I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI)
Vo
l.
9
,
No
.
3
,
Sep
tem
b
er
2020
,
p
p
.
5
53
~
56
0
I
SS
N:
2
2
5
2
-
8938
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
ai.
v
9
.i
3
.
p
p
5
53
-
56
0
553
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Para
llel proces
sin
g
using
big da
ta a
nd
m
a
chine
lea
rn
ing
techniqu
es for
int
rusio
n det
e
ction
Ala
eddin
e
B
o
u
k
ha
lf
a
1
,
Na
bil
H
m
i
na
2
,
H
a
bib
a
Cha
o
ui
3
1
,
3
S
y
ste
m
En
g
in
e
e
rin
g
Lab
o
ra
to
ry
,
A
DSI
T
e
a
m
,
Na
ti
o
n
a
l
S
c
h
o
o
l
o
f
A
p
p
li
e
d
S
c
ien
c
e
s,
Ib
n
T
o
f
a
il
Un
i
v
e
rsit
y
,
M
o
ro
c
c
o
2
S
u
lt
a
n
M
o
u
lay
S
li
m
a
n
e
Un
iv
e
rsity
,
Be
n
i
M
e
ll
a
l,
M
o
r
o
c
c
o
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
1
,
2
0
20
R
ev
i
s
ed
J
u
n
2
2
,
2
0
20
A
cc
ep
ted
J
u
l 9
,
2
0
20
Cu
rre
n
tl
y
,
in
f
o
r
m
a
ti
o
n
tec
h
n
o
l
o
g
y
is
u
se
d
in
a
ll
th
e
li
fe
d
o
m
a
in
s.
M
a
n
y
d
e
v
ice
s
a
n
d
e
q
u
i
p
m
e
n
t
p
ro
d
u
c
e
d
a
ta
a
n
d
tra
n
sf
e
r
th
e
m
a
c
ro
ss
th
e
n
e
tw
o
rk
.
T
h
e
se
tran
s
f
e
rs
a
re
n
o
t
a
lw
a
y
s
s
e
c
u
re
d
a
n
d
c
a
n
c
o
n
tain
n
e
w
m
e
n
a
c
e
s
a
n
d
a
tt
a
c
k
s
in
v
isib
le
b
y
th
e
c
u
rre
n
t
se
c
u
rit
y
to
o
ls.
M
o
re
o
v
e
r,
th
e
larg
e
a
m
o
u
n
t
a
n
d
v
a
riet
y
o
f
th
e
e
x
c
h
a
n
g
e
d
d
a
ta
m
a
k
e
th
e
id
e
n
ti
f
ica
ti
o
n
o
f
th
e
in
tr
u
sio
n
s
m
o
re
d
iff
icu
lt
in
term
s
o
f
d
e
tec
ti
o
n
ti
m
e
.
T
o
so
lv
e
th
e
se
issu
e
s,
w
e
su
g
g
e
st
in
th
is
p
a
p
e
r,
a
n
e
w
a
p
p
ro
a
c
h
b
a
se
d
o
n
st
o
r
in
g
th
e
lar
g
e
a
m
o
u
n
t
a
n
d
v
a
riety
o
f
n
e
tw
o
rk
tra
ff
i
c
d
a
ta
e
m
p
lo
y
in
g
b
ig
d
a
ta
t
e
c
h
n
iq
u
e
s,
a
n
d
a
n
a
ly
z
in
g
th
e
se
d
a
ta
w
it
h
m
a
c
h
in
e
lea
rn
in
g
a
lg
o
rit
h
m
s,
in
a
d
istri
b
u
ted
a
n
d
p
a
ra
ll
e
l
w
a
y
,
in
o
rd
e
r
to
d
e
tec
t
n
e
w
h
id
d
e
n
i
n
tru
si
o
n
s
w
it
h
les
s
p
ro
c
e
ss
in
g
ti
m
e
.
A
c
c
o
rd
in
g
to
t
h
e
re
su
lt
s
o
f
th
e
e
x
p
e
ri
m
e
n
ts,
th
e
d
e
tec
ti
o
n
a
c
c
u
ra
c
y
o
f
th
e
m
a
c
h
in
e
lea
rn
in
g
m
e
th
o
d
s
re
a
c
h
e
s
u
p
to
9
9
.
9
%
,
a
n
d
th
e
ir
p
ro
c
e
ss
in
g
ti
m
e
h
a
s b
e
e
n
re
d
u
c
e
d
c
o
n
sid
e
ra
b
ly
b
y
a
p
p
l
y
in
g
th
e
m
i
n
a
p
a
ra
ll
e
l
a
n
d
d
istri
b
u
te
d
w
a
y
,
w
h
ich
p
ro
v
e
s
th
a
t
o
u
r
p
r
o
p
o
se
d
m
o
d
e
l
is
v
e
r
y
e
ff
e
c
ti
v
e
f
o
r
th
e
d
e
tec
ti
o
n
o
f
n
e
w
h
id
d
e
n
i
n
tru
sio
n
s.
K
ey
w
o
r
d
s
:
B
ig
d
ata
I
n
tr
u
s
io
n
d
etec
tio
n
Ma
ch
i
n
e
lear
n
i
n
g
P
ar
allel
p
r
o
ce
s
s
in
g
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
:
A
laed
d
in
e
B
o
u
k
h
alf
a,
S
y
s
te
m
E
n
g
in
ee
r
i
n
g
L
ab
o
r
ato
r
y
,
A
DSI
T
ea
m
,
Natio
n
al
Sc
h
o
o
l o
f
A
p
p
lied
Scien
ce
s
,
I
b
n
T
o
f
ail
Un
iv
er
s
i
t
y
,
Ke
n
itra
,
Mo
r
o
cc
o
.
E
m
ail: a
laed
d
in
e.
b
o
u
k
h
a
lf
a
@
g
m
a
il.c
o
m
1.
I
NT
RO
D
UCT
I
O
N
No
w
ad
a
y
s
,
in
f
o
r
m
a
tio
n
tec
h
n
o
lo
g
y
i
s
e
m
p
lo
y
ed
in
a
ll
ar
ea
s
o
f
li
f
e
(
f
i
n
a
n
ce
,
ed
u
ca
tio
n
,
w
ea
t
h
er
,
etc.
.
.
)
,
v
ar
io
u
s
eq
u
i
p
m
e
n
t,
n
a
m
el
y
,
co
m
p
u
ter
s
,
s
er
v
er
s
,
tab
lets
a
n
d
o
th
er
s
d
ev
ices,
ar
e
p
r
o
d
u
cin
g
co
n
tin
u
o
u
s
l
y
d
ata
a
n
d
e
x
ch
an
g
i
n
g
i
t
t
h
r
o
u
g
h
t
h
e
n
et
wo
r
k
.
Ho
w
e
v
er
,
t
h
ese
ex
c
h
a
n
g
es
b
et
w
ee
n
th
e
s
e
eq
u
ip
m
e
n
ts
ar
e
n
o
t
al
w
a
y
s
s
e
cu
r
ed
,
an
d
t
h
e
y
ca
n
co
n
tai
n
n
e
w
h
id
d
en
attac
k
s
.
