
r
h
i
g
h
e
r
a
c
c
u
r
a
c
y
P
o
t
e
n
t
i
a
l
l
y
l
o
w
e
r
a
c
c
u
r
a
c
y
I
n
t
e
r
p
r
e
t
a
b
i
l
i
t
y
M
o
r
e
i
n
t
e
r
p
r
e
t
a
b
l
e
C
h
a
l
l
e
n
g
i
n
g
t
o
i
n
t
e
r
p
r
e
t
(
B
l
a
c
k
b
o
x
n
a
t
u
r
e
)
F
l
e
x
i
b
i
l
i
t
y
A
b
l
e
t
o
i
mp
r
o
v
e
p
e
r
f
o
r
ma
n
c
e
i
n
s
p
e
c
i
f
i
c
sc
e
n
a
r
i
o
s
Le
ss fl
e
x
i
b
i
l
i
t
y
D
a
t
a
sca
r
c
i
t
y
R
e
q
u
i
r
e
s
l
a
r
g
e
d
a
t
a
s
e
t
s
o
f
mem
o
r
y
a
c
c
e
ss
p
a
t
t
e
r
n
s
f
o
r
t
r
a
i
n
i
n
g
p
r
o
c
e
ss
R
e
q
u
i
r
e
s s
mal
l
e
r
d
a
t
a
s
e
t
s
o
f
mem
o
r
y
a
c
c
e
ss
p
a
t
t
e
r
n
s f
o
r
t
r
a
i
n
i
n
g
p
r
o
c
e
ss
A
d
a
p
t
a
b
i
l
i
t
y
A
d
a
p
t
a
b
l
e
t
o
d
i
f
f
e
r
e
n
t
d
o
m
a
i
n
s
A
d
a
p
t
a
b
l
e
t
o
d
i
f
f
e
r
e
n
t
d
o
m
a
i
n
s
Li
mi
t
e
d
a
v
a
i
l
a
b
i
l
i
t
y
Th
e
mo
d
e
l
s
a
r
e
l
e
ss
c
o
mm
o
n
l
y
u
se
d
W
e
l
l
-
e
st
a
b
l
i
s
h
e
d
a
n
d
w
i
d
e
l
y
u
se
d
I
n
g
en
er
al,
ch
o
o
s
in
g
t
h
e
b
est
ap
p
r
o
ac
h
d
e
p
en
d
s
o
n
th
e
s
p
ec
if
ic
r
eq
u
ir
em
en
ts
a
n
d
p
r
io
r
i
ties
o
f
th
e
ap
p
licatio
n
.
I
f
m
a
x
im
izin
g
ac
cu
r
ac
y
an
d
in
ter
p
r
etab
ilit
y
ar
e
cr
u
cial,
an
d
th
e
ch
allen
g
es
o
f
d
ata
s
ca
r
city
an
d
lim
ited
av
ailab
ilit
y
ca
n
b
e
ad
d
r
ess
ed
,
a
m
em
o
r
y
p
r
e
d
ictio
n
m
o
d
el
with
f
ea
tu
r
e
an
aly
s
is
m
ig
h
t
b
e
a
b
etter
ch
o
ice.
Ho
wev
e
r
,
if
estab
lis
h
ed
tech
n
o
lo
g
y
,
ad
a
p
tab
ilit
y
,
a
n
d
wid
er
r
eso
u
r
ce
av
ailab
ilit
y
ar
e
m
o
r
e
im
p
o
r
tan
t
co
n
s
id
er
atio
n
s
,
an
L
STM
with
f
ea
tu
r
e
an
aly
s
is
m
ig
h
t b
e
a
s
u
itab
le
s
o
lu
tio
n
.
I
t is im
p
o
r
tan
t
to
n
o
te
th
at
n
eith
er
ap
p
r
o
ac
h
is
d
ef
in
itiv
ely
“b
ett
er
”
as
th
e
o
p
tim
al
c
h
o
ice
d
ep
en
d
s
o
n
t
h
e
s
p
ec
if
ic
ap
p
licatio
n
an
d
its
p
r
io
r
ities
.
C
o
m
b
in
in
g
elem
e
n
ts
o
f
b
o
th
ap
p
r
o
ac
h
es
o
r
e
x
p
lo
r
i
n
g
o
t
h
er
em
e
r
g
in
g
tech
n
i
q
u
es
m
i
g
h
t
also
b
e
wo
r
th
in
v
esti
g
atin
g
d
ep
e
n
d
in
g
o
n
s
p
ec
if
ic
n
ee
d
s
an
d
r
esear
ch
g
o
als.