W
h
ile
t
h
e
ex
i
s
tin
g
to
o
ls
an
d
s
tr
ateg
ie
s
o
f
s
ec
u
r
it
y
ar
e
esta
b
lis
h
ed
o
n
p
r
ed
ef
i
n
ed
m
et
h
o
d
s
an
d
alg
o
r
it
h
m
s
to
id
en
t
if
y
i
n
tr
u
s
io
n
s
,
th
e
y
d
o
n
'
t
h
av
e
t
h
e
ab
ilit
y
to
d
etec
t
n
e
w
th
r
ea
ts
.
T
h
is
p
u
s
h
es
u
s
to
th
i
n
k
ab
o
u
t
n
e
w
m
e
th
o
d
s
an
d
te
ch
n
iq
u
es
t
h
at
ca
n
ev
o
lv
e
to
d
is
clo
s
e
n
e
w
m
e
n
ac
es.
I
n
ad
d
itio
n
,
th
e
d
ata
p
ass
in
g
th
r
o
u
g
h
t
h
e
n
et
w
o
r
k
is
s
o
lar
g
e
an
d
ca
n
b
e
o
f
s
e
v
er
al
t
y
p
es,
w
h
ic
h
p
r
o
v
o
k
e
d
if
f
ic
u
ltie
s
r
ela
ted
to
th
r
ea
ts
d
etec
tio
n
ti
m
e
f
o
r
th
e
cu
r
r
en
t
d
ev
ice
s
o
f
s
ec
u
r
it
y
.
W
ith
t
h
e
f
a
s
t
g
r
o
w
t
h
o
f
t
h
e
g
en
er
ated
d
ata
li
k
e
v
id
eo
s
,
s
o
u
n
d
s
,
e
m
ail
s
,
etc.
.
.
,
i
n
all
s
ec
to
r
s
,
th
e
o
ld
d
ata
m
an
a
g
e
m
e
n
t
to
o
ls
h
a
v
e
b
ec
o
m
e
o
b
s
o
lete,
t
h
e
y
ar
e
n
o
t
ab
le
to
s
to
r
e
o
r
m
a
n
ag
e
th
is
lar
g
e
a
m
o
u
n
t
o
f
d
ata,
a
s
a
co
n
s
eq
u
e
n
ce
,
a
n
e
w
co
n
ce
p
t
w
as
co
n
ce
i
v
ed
ca
lled
B
ig
Data
to
d
e
f
in
e
n
e
w
r
u
les
f
o
r
t
h
e
m
an
a
g
e
m
en
t
a
n
d
s
to
r
ag
e
o
f
th
i
s
lar
g
e
m
a
s
s
o
f
d
ata.
Als
o
,
a
co
llectio
n
o
f
cla
s
s
i
f
ica
tio
n
m
eth
o
d
s
o
f
th
e
Ma
ch
i
n
e
L
ea
r
n
in
g
d
o
m
ai
n
h
av
e
ap
p
ea
r
ed
r
ec
en
tl
y
,
a
n
d
w
a
s
ass
o
ciate
d
w
it
h
t
h
is
n
e
w
b
o
r
n
B
ig
Da
ta,
in
o
r
d
er
to
ex
tr
ac
t
in
v
i
s
ib
le
in
f
o
r
m
atio
n
f
r
o
m
it.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
5
53
–
56
0
554
T
h
er
ef
o
r
e,
to
d
ea
l
w
ith
i
s
s
u
e
s
cited
ab
o
v
e,
w
e
s
u
g
g
es
t
in
th
is
p
ap
er
a
n
e
w
id
ea
to
an
al
y
ze
an
d
ev
alu
a
te
n
et
w
o
r
k
tr
af
f
ic,
b
as
ed
o
n
co
llectin
g
an
d
s
to
r
in
g
its
h
u
g
e
d
ata
e
m
p
lo
y
i
n
g
b
i
g
d
ata
tech
n
iq
u
es,
an
d
ap
p
ly
in
g
p
a
r
allel
p
r
o
ce
s
s
in
g
o
f
clas
s
i
f
icatio
n
a
lg
o
r
it
h
m
s
o
n
th
e
s
e
d
ata,
in
o
r
d
er
to
d
etec
t
n
e
w
h
id
d
en
attac
k
s
w
i
th
le
s
s
ti
m
e
co
n
s
u
m
p
tio
n
.
T
h
e
s
tu
d
y
h
a
s
t
h
e
f
o
llo
w
i
n
g
s
tr
u
ct
u
r
e,
w
e
g
i
v
e
a
s
u
m
m
ar
y
o
f
r
elate
d
w
o
r
k
i
n
Sectio
n
2
,
Sectio
n
3
is
r
e
s
er
v
ed
f
o
r
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
,
ex
p
er
im
en
tal
e
n
v
ir
o
n
m
en
t
is
d
escr
ib
ed
in
Sec
tio
n
4
,
Sectio
n
5
is
d
ed
icate
d
f
o
r
r
esu
lts
an
d
an
a
l
y
s
is
,
s
ec
t
io
n
6
an
n
o
u
n
ce
s
t
h
e
co
n
cl
u
s
io
n
an
d
f
u
t
u
r
e
w
o
r
k
.
2.
RE
L
AT
E
D
WO
RK
S
T
h
e
id
ea
o
f
in
tr
u
s
io
n
d
etec
tio
n
i
n
a
b
ig
d
ata
e
n
v
ir
o
n
m
en
t
was
d
eb
ated
in
p
ap
e
r
[
1
]
,
th
e
au
th
o
r
s
h
a
v
e
s
h
o
w
n
t
h
at
cu
r
r
e
n
t
r
esear
ch
is
n
o
t
v
er
y
in
ter
ested
i
n
m
et
h
o
d
s
th
at
ca
n
e
x
tr
ac
t
i
n
f
o
r
m
at
io
n
ab
o
u
t
t
h
e
an
o
m
alie
s
f
o
u
n
d
,
an
d
th
e
y
a
f
f
ir
m
ed
t
h
at
t
h
is
i
n
f
o
r
m
atio
n
i
s
v
er
y
u
s
e
f
u
l
f
o
r
r
ed
u
cin
g
t
h
e
ti
m
e
b
et
w
ee
n
t
h
e
id
en
ti
f
icatio
n
o
f
t
h
e
an
o
m
al
y
an
d
th
e
r
ea
ctio
n
.
Usi
n
g
m
ac
h
i
n
e
lear
n
i
n
g
m
eth
o
d
s
,
t
h
e
y
p
r
o
p
o
s
ed
a
n
e
w
m
et
h
o
d
ca
lled
m
u
lt
iv
ar
iate
b
i
g
d
ata
an
al
y
s
i
s
(
MB
DA
)
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
,
b
ased
o
n
f
iv
e
s
ta
g
e
s
,
th
e
f
ir
s
t
s
tep
is
to
tr
an
s
f
o
r
m
th
e
s
o
u
r
ce
d
ata
in
to
q
u
an
titati
v
e
ch
ar
ac
ter
is
t
ics,
th
e
s
ec
o
n
d
s
tep
is
to
c
r
ea
te
a
s
in
g
le
d
ata
f
lo
w
f
r
o
m
ch
ar
ac
ter
is
tic
s
o
f
m
u
ltip
le
s
o
u
r
ce
s
o
f
d
ata,
th
e
th
ir
d
s
tep
is
to
f
in
d
th
e
a
n
o
m
alie
s
at
th
e
r
ig
h
t
m
o
m
en
t,
th
e
f
o
u
r
th
s
tep
is
to
id
e
n
ti
f
y
t
h
e
f
ea
t
u
r
es o
f
t
h
e
a
n
o
m
al
y
,
th
e
f
i
f
th
a
n
d
last
s
tep
is
to
d
etec
t a
n
d
p
r
esen
t t
h
e
d
ata
ca
u
s
i
n
g
th
e
a
n
o
m
al
y
.
T
h
e
p
r
o
p
o
s
ed
id
ea
is
r
em
ar
k
ab
le,
ex
c
ep
t
th
at
th
e
a
u
t
h
o
r
s
d
id
n
o
t
p
r
esen
t
h
o
w
a
n
d
w
h
y
th
e
y
u
s
ed
t
h
e
m
ac
h
i
n
e
lear
n
i
n
g
m
et
h
o
d
s
.
An
o
th
er
co
n
ce
p
t
f
o
r
an
al
y
zin
g
t
h
e
b
i
g
tr
af
f
ic
o
f
t
h
e
n
et
w
o
r
k
w
a
s
d
is
c
u
s
s
ed
in
p
ap
er
[
2
]
,
th
e
a
u
t
h
o
r
s
co
n
f
ir
m
ed
t
h
at
t
h
e
n
et
w
o
r
k
t
r
af
f
ic
is
v
er
y
lar
g
e,
w
h
ic
h
p
u
s
h
es
to
f
in
d
n
e
w
m
ea
n
s
ab
l
e
to
d
etec
t
t
h
r
ea
ts
w
it
h
p
r
ec
is
io
n
.
T
h
e
y
s
u
g
g
e
s
te
d
a
s
et
o
f
m
et
h
o
d
s
o
f
a
n
al
y
s
is
u
s
in
g
R
lan
g
u
ag
e
to
r
e
m
ed
y
p
r
o
b
lem
s
r
elate
d
to
th
e
v
o
lu
m
e,
v
er
ac
i
t
y
,
a
n
d
v
ar
i
et
y
o
f
lar
g
e
a
m
o
u
n
ts
o
f
d
ata.
T
o
test
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
s
,
th
e
y
ap
p
lied
th
e
m
o
n
th
e
KDD
C
u
p
9
9
d
ataset
[
3
]
,
w
h
ich
i
s
an
ea
r
lier
v
er
s
i
o
n
o
f
NS
L
KDD
[
4
]
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
s
ar
e
in
ter
esti
n
g
,
ex
ce
p
t
th
a
t
th
e
y
a
r
e
o
r
ien
ted
to
w
ar
d
s
i
m
p
r
o
v
i
n
g
th
e
q
u
a
lit
y
o
f
th
e
lar
g
e
q
u
a
n
tit
y
o
f
d
ata
an
d
n
o
t
an
ac
tio
n
a
g
ain
s
t t
h
r
ea
ts
.
On
e
m
o
r
e
ap
p
r
o
ac
h
to
d
etec
t
an
o
m
alie
s
i
n
a
telec
o
m
m
u
n
ica
tio
n
d
o
m
ai
n
w
as
d
eb
ated
b
y
t
h
e
au
t
h
o
r
s
in
th
e
m
a
n
u
s
cr
ip
t
[
5
]
,
th
e
y
a
d
m
it
th
a
t
cu
r
r
en
t
l
y
t
h
er
e
is
a
lar
g
e
a
m
o
u
n
t
o
f
t
h
e
p
r
o
d
u
ce
d
d
ata
b
y
m
u
lt
ip
le
d
ev
ices,
t
h
e
y
p
r
o
p
o
s
ed
a
n
e
w
s
y
s
te
m
b
ased
o
n
f
o
u
r
s
tep
s
,
f
ir
s
tl
y
i
t
co
llects
d
ata,
s
ec
o
n
d
l
y
it
p
r
ep
ar
es
d
ata
b
ef
o
r
e
tr
ea
t
m
en
t,
t
h
ir
d
l
y
i
t
ap
p
lies
an
u
n
s
u
p
er
v
i
s
ed
clu
s
ter
in
g
al
g
o
r
ith
m
,
f
i
n
all
y
it
s
h
o
w
s
t
h
e
f
o
u
n
d
an
o
m
alie
s
.
T
h
e
s
u
g
g
es
ted
s
y
s
t
e
m
co
n
s
t
itu
te
s
a
n
e
w
ap
p
r
o
ac
h
to
v
is
u
alize
an
o
m
alies
b
u
t
it
is
n
o
t
d
ed
icate
d
to
d
etec
t n
e
w
a
ttack
s
.
I
n
p
ap
er
[
6
]
,
th
e
au
t
h
o
r
s
d
is
c
u
s
s
a
n
e
w
id
ea
o
f
r
ed
u
ci
n
g
th
e
ti
m
e
o
f
d
etec
tio
n
o
f
in
tr
u
s
io
n
s
i
n
B
i
g
Data
en
v
ir
o
n
m
en
t,
t
h
e
y
cla
i
m
th
at
attac
k
s
i
n
cr
ea
s
e
as
t
h
e
q
u
an
t
it
y
o
f
p
r
o
d
u
ce
d
d
ata
in
cr
ea
s
es,
also
th
e
lar
g
e
a
m
o
u
n
t
o
f
d
ata
ca
u
s
es
d
i
f
f
icu
ltie
s
r
elate
d
to
th
e
d
u
r
at
io
n
o
f
t
h
e
a
n
al
y
s
i
s
f
o
r
in
t
r
u
s
io
n
s
d
etec
tio
n
,
th
e
y
p
r
o
p
o
s
ed
a
n
e
w
ar
c
h
ite
ctu
r
e
b
ased
o
n
a
d
is
tr
ib
u
ted
s
tr
ea
m
i
n
g
p
latf
o
r
m
ca
l
led
A
p
ac
h
e
Kaf
k
a
[
7
]
,
an
d
a
co
m
p
o
n
e
n
t
o
f
Sp
ar
k
[
8
]
u
s
ed
f
o
r
d
ata
s
tr
ea
m
p
r
o
ce
s
s
i
n
g
ca
lled
Sp
ar
k
S
tr
ea
m
i
n
g
[
9
]
.
T
h
e
m
o
d
el
g
r
o
u
p
s
lo
ad
in
g
n
et
w
o
r
k
tr
af
f
ic
f
r
o
m
C
SV
f
ile,
in
g
esti
n
g
d
ata
u
s
in
g
Ka
f
k
a,
p
r
o
c
ess
i
n
g
d
ata
u
s
i
n
g
A
p
ac
h
e
Sp
ar
k
Stre
a
m
i
n
g
.
T
h
e
ex
p
er
i
m
en
t
w
as
r
ea
lized
an
d
s
h
o
w
ed
g
o
o
d
r
esu
lts
.
T
h
e
p
r
o
p
o
s
ed
id
ea
r
ep
r
esen
ts
a
p
r
o
m
is
i
n
g
ap
p
r
o
ac
h
to
r
ed
u
ce
p
r
o
ce
s
s
in
g
t
i
m
e
i
n
a
b
i
g
d
ata
e
n
v
ir
o
n
m
en
t,
b
u
t
th
e
a
u
t
h
o
r
s
d
id
n
o
t
s
p
ec
if
y
th
e
u
s
ed
m
eth
o
d
s
f
o
r
in
tr
u
s
io
n
s
d
etec
tio
n
.