C
o
m
b
in
in
g
m
em
o
r
y
p
r
ed
icti
o
n
m
o
d
els
with
th
e
R
FE
f
o
r
u
s
er
b
eh
a
v
io
r
an
aly
s
is
s
o
u
n
d
s
lik
e
a
p
r
o
m
is
in
g
a
p
p
r
o
ac
h
,
a
n
d
ac
h
i
ev
in
g
h
ig
h
ac
c
u
r
ac
y
is
d
ef
in
it
ely
a
p
o
s
itiv
e
o
u
tco
m
e.
Ho
we
v
er
,
it
is
im
p
o
r
tan
t
to
co
n
s
id
er
th
e
r
am
if
icatio
n
s
o
f
th
e
f
in
d
in
g
s
b
ey
o
n
d
ju
s
t
ac
cu
r
ac
y
in
clu
d
in
g
Fra
u
d
d
etec
tio
n
,
i.e
.
:
m
o
r
e
ac
cu
r
ate
u
s
er
b
eh
a
v
io
r
m
o
d
els
ca
n
h
elp
id
e
n
tify
u
n
u
s
u
al
ac
tiv
ities
th
at
m
ig
h
t
in
d
icate
f
r
au
d
u
len
t
tr
an
s
ac
tio
n
s
o
r
ac
co
u
n
t
co
m
p
r
o
m
is
es;
User
ex
p
er
ien
ce
im
p
r
o
v
em
en
t,
i.e
.
h
ig
h
ly
ac
c
u
r
ate
u
s
er
b
eh
av
io
r
an
aly
s
is
ca
n
lead
to
b
etter
p
er
s
o
n
aliza
tio
n
an
d
r
ec
o
m
m
en
d
atio
n
s
ac
r
o
s
s
v
ar
i
o
u
s
p
latf
o
r
m
s
;
E
n
h
an
ce
d
tar
g
etin
g
,
i.e
.
:
tar
g
eted
ad
v
er
tis
in
g
b
ased
o
n
ac
cu
r
ate
u
s
er
p
r
ed
ictio
n
s
ca
n
b
e
m
o
r
e
e
f
f
ec
tiv
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
7
9
3
-
1
8
0
4
1802
B
u
ild
in
g
o
n
th
e
s
u
cc
ess
with
a
h
ig
h
ly
ac
c
u
r
ate
u
s
er
b
e
h
av
io
r
a
n
aly
s
is
m
o
d
el
u
s
in
g
m
em
o
r
y
p
r
ed
ictio
n
a
n
d
R
FE,
s
o
m
e
ar
e
as a
r
e
id
en
tifie
d
th
at
will b
e
v
alu
ab
le
f
o
r
f
u
tu
r
e
ad
v
a
n
ce
m
en
ts
,
in
clu
d
e:
a.
C
o
n
tex
t
awa
r
en
ess
:
th
e
f
u
tu
r
e
lies
in
in
co
r
p
o
r
atin
g
m
o
r
e
d
ata
s
o
u
r
ce
s
b
ey
o
n
d
tr
a
d
itio
n
al
u
s
er
ac
tio
n
s
,
in
clu
d
in
g
u
s
er
d
em
o
g
r
ap
h
ics,
s
en
tim
en
t a
n
aly
s
is
f
r
o
m
tex
t d
ata,
wh
ich
lead
s
to
m
o
r
e
n
u
an
ce
d
p
r
ed
ictio
n
s
.
b.
C
o
n
tin
u
o
u
s
lear
n
i
n
g
an
d
im
p
r
o
v
em
en
t:
d
e
v
elo
p
s
tr
ateg
ies
f
o
r
co
n
tin
u
o
u
s
lear
n
in
g
an
d
i
m
p
r
o
v
em
e
n
t.
T
h
is
m
ig
h
t in
v
o
l
v
e
in
co
r
p
o
r
atin
g
n
ew
d
ata
s
o
u
r
ce
s
,
u
s
er
f
ee
d
b
ac
k
lo
o
p
s
,
o
r
r
etr
ain
in
g
t
h
e
m
o
d
el
p
er
io
d
ically
to
ad
ap
t to
ev
o
lv
in
g
u
s
er
b
e
h
av
io
r
p
atter
n
s
.
c.
Mo
d
el
g
en
e
r
aliza
b
ilit
y
:
en
s
u
r
e
th
e
m
o
d
el
p
e
r
f
o
r
m
s
well
o
n
u
n
s
ee
n
d
ata
(
g
en
er
aliza
b
ilit
y
)
.