I
n
th
e
s
a
m
e
b
ig
d
ata
e
n
v
ir
o
n
m
e
n
t,
a
n
o
th
er
s
y
s
te
m
to
id
en
ti
f
y
in
tr
u
s
io
n
b
ased
o
n
a
m
et
h
o
d
o
f
s
tatis
t
ics
ca
lled
f
i
n
ite
D
ir
ich
l
et
m
i
x
tu
r
e
m
o
d
el
w
a
s
p
r
o
p
o
s
ed
in
p
a
p
er
[
1
0
]
,
th
e
au
th
o
r
s
an
n
o
u
n
ce
d
th
at
a
s
y
s
te
m
th
a
t
d
etec
ts
n
o
th
r
ea
t
s
p
er
d
ay
is
o
b
s
o
lete,
t
h
e
y
h
av
e
s
et
u
p
a
n
e
w
f
r
a
m
e
w
o
r
k
co
m
p
o
s
ed
o
f
th
r
e
e
co
m
p
o
n
e
n
t
s
,
th
e
f
ir
s
t
co
m
p
o
n
en
t
ca
p
tu
r
e
an
d
lo
g
n
et
w
o
r
k
d
ata,
th
e
s
ec
o
n
d
co
m
p
o
n
e
n
t
p
er
f
o
r
m
s
a
n
al
y
s
i
s
an
d
f
iltra
tio
n
o
p
er
atio
n
s
to
p
r
ep
ar
e
th
e
d
ata
f
o
r
t
h
e
n
e
x
t
co
m
p
o
n
en
t,
th
e
th
ir
d
an
d
las
t
co
m
p
o
n
en
t
is
d
ed
icate
d
to
ap
p
ly
D
ir
ich
let
m
i
x
t
u
r
e
m
o
d
el
m
et
h
o
d
in
o
r
d
er
to
d
etec
t
in
tr
u
s
io
n
s
.
T
h
e
s
y
s
te
m
h
as
b
ee
n
test
ed
o
n
t
w
o
d
atasets
NS
L
K
DD
[
4
]
an
d
UNSW
-
NB
1
5
[
1
1
]
,
th
e
in
tr
u
s
i
o
n
d
etec
tio
n
r
ates
w
er
e
h
ig
h
.
T
h
e
p
r
o
p
o
s
al
co
n
s
tit
u
tes
a
r
e
m
ar
k
ab
le
n
e
w
ap
p
r
o
ac
h
f
o
r
th
e
d
etec
tio
n
o
f
i
n
tr
u
s
io
n
s
,
alt
h
o
u
g
h
it
s
p
er
f
o
r
m
a
n
ce
s
h
a
v
e
n
o
t
b
ee
n
co
m
p
ar
ed
to
th
at
o
f
th
e
o
th
er
m
et
h
o
d
s
.
T
h
e
s
y
s
te
m
i
s
n
o
t d
ed
icate
d
to
ev
o
lv
e
to
d
etec
t n
e
w
th
r
ea
t
s
.
Th
e
au
th
o
r
s
ex
p
o
s
e,
in
th
e
s
t
u
d
y
[
1
2
]
,
a
n
e
w
ap
p
r
o
ac
h
to
an
al
y
ze
th
e
b
ig
tr
a
f
f
ic
o
f
t
h
e
n
et
w
o
r
k
,
t
h
e
y
d
ec
lar
e
th
at
th
e
e
x
is
t
in
g
to
o
ls
o
f
s
ec
u
r
it
y
m
u
s
t
a
n
al
y
ze
th
e
c
o
llected
d
ata
in
o
r
d
er
to
ev
o
lv
e
to
ca
tch
th
r
ea
ts
,
th
e
y
s
u
g
g
e
s
ted
a
d
is
tr
ib
u
ted
ar
ch
itect
u
r
e
i
n
clo
u
d
co
m
p
u
ti
n
g
ar
ea
,
f
o
r
t
h
e
p
u
r
p
o
s
e
o
f
e
x
p
lo
itin
g
t
h
e
g
r
ea
t
ca
p
ac
it
y
o
f
s
to
r
ag
e
a
n
d
p
r
o
c
ess
i
n
g
,
t
h
e
ar
c
h
itect
u
r
e
is
e
s
tab
lis
h
ed
o
n
co
llect
in
g
tr
a
f
f
i
c
o
f
th
e
n
et
w
o
r
k
,
s
to
r
in
g
it,
a
n
d
an
al
y
zi
n
g
it
e
m
p
lo
y
i
n
g
t
h
e
f
a
m
o
u
s
p
ar
allel
p
r
o
ce
s
s
in
g
p
r
o
g
r
a
m
Ma
p
R
ed
u
ce
.
T
h
e
au
th
o
r
s
h
a
v
e
n
o
t c
ar
r
ied
o
u
t e
x
p
er
i
m
e
n
ts
to
v
alid
ate
th
e
ap
p
r
o
ac
h
.
A
ls
o
,
th
e
id
ea
is
n
o
t scala
b
le
to
d
etec
t n
e
w
t
h
r
ea
ts
.
Ou
r
ap
p
r
o
ac
h
is
to
tall
y
d
i
f
f
er
en
t
co
m
p
ar
ed
to
th
e
cited
w
o
r
k
s
.
I
t
f
o
c
u
s
es
o
n
d
etec
ti
n
g
n
e
w
h
id
d
en
in
tr
u
s
io
n
s
w
it
h
les
s
p
r
o
ce
s
s
i
n
g
ti
m
e,
in
a
n
en
v
ir
o
n
m
e
n
t
w
h
er
e
th
e
d
ata
is
v
er
y
lar
g
e
a
n
d
v
ar
ied
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
P
a
r
a
llel
p
r
o
ce
s
s
in
g
u
s
in
g
b
ig
d
a
ta
a
n
d
ma
c
h
in
e
lea
r
n
in
g
tec
h
n
iq
u
es fo
r
... (
A
la
ed
d
i
n
e
B
o
u
k
h
a
lfa
)
555
3.
P
RO
P
O
SE
D
AP
P
RO
ACH
Ou
r
s
u
g
g
e
s
ted
ap
p
r
o
ac
h
is
illu
s
tr
ated
in
Fi
g
u
r
e
1
,
it
i
s
an
ar
ch
itec
tu
r
e
o
f
a
lo
ca
l
b
u
s
i
n
es
s
n
et
w
o
r
k
th
a
t
i
s
co
m
p
o
s
ed
o
f
f
o
u
r
m
ai
n
co
m
p
o
n
e
n
ts
,
n
a
m
e
l
y
,
a
co
llecto
r
,
an
ex
tr
ac
t
-
tr
a
n
s
f
o
r
m
lo
ad
(
E
T
L
)
,
a
b
ig
d
ata
clu
s
ter
,
an
d
an
a
n
al
y
s
i
s
m
ac
h
i
n
e.
Fig
u
r
e
1
.
B
ig
d
ata
an
d
m
ac
h
i
n
e
lear
n
in
g
ar
ch
itec
tu
r
e
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
3
.
1
.
Co
llect
o
r
T
h
e
co
llecto
r
is
a
tr
af
f
ic
lis
te
n
er
,
it
is
a
s
o
f
t
w
ar
e
th
a
t
co
llect
th
e
tr
a
f
f
ic
p
ass
in
g
t
h
r
o
u
g
h
th
e
n
et
w
o
r
k
,
it
is
i
n
s
tal
led
o
n
a
n
et
w
o
r
k
m
ac
h
in
e,
i
t
lis
te
n
s
,
ca
p
tu
r
es,
an
d
s
av
es
n
et
w
o
r
k
tr
a
f
f
ic
o
n
t
h
e
s
a
m
e
m
ac
h
in
e
i
n
o
r
d
er
t
o
lo
ad
it to
th
e
b
ig
d
ata
clu
s
ter
v
ia
th
e
E
T
L
.
3
.
2
.
E
x
t
ra
ct
t
r
a
ns
f
o
r
m
lo
a
d
(
E
T
L
)
An
E
T
L
is
a
s
o
f
t
w
ar
e
th
at
ai
m
s
to
ex
tr
ac
t
d
ata
f
r
o
m
a
s
o
u
r
ce
,
tr
an
s
f
o
r
m
it,
t
h
en
l
o
ad
it
to
a
d
esti
n
atio
n
[
1
3
]
,
s
o
,
it
is
in
s
t
alled
o
n
th
e
s
a
m
e
m
ac
h
i
n
e
o
f
th
e
co
llecto
r
,
it
is
r
esp
o
n
s
i
b
le
f
o
r
lo
ad
in
g
th
e
ca
u
g
h
t
tr
a
f
f
ic
f
r
o
m
t
h
e
n
et
w
o
r
k
b
y
t
h
e
co
llecto
r
,
to
th
e
b
ig
d
ata
clu
s
ter
.
3
.
3
.
B
ig
da
t
a
clus
t
er
B
ec
au
s
e
o
f
t
h
e
lar
g
e
a
m
o
u
n
t
an
d
v
ar
iet
y
o
f
tr
a
f
f
ic
d
ata
e
x
ch
an
g
ed
all
t
h
e
ti
m
e
b
et
w
ee
n
th
e
lo
ca
l
n
et
w
o
r
k
an
d
th
e
I
n
ter
n
et,
w
e
h
av
e
s
et
u
p
a
B
ig
Data
clu
s
ter
.
T
h
e
tw
o
m
o
s
t
u
s
ed
b
ig
d
ata
m
an
a
g
e
m
e
n
t
f
r
a
m
e
w
o
r
k
s
ar
e
Had
o
o
p
[
1
4
]
an
d
Sp
ar
k
[
8
]
,
th
ey
ar
e
co
m
p
o
s
ed
o
f
t
w
o
co
m
p
o
n
en
ts
,
th
e
f
ir
s
t
ca
lled
Had
o
o
p
d
is
tr
ib
u
ted
f
ile
s
y
s
te
m
(
H
DFS
)
is
r
eser
v
ed
f
o
r
s
to
r
in
g
d
ata,
th
e
s
ec
o
n
d
is
r
eser
v
ed
f
o
r
d
is
t
r
ib
u
ted
p
r
o
ce
s
s
in
g
o
f
d
ata
v
i
a
th
e
Ma
p
R
ed
u
ce
p
r
o
g
r
a
m
[
1
5
]
.
W
e
u
s
ed
Had
o
o
p
b
ec
au
s
e
it
is
m
o
r
e
p
o
w
er
f
u
l
t
h
an
Sp
ar
k
in
ter
m
s
o
f
d
ata
s
ec
u
r
it
y
[
1
6
]
.
3
.
4
.
Ana
ly
s
is
m
a
chi
ne
Du
e
to
th
e
lar
g
e
a
m
o
u
n
t
a
n
d
v
ar
iet
y
o
f
d
ata
t
h
at
ca
n
b
e
co
l
lecte
d
ac
r
o
s
s
t
h
e
n
et
w
o
r
k
,
it
h
as
b
ec
o
m
e
d
if
f
ic
u
lt
to
p
r
o
ce
s
s
th
e
m
w
it
h
t
h
e
o
ld
an
al
y
s
i
s
m
e
th
o
d
s
an
d
to
o
ls
o
f
s
ec
u
r
it
y
[
17]
,
c
o
n
tr
ar
i
w
is
e
,
Ma
ch
i
n
e
L
ea
r
n
in
g
m
eth
o
d
s
h
av
e
th
e
ca
p
ac
it
y
to
ex
tr
ac
t
i
n
f
o
r
m
at
io
n
h
id
d
en
i
n
th
is
lar
g
e
v
o
lu
m
e
an
d
v
ar
iet
y
o
f
d
ata
[
1
8
]
,
th
at
's
w
h
y
w
e
w
il
l
u
s
e
th
e
m
to
p
r
o
ce
s
s
n
et
w
o
r
k
tr
a
f
f
ic.
So
,
t
h
e
a
n
al
y
s
i
s
m
ac
h
i
n
e
i
s
al
s
o
a
m
ac
h
in
e
o
n
t
h
e
lo
ca
l
n
et
w
o
r
k
,
o
n
w
h
ic
h
w
e
h
a
v
e
i
n
s
tall
ed
s
o
f
t
w
ar
e
t
h
at
w
i
ll
la
u
n
ch
Ma
ch
i
n
e
L
ea
r
n
in
g
alg
o
r
ith
m
s
,
i
n
o
r
d
er
to
p
r
o
ce
s
s
th
e
d
ata
alr
ea
d
y
s
to
r
ed
in
th
e
B
ig
Data
clu
s
ter
.
4.
E
XP
E
R
I
M
E
NT
A
L
E
NV
I
R
O
NM
E
NT
I
n
th
is
p
ar
t,
w
e
p
r
esen
t
t
h
e
u
s
ed
m
et
h
o
d
s
f
o
r
th
e
an
al
y
s
i
s
,
th
e
ch
o
s
e
n
d
ata
f
o
r
th
e
ex
p
er
im
en
tatio
n
,
th
e
v
al
id
atio
n
m
eth
o
d
,
th
e
e
v
a
lu
atio
n
m
etr
ic
s
,
an
d
th
e
w
o
r
k
en
v
ir
o
n
m
e
n
t.
4
.
1
.
Ana
ly
s
is
m
et
ho
d
s
T
h
er
e
ar
e
s
ev
er
al
m
ac
h
i
n
e
l
ea
r
n
in
g
m
et
h
o
d
s
,
s
o
it
's
n
o
t
ea
s
y
to
t
est
th
e
m
a
ll,
w
e
tr
ied
to
tes
t
o
n
l
y
t
h
e
m
o
s
t
k
n
o
w
n
an
d
u
s
e
d
o
f
th
e
m
,
w
h
ic
h
ar
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
S
VM
)
[
1
9
]
,
K
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
[
2
0
]
,
an
d
d
ec
is
io
n
tr
ee
[
2
1
]
.
Su
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
SVM)
:
it
is
a
m
ac
h
i
n
e
lear
n
in
g
m
eth
o
d
,
w
h
ic
h
is
in
t
en
d
ed
to
s
o
lv
e
b
in
ar
y
a
n
d
m
u
ltip
le
cla
s
s
i
f
ica
tio
n
p
r
o
b
lem
s
,
it
i
s
b
ased
o
n
m
ar
g
i
n
s
,
it
ta
k
es
f
e
w
s
a
m
p
le
s
an
d
it
ac
h
iev
e
s
g
o
o
d
r
esu
lts
[
2
2
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
5
53
–
56
0
556
K
-
n
ea
r
est
n
ei
g
h
b
o
r
s
(
KNN)
:
it
is
an
ef
f
ec
ti
v
e
m
et
h
o
d
o
f
m
a
ch
in
e
lear
n
i
n
g
t
h
at
is
ap
p
lied
to
class
if
ica
tio
n
an
d
r
eg
r
es
s
io
n
p
r
o
b
le
m
s
.