Utilize
r
ig
o
r
o
u
s
test
in
g
with
d
iv
er
s
e
d
atasets
to
m
ain
tain
ac
cu
r
ac
y
in
r
ea
l
-
wo
r
ld
s
ce
n
ar
io
s
.
d.
R
ea
l
-
tim
e
p
er
s
o
n
aliza
tio
n
an
d
r
ec
o
m
m
en
d
atio
n
s
y
s
tem
s
:
lev
er
ag
e
th
e
m
o
d
el
f
o
r
r
ea
l
-
tim
e
p
er
s
o
n
aliza
tio
n
th
at
ca
n
s
ig
n
if
ican
tly
e
n
h
an
ce
u
s
er
ex
p
er
ien
ce
ac
r
o
s
s
v
ar
io
u
s
ap
p
licatio
n
s
.
e.
E
th
ical
co
n
s
id
er
atio
n
s
a
n
d
b
i
as
m
itig
atio
n
:
eth
ical
co
n
s
id
e
r
atio
n
s
b
ec
o
m
e
p
ar
am
o
u
n
t
wh
en
th
e
m
em
o
r
y
p
r
ed
ictio
n
m
o
d
el
b
ec
o
m
es
m
o
r
e
p
o
wer
f
u
l.
Fu
tu
r
e
s
u
cc
ess
will
in
v
o
lv
e
p
r
o
ac
tiv
ely
ad
d
r
ess
in
g
p
o
ten
tial
b
iases
in
th
e
d
ata
an
d
tr
ain
in
g
p
r
o
ce
s
s
.
Fair
n
ess
m
etr
ics
im
p
lem
en
tatio
n
will
en
s
u
r
e
th
at
th
e
m
o
d
el
tr
ea
ts
all
u
s
er
s
eq
u
itab
ly
.
T
h
e
p
r
o
p
o
s
ed
MPM
-
R
FE
m
o
d
el
is
n
o
t
d
esig
n
ed
f
o
r
lo
n
g
-
ter
m
d
ep
en
d
e
n
t
lear
n
in
g
o
f
h
ig
h
-
o
r
d
er
m
em
o
r
y
s
ets,
b
ec
au
s
e
it
will
r
eq
u
ir
e
a
lo
n
g
p
r
o
ce
s
s
in
g
tim
e,
as
r
eq
u
ir
ed
b
y
t
h
e
h
ier
ar
c
h
ical
tem
p
o
r
al
m
em
o
r
y
m
o
d
el
[
3
4
]
.
T
h
is
f
ac
t
m
ay
b
e
co
n
s
id
er
ed
as
lim
itatio
n
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
.
C
o
n
s
id
e
r
in
g
th
e
p
r
o
ce
s
s
in
g
tim
e
an
d
co
m
p
le
x
ity
o
f
th
e
m
o
d
el,
th
e
MPM
-
R
FE
m
o
d
el
h
as
th
e
p
o
ten
tial
to
b
e
ad
o
p
ted
as
a
way
to
an
aly
ze
th
e
b
eh
av
i
o
r
en
tity
to
p
r
e
v
en
t
in
s
id
er
s
’
attac
k
s
in
r
ea
l
-
tim
e
f
ash
io
n
.
4.
CO
NCLU
SI
O
N
I
n
th
is
p
a
p
er
,
th
e
a
u
th
o
r
s
h
av
e
in
tr
o
d
u
ce
d
th
e
co
m
b
in
e
d
u
s
e
o
f
R
FE
as
a
to
o
l
f
o
r
f
ea
tu
r
e
s
elec
tio
n
an
d
m
em
o
r
y
p
r
ed
ictio
n
m
o
d
el
(
MPM
-
R
FE)
to
d
etec
t
an
o
m
a
lo
u
s
tr
af
f
ic/cy
b
er
attac
k
s
in
v
o
l
v
in
g
i
n
s
id
er
s
.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
s
h
o
w
th
at
th
e
MPM
-
R
FE
m
o
d
el
i
s
ca
p
ab
le
o
f
d
em
o
n
s
tr
atin
g
th
e
d
etec
tio
n
o
f
attac
k
s
/an
o
m
aly
in
v
o
lv
in
g
in
s
id
er
s
with
an
ac
cu
r
ac
y
r
ate
o
f
u
p
to
9
4
.
0
1
%
f
o
r
th
e
d
et
ec
tio
n
o
f
u
n
k
n
o
wn
tr
af
f
ic
class
es.
I
n
g
en
er
al,
MPM
-
R
FE
ac
h
iev
es
an
ac
cu
r
ac
y
o
f
9
0
.
1
5
%
to
9
4
.