T
o
esti
m
ate
t
h
e
o
u
tp
u
t
a
s
s
o
ciate
d
w
it
h
a
n
e
w
i
n
p
u
t
X,
(
KNN)
c
o
n
s
is
ts
i
n
tak
i
n
g
in
to
ac
co
u
n
t th
e
K
tr
ai
n
i
n
g
s
a
m
p
les
w
h
o
s
e
i
n
p
u
t is clo
s
est t
o
th
e
n
e
w
i
n
p
u
t X
[
2
3
]
.
Dec
is
io
n
tr
ee
:
it
is
a
m
et
h
o
d
o
f
d
e
cisi
o
n
m
a
k
i
n
g
a
n
d
class
if
icatio
n
,
th
e
d
if
f
er
en
t
d
ec
is
io
n
s
p
o
s
s
ib
le
ar
e
lo
ca
ted
at
th
e
ter
m
in
al
n
o
d
es
(
w
h
ich
r
ep
r
ese
n
t
t
h
e
lea
v
es
o
f
t
h
e
tr
ee
)
an
d
ar
e
o
b
tain
ed
a
cc
o
r
d
in
g
to
th
e
d
ec
is
io
n
s
r
ea
ch
ed
at
ea
c
h
s
ta
g
e
[
2
4
]
.
4
.
2
.
Da
t
a
s
et
T
o
ev
alu
ate
o
u
r
ap
p
r
o
ac
h
,
w
e
ch
o
s
e
th
e
f
a
m
o
u
s
NS
L
K
DD
d
ataset
[
4
]
,
w
h
ic
h
is
an
ad
v
an
ce
d
v
er
s
io
n
o
f
KDD
C
u
p
9
9
[
3
]
.
NSL
K
DD
g
at
h
er
s
w
i
th
o
u
t
r
ed
u
n
d
an
c
y
n
et
w
o
r
k
tr
af
f
ic
d
ata
f
r
o
m
a
m
ilit
ar
y
en
v
ir
o
n
m
e
n
t,
it i
s
co
m
p
o
s
ed
o
f
n
o
r
m
a
l a
n
d
attac
k
r
ec
o
r
d
s
,
n
a
m
el
y
:
Do
S (
Den
ial
-
of
-
Ser
v
ice)
: T
h
is
m
ak
e
s
th
e
s
er
v
ice
u
n
a
v
ailab
le
.
P
r
o
b
e:
w
h
ic
h
tr
ies to
d
is
clo
s
e
in
f
o
r
m
atio
n
ab
o
u
t a
n
et
w
o
r
k
a
n
d
f
i
n
d
s
y
s
te
m
v
u
l
n
er
ab
ilit
ies.
U2
R
(
User
to
R
o
o
t)
:
w
h
ich
p
r
o
f
it f
r
o
m
v
u
l
n
er
ab
ilit
ies i
n
th
e
s
y
s
te
m
to
g
et
s
u
p
er
u
s
er
p
r
iv
il
eg
es.
R
2
L
(
R
e
m
o
te
to
L
o
ca
l)
:
w
h
ic
h
tr
ies
to
at
tack
a
m
ac
h
in
e
an
d
ca
u
s
es
v
u
l
n
er
ab
ilit
ies
to
o
b
tai
n
s
ec
u
r
e
in
f
o
r
m
a
tio
n
.
T
ab
le
s
1
-
3
r
e
p
r
esen
t
th
e
n
u
m
b
er
o
f
r
ec
o
r
d
s
f
o
r
ea
ch
t
y
p
e.
T
a
b
le
1
s
h
o
w
s
t
h
e
d
is
tr
ib
u
tio
n
o
f
t
h
e
d
ataset
in
t
w
o
cla
s
s
e
s
.
T
ab
le
2
s
h
o
w
s
t
h
e
d
is
tr
ib
u
tio
n
o
f
t
h
e
d
ataset
in
f
i
v
e
clas
s
es,
a
n
d
T
ab
le
3
s
h
o
w
s
t
h
e
d
is
tr
ib
u
tio
n
o
f
th
e
d
ataset
i
n
t
w
e
n
t
y
-
th
r
ee
clas
s
es.
T
ab
l
e
1
.
Dis
tr
ib
u
tio
n
o
f
d
atase
t in
t
w
o
class
e
s
T
r
a
f
f
i
c
N
u
mb
e
r
o
f
sam
p
l
e
s
N
o
r
mal
6
7
3
4
3
A
t
t
a
c
k
5
8
6
3
0
T
o
t
a
l
1
2
5
9
7
3
T
ab
le
2
.
Dis
tr
ib
u
tio
n
o
f
d
atase
t in
f
iv
e
cla
s
s
es
T
r
a
f
f
i
c
N
u
mb
e
r
o
f
sam
p
l
e
s
N
o
r
mal
6
7
3
4
3
A
t
t
a
c
k
D
o
S
4
5
9
2
7
P
r
o
b
e
1
1
6
5
6
R
2
L
9
9
5
U
2
R
52
T
o
t
a
l
1
2
5
9
7
3
T
ab
le
3
.
Dis
tr
ib
u
tio
n
o
f
d
atase
t in
t
w
e
n
t
y
-
t
h
r
ee
clas
s
es
T
r
a
f
f
i
c
N
u
mb
e
r
o
f
sam
p
l
e
s
N
o
r
mal
6
7
3
4
3
A
t
t
a
c
k
N
EPT
U
N
E
(
D
O
S
)
4
1
2
1
4
S
A
TA
N
(
P
r
o
b
e
)
3
6
3
3
I
P
S
W
EEP (P
r
o
b
e
)
3
5
9
9
P
O
R
T
S
W
EEP (P
r
o
b
e
)
2
9
3
1
S
M
U
R
F
(
D
o
S
)
2
6
4
6
N
M
A
P
(
P
r
o
b
e
)
1
4
9
3
B
A
C
K
(
D
o
S
)
9
5
6
T
E
A
R
D
R
O
P
(
D
o
S
)
8
9
2
W
A
R
EZ
C
L
I
EN
T
(
R
2
L
)
8
9
0
P
O
D
(
D
o
S
)
2
0
1
G
U
ESS
_
P
A
S
S
W
D
(
R
2
L
)
53
B
U
F
F
ER
_
O
V
ER
F
L
O
W
(
U
2
R
)
30
W
A
R
EZ
M
A
S
T
ER
(
R
2
L
)
20
L
A
N
D
(
D
o
S
)
18
I
M
A
P
(
R
2
L
)
11
R
O
O
TK
I
T
(
U
2
R
)
10
L
O
A
D
M
O
D
U
L
E
(
U
2
R
)
9
F
T
P
_
W
R
I
T
E
(
R
2
L
)
8
M
U
L
TI
H
O
P
(
R
2
L
)
7
P
H
F
(
R
2
L
)
4
P
ER
L
(
U
2
R
)
3
S
P
Y
(
R
2
L
)
2
T
o
t
a
l
1
2
5
9
7
3
4
.
3
.
Va
lid
a
t
i
o
n
m
et
ho
d
T
o
ass
ess
o
u
r
m
o
d
el,
w
e
c
h
o
s
e
t
h
e
cr
o
s
s
-
v
a
lid
atio
n
m
eth
o
d
,
it
is
a
tec
h
n
iq
u
e
t
h
a
t
ass
es
s
es
th
e
d
etec
tio
n
ca
p
ac
it
y
o
f
a
class
if
ier
b
y
d
i
v
id
in
g
th
e
d
ata
s
et
i
n
to
t
w
o
s
u
b
s
ets,
t
h
e
tr
ai
n
i
n
g
s
u
b
s
et
an
d
th
e
test
s
u
b
s
et.
Firstl
y
,
th
e
class
if
ier
is
tr
ain
ed
o
n
th
e
tr
ain
i
n
g
d
ataset,
s
ec
o
n
d
l
y
,
it
is
ap
p
lied
o
n
th
e
test
d
ataset
i
n
o
r
d
er
to
m
ea
s
u
r
e
its
d
e
g
r
ee
o
f
s
u
cc
es
s
.
T
h
e
p
r
o
ce
s
s
is
r
ep
ea
ted
N
ti
m
es
i
n
d
ep
en
d
en
tl
y
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
P
a
r
a
llel
p
r
o
ce
s
s
in
g
u
s
in
g
b
ig
d
a
ta
a
n
d
ma
c
h
in
e
lea
r
n
in
g
tec
h
n
iq
u
es fo
r
... (
A
la
ed
d
i
n
e
B
o
u
k
h
a
lfa
)
557
th
e
av
er
ag
e
o
f
t
h
e
N
p
er
f
o
r
m
an
ce
s
is
r
etu
r
n
ed
.
T
h
e
s
tr
o
n
g
p
o
in
t
o
f
th
is
tec
h
n
iq
u
e
is
th
at
all
d
ata
is
u
s
ed
f
o
r
tr
ain
i
n
g
a
n
d
test
in
g
,
w
h
ic
h
m
ak
e
s
t
h
e
as
s
es
s
m
e
n
t
m
o
r
e
p
r
ec
is
e.
W
e
em
p
lo
y
ed
5
-
f
o
ld
cr
o
s
s
v
alid
atio
n
to
ass
es
s
o
u
r
ap
p
r
o
ac
h
,
if
w
e
i
n
cr
ea
s
e
th
e
N,
th
e
n
u
m
b
er
o
f
attac
k
s
f
o
r
s
o
m
e
t
y
p
e
s
li
k
e
R
2
L
a
n
d
U2
R
w
ill
d
ec
r
ea
s
e
f
o
r
ea
ch
s
u
b
s
et,
a
n
d
th
e
y
m
a
y
b
e
n
eg
lec
ted
d
u
r
in
g
p
r
o
ce
s
s
in
g
[
2
5
]
.
4
.
4
.
E
v
a
lua
t
i
o
n
m
et
rics
T
o
ass
ess
th
e
d
etec
t
io
n
e
f
f
ic
ien
c
y
o
f
o
u
r
p
r
o
p
o
s
ed
alg
o
r
ith
m
s
,
w
e
c
h
o
o
s
e
t
h
e
m
etr
ics
ac
cu
r
ac
y
,
s
en
s
iti
v
it
y
,
s
p
ec
i
f
icit
y
,
f
al
s
e
p
o
s
itiv
e
r
ate
(
FP
R
)
an
d
ar
ea
u
n
d
er
cu
r
v
e
(
AUC),
th
e
d
ef
i
n
itio
n
s
o
f
th
e
s
e
m
etr
ics ar
e:
A
cc
u
r
ac
y
=
T
P
+
T
N/(
T
P
+
T
N+
FP
+FN)
(
1
)
I
t r
ep
r
esen
ts
th
e
f
r
ac
tio
n
o
f
tr
u
e
id
en
tific
atio
n
o
v
er
all
d
ata
i
n
s
tan
ce
s
.
Sen
s
iti
v
it
y
=
T
P
/(
T
P
+FN)
(
2
)
I
t
is
al
s
o
ca
lled
tr
u
e
p
o
s
i
tiv
e
r
ate
(
T
P
R
)
,
it
m
ea
s
u
r
es
t
h
e
r
atio
o
f
p
o
s
iti
v
e
i
n
s
ta
n
ce
s
t
h
at
ar
e
co
r
r
ec
tly
cla
s
s
i
f
ied
.
Sp
ec
if
icit
y
=
T
N/(
T
N+
F
P
)
(
3
)
I
t
m
ea
s
u
r
es th
e
r
atio
o
f
n
eg
a
ti
v
e
in
s
ta
n
ce
s
t
h
at
ar
e
co
r
r
ec
tly
class
i
f
ied
.
Fals
e
P
o
s
itiv
e
R
ate
=
FP
/(
FP
+T
N)
(
4
)
I
t r
ep
r
esen
ts
th
e
p
r
o
b
ab
ilit
y
o
f
f
alsel
y
r
ej
ec
tin
g
t
h
e
n
u
ll
h
y
p
o
th
esi
s
.
A
U
C
is
t
h
e
p
r
o
b
a
b
ilit
y
th
at
a
r
an
d
o
m
l
y
s
elec
ted
p
o
s
itiv
e
in
s
ta
n
ce
is
r
an
k
ed
h
i
g
h
er
th
a
n
a
r
an
d
o
m
l
y
s
elec
ted
n
e
g
ati
v
e
i
n
s
ta
n
ce
.
I
t
is
t
h
e
m
ea
s
u
r
e
m
e
n
t
o
f
t
h
e
s
u
r
f
ac
e
ar
ea
u
n
d
er
th
e
r
ec
ei
v
er
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
cu
r
v
e
w
h
ich
p
lo
ts
th
e
tr
u
e
p
o
s
iti
v
e
r
ate
(
T
P
R
)
ag
ain
s
t t
h
e
f
a
ls
e
p
o
s
iti
v
e
r
ate
(
FP
R
)
.
T
P
,
T
N,
FP
an
d
FN
ar
e
ex
tr
ac
ted
f
r
o
m
t
h
e
co
n
f
u
s
io
n
m
a
tr
ix
af
ter
th
e
cla
s
s
i
f
icatio
n
o
p
er
atio
n
,
th
e
y
m
ea
n
r
esp
ec
tiv
el
y
,
T
r
u
e
P
o
s
itiv
e,
T
r
u
e
Neg
at
iv
e,
Fal
s
e
P
o
s
itiv
e
a
n
d
Fals
e
Neg
ati
v
e.
T
r
u
e
Po
s
itiv
e
(
T
P)
in
d
icate
s
th
e
n
u
m
b
er
o
f
in
s
ta
n
ce
s
co
r
r
ec
tly
clas
s
if
ied
.
Fals
e
P
o
s
itiv
e
(
FP
)
in
d
icate
s
t
h
e
n
u
m
b
er
o
f
i
n
s
tan
ce
s
in
co
r
r
ec
tl
y
clas
s
i
f
ied
.
T
r
u
e
Neg
ativ
e
(
T
N)
in
d
icate
s
th
e
n
u
m
b
er
o
f
i
n
s
tan
ce
s
co
r
r
ec
tl
y
r
ej
ec
ted
.
Fals
e
Ne
g
ati
v
e
(
FN)
in
d
icate
s
th
e
n
u
m
b
er
o
f
i
n
s
tan
ce
s
in
co
r
r
ec
tl
y
r
ej
ec
ted
.