0
1
%.
B
esid
es
th
at,
MPM
-
R
FE
is
also
ab
le
to
p
r
o
v
i
d
e
b
etter
ac
cu
r
ac
y
th
an
th
e
L
STM
-
R
FE
m
o
d
el,
b
ec
a
u
s
e
it
is
ab
le
to
ca
r
r
y
o
u
t
o
n
-
t
h
e
-
f
ly
lear
n
in
g
,
s
o
th
at
th
e
s
y
s
tem
c
an
r
ec
o
g
n
ize
n
ew
p
atter
n
s
o
f
attac
k
s
/an
o
m
alies.
T
h
er
ef
o
r
e,
MPM
-
R
FE
ca
n
b
e
im
p
lem
en
ted
as
a
s
u
b
-
s
y
s
tem
to
s
u
p
p
o
r
t
a
n
in
tellig
en
t
an
d
h
o
lis
tic
cy
b
er
s
ec
u
r
ity
p
latf
o
r
m
,
wh
ich
is
b
ein
g
d
ev
elo
p
e
d
at
th
e
Netwo
r
k
e
d
C
o
m
p
u
tin
g
L
ab
,
De
f
en
s
e
Ma
th
e
m
atics
Stu
d
y
Pro
g
r
am
,
Def
en
s
e
Un
iv
er
s
ity
.
T
h
is
p
latf
o
r
m
is
p
r
o
jecte
d
to
b
e
u
s
ed
f
o
r
b
o
th
g
o
v
er
n
m
en
t
a
n
d
p
r
iv
ate
in
s
titu
tio
n
s
.
T
h
e
lim
itatio
n
s
o
f
th
e
MPM
-
R
FE
m
o
d
el
ar
e
th
e
g
en
e
r
al
lim
itatio
n
s
o
f
th
e
m
ac
h
in
e
l
ea
r
n
in
g
m
o
d
el,
b
ec
au
s
e
th
is
m
o
d
el
in
cl
u
d
es
a
n
u
n
s
u
p
er
v
is
ed
,
all
-
tim
e,
an
d
c
o
n
tin
u
o
u
s
lear
n
in
g
ap
p
r
o
ac
h
,
wh
er
e
th
e
MPM
-
R
FE
ca
r
r
ies
o
u
t
lear
n
in
g
f
r
o
m
u
s
er
d
ata
with
n
o
r
m
al
b
eh
a
v
io
r
.
Ov
er
all,
th
e
s
tate
o
f
th
e
ar
t
in
r
ea
l
-
tim
e
attac
k
/an
o
m
aly
d
et
ec
tio
n
f
o
r
n
etwo
r
k
s
ec
u
r
ity
is
p
r
o
m
is
in
g
,
with
co
n
tin
u
o
u
s
im
p
r
o
v
em
en
ts
in
ac
cu
r
ac
y
,
ef
f
icien
c
y
,
an
d
ad
ap
tab
ilit
y
.
Ho
wev
e
r
,
th
er
e
ar
e
s
till
o
n
g
o
in
g
ch
allen
g
es lik
e
b
alan
cin
g
tr
ad
eo
f
f
s
an
d
k
ee
p
in
g
p
ac
e
with
e
v
o
lv
in
g
t
h
r
ea
ts
.
R
esear
ch
an
d
d
ev
elo
p
m
e
n
t e
f
f
o
r
ts
ar
e
ac
tiv
ely
ex
p
lo
r
in
g
n
ew
tech
n
iq
u
es a
n
d
ap
p
r
o
ac
h
es to
en
h
an
ce
th
e
ef
f
ec
tiv
en
ess
an
d
r
e
liab
ilit
y
o
f
r
ea
l
-
tim
e
attac
k
/
an
o
m
aly
d
etec
tio
n
in
s
ec
u
r
in
g
n
etwo
r
k
s
.
W
h
ile
p
r
o
m
is
in
g
ad
v
an
ce
m
e
n
ts
ar
e
b
ei
n
g
m
ad
e,
m
em
o
r
y
p
r
ed
ictio
n
f
r
am
ewo
r
k
s
ar
e
s
till
u
n
d
er
ac
tiv
e
d
ev
elo
p
m
en
t.
R
esear
ch
is
o
n
g
o
in
g
to
o
v
e
r
c
o
m
e
lim
itatio
n
s
in
ac
cu
r
ac
y
,
g
en
er
aliza
b
ilit
y
,
a
n
d
in
ter
p
r
eta
b
ilit
y
.
As
th
e
f
ield
ev
o
lv
es,
we
ca
n
e
x
p
ec
t
to
s
e
e
th
ese
f
r
am
ewo
r
k
s
b
ec
o
m
e
i
n
cr
ea
s
in
g
ly
in
teg
r
at
ed
in
to
v
ar
io
u
s
co
m
p
u
tin
g
s
y
s
tem
s
to
im
p
r
o
v
e
p
e
r
f
o
r
m
an
ce
an
d
r
eso
u
r
ce
u
tili
za
tio
n
.
Sig
n
if
ican
t
ad
v
a
n
ce
m
en
ts
in
m
o
d
el
ar
c
h
itectu
r
es,
h
ar
d
war
e
u
tili
za
tio
n
,
an
d
tr
an
s
f
er
lear
n
i
n
g
tech
n
iq
u
es
ar
e
am
o
n
g
f
o
cu
s
ar
ea
in
t
h
e
n
ea
r
f
u
tu
r
e.
T
h
e
UB
A
is
a
r
ap
id
ly
e
v
o
lv
in
g
f
ield
with
co
n
tin
u
o
u
s
ad
v
an
ce
m
e
n
ts
in
tech
n
o
lo
g
y
an
d
ca
p
ab
ilit
ies.
B
y
lev
er
ag
in
g
ad
v
an
ce
d
an
aly
tics
,
b
eh
av
i
o
r
al
b
io
m
etr
ics,
an
d
r
is
k
-
b
ased
ass
ess
m
en
ts
,
U
B
A
em
p
o
wer
s
o
r
g
an
izatio
n
s
to
p
r
o
ac
tiv
ely
d
etec
t
an
d
r
esp
o
n
d
to
in
s
id
er
th
r
ea
ts
,
co
m
p
r
o
m
is
ed
ac
co
u
n
ts
,
an
d
o
th
er
s
o
p
h
is
ticated
cy
b
e
r
-
atta
ck
s
.
Ho
wev
er
,
en
s
u
r
in
g
d
ata
p
r
iv
ac
y
,
ad
d
r
ess
in
g
f
alse
p
o
s
itiv
es,
an
d
s
tay
in
g
ah
ea
d
o
f
ev
o
l
v
in
g
th
r
ea
ts
r
em
ain
o
n
g
o
in
g
ch
allen
g
es
th
at
r
esea
r
ch
er
s
an
d
s
ec
u
r
ity
p
r
o
f
ess
io
n
als ar
e
ac
tiv
ely
a
d
d
r
ess
in
g
.
RE
F
E
R
E
NC
E
S
[
1
]
H
.
Ei
c
h
e
n
b
a
u
m,
“
M
e
m
o
r
y
s
y
st
e
ms,
”
Wi
l
e
y
I
n
t
e
rd
i
sc
i
p
l
i
n
a
ry
R
e
v
i
e
w
s:
C
o
g
n
i
t
i
v
e
S
c
i
e
n
c
e
,
v
o
l
.
1
,
n
o
.
4
,
p
p
.
4
7
8
–
4
9
0
,
M
a
r
.
2
0
1
0
,
d
o
i
:
1
0
.
1
0
0
2
/
w
c
s.
4
9
.
[
2
]
H
.
Z
h
a
n
g
,
M
.
W
a
n
g
,
L
.
Y
a
n
g
,
a
n
d
H
.
Z
h
u
,
“
A
n
o
v
e
l
u
ser
b
e
h
a
v
i
o
r
a
n
a
l
y
si
s
a
n
d
p
r
e
d
i
c
t
i
o
n
a
l
g
o
r
i
t
h
m
b
a
s
e
d
o
n
m
o
b
i
l
e
s
o
c
i
a
l
e
n
v
i
r
o
n
m
e
n
t
,
”
Wi
re
l
e
ss
N
e
t
w
o
r
k
s
,
v
o
l
.
2
5
,
n
o
.
2
,
p
p
.
7
9
1
–
8
0
3
,
O
c
t
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
1
2
7
6
-
0
1
7
-
1
5
9
2
-
0.
[
3
]
D
.
S
t
i
a
w
a
n
,
A
.
H
.
A
b
d
u
l
l
a
h
,
a
n
d
M
.
Y
.
I
d
r
i
s,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
o
f
h
a
b
i
t
u
a
l
a
c
t
i
v
i
t
i
e
s
i
n
b
e
h
a
v
i
o
r
-
b
a
s
e
d
n
e
t
w
o
r
k
d
e
t
e
c
t
i
o
n
,
”
J
o
u
rn
a
l
o
f
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
,
n
o
.
8
,
p
p
.
1
–
7
,
2
0
1
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.