T
o
ass
ess
th
e
p
r
o
ce
s
s
i
n
g
ti
m
e
,
w
e
u
s
e
t
h
e
tr
ai
n
i
n
g
a
n
d
v
al
i
d
atio
n
ti
m
e
ac
co
r
d
in
g
to
t
h
e
n
u
m
b
er
o
f
n
o
d
es in
t
h
e
B
ig
Data
cl
u
s
ter
.
4
.
5
.
Wo
rk
e
nv
iro
n
m
e
nt
T
ab
le
4
s
u
m
m
ar
izes
o
u
r
w
o
r
k
en
v
ir
o
n
m
e
n
t,
it
p
r
esen
ts
t
h
e
h
ar
d
w
ar
e
a
n
d
th
e
s
o
f
t
w
ar
e
w
it
h
t
h
e
co
n
f
i
g
u
r
atio
n
o
r
v
er
s
io
n
.
T
ab
le
4
.
Har
d
w
ar
e
an
d
s
o
f
t
w
a
r
e
co
n
f
ig
u
r
atio
n
H
a
r
d
w
a
r
e
o
r
so
f
t
w
a
r
e
C
o
n
f
i
g
u
r
a
t
i
o
n
o
r
v
e
r
si
o
n
O
p
e
r
a
t
i
n
g
sy
st
e
m
W
i
n
d
o
w
s 1
0
P
r
o
c
e
sso
r
t
y
p
e
i
3
2
3
7
0
M
C
P
U
@
2
.
4
0
G
H
Z
C
o
r
e
c
o
u
n
t
3
c
o
r
e
s
R
A
M
4
0
9
6
M
B
D
i
sk
2
0
0
G
B
C
l
u
st
e
r
H
a
d
o
o
p
5
n
o
d
e
s
H
a
d
o
o
p
V
e
r
si
o
n
2
.
7
.
6
5.
RE
SU
L
T
S AN
D
AN
AL
Y
SI
S
T
h
is
p
ar
t
is
r
eser
v
ed
to
p
r
esen
t
an
d
d
is
c
u
s
s
th
e
d
i
f
f
er
en
t
r
esu
lt
s
o
b
tain
ed
.
W
e
h
av
e
as
s
ess
ed
t
h
e
ap
p
r
o
ac
h
ac
co
r
d
in
g
to
th
e
th
r
ee
d
is
tr
ib
u
tio
n
s
o
f
th
e
d
ataset
NSL
K
DD
as
d
escr
ib
ed
ab
o
v
e,
s
o
w
e
ca
r
r
ied
o
u
t
th
r
ee
t
y
p
es
o
f
cla
s
s
i
f
icatio
n
,
n
a
m
el
y
,
clas
s
i
f
icatio
n
o
f
t
w
o
clas
s
es,
clas
s
if
icatio
n
o
f
f
iv
e
cla
s
s
es
a
n
d
class
i
f
icatio
n
o
f
t
w
e
n
t
y
-
th
r
ee
class
e
s
.
Fo
r
2
-
clas
s
es
c
lass
if
icatio
n
,
t
h
e
d
atase
t
i
s
d
i
v
id
ed
in
to
t
w
o
s
u
b
s
ets
,
a
s
u
b
s
e
t o
f
n
o
r
m
a
l d
ata
an
d
a
s
u
b
s
et
o
f
at
tack
d
ata.
Fo
r
5
-
cl
ass
es
cl
as
s
if
icatio
n
,
t
h
e
d
ata
s
e
t is d
i
v
id
ed
in
to
f
i
v
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
5
53
–
56
0
558
s
u
b
s
et
s
,
o
n
e
s
u
b
s
et
o
f
n
o
r
m
a
l
d
ata
an
d
f
o
u
r
s
u
b
s
et
s
o
f
at
tack
d
ata,
n
a
m
el
y
,
U2
R
,
R
2
L
,
P
r
o
b
e
an
d
Do
S.
Fo
r
2
3
-
class
es
class
if
icatio
n
,
t
h
e
d
ataset
is
d
iv
id
ed
in
to
t
w
e
n
t
y
-
t
h
r
ee
s
u
b
s
ets,
o
n
e
s
u
b
s
et
o
f
n
o
r
m
al
d
ata
an
d
t
w
e
n
t
y
-
t
w
o
s
u
b
s
ets o
f
at
tack
d
ata
d
er
iv
ed
also
f
r
o
m
U2
R
,
R
2
L
,
P
r
o
b
e
an
d
Do
S a
ttack
s
.
Ou
r
ex
p
er
i
m
e
n
ts
w
er
e
ca
r
r
ied
o
u
t
in
s
e
v
er
al
s
tep
s
,
at
ea
c
h
s
tep
w
e
i
n
cr
ea
s
e
t
h
e
n
u
m
b
er
o
f
n
o
d
es
co
n
s
tit
u
ti
n
g
t
h
e
b
ig
d
ata
clu
s
t
er
an
d
w
e
s
to
r
e
o
u
r
d
ata
th
er
e,
w
e
ap
p
l
y
t
h
e
m
ac
h
i
n
e
lear
n
in
g
c
lass
if
ier
s
i
n
a
d
is
tr
ib
u
ted
a
n
d
p
ar
allel
w
a
y
o
n
t
h
e
b
ig
d
ata
cl
u
s
ter
,
th
e
n
w
e
ca
lcu
late
t
h
e
i
n
d
icato
r
s
ac
cu
r
ac
y
,
s
p
ec
if
ic
it
y
,
s
en
s
iti
v
it
y
,
f
al
s
e
p
o
s
iti
v
e
r
ate
(
FP
R
)
,
ar
ea
u
n
d
er
cu
r
v
e
(
A
U
C
)
,
an
d
th
e
p
r
o
ce
s
s
i
n
g
ti
m
e,
i
n
o
r
d
er
to
ev
alu
ate
th
e
p
er
f
o
r
m
a
n
ce
s
.
Fig
u
r
e
2
s
h
o
w
s
cla
s
s
i
f
icat
io
n
A
cc
u
r
ac
y
f
o
r
t
w
o
clas
s
es,
f
i
v
e
cla
s
s
es
an
d
t
w
e
n
t
y
-
th
r
ee
class
e
s
.
T
ab
le
5
s
h
o
w
s
cla
s
s
i
f
icat
io
n
m
et
r
ics:
s
e
n
s
iti
v
it
y
,
s
p
ec
i
f
icit
y
,
A
U
C
a
n
d
FP
R
f
o
r
t
w
o
cla
s
s
es,
f
i
v
e
cla
s
s
e
s
a
n
d
t
w
e
n
t
y
-
th
r
ee
clas
s
es.
W
h
ile
Fi
g
u
r
e
s
3
d
escr
ib
es
th
e
ev
o
lu
tio
n
o
f
th
e
tr
ain
i
n
g
an
d
v
alid
atio
n
ti
m
e
o
f
m
ac
h
i
n
e
lear
n
in
g
clas
s
i
f
ier
s
ac
co
r
d
in
g
t
o
th
e
n
u
m
b
er
o
f
n
o
d
es o
f
th
e
b
ig
d
ata
clu
s
ter
.
Fig
u
r
e
2
.
C
lass
if
ica
tio
n
A
cc
u
r
ac
y
o
f
t
w
o
clas
s
es,
f
iv
e
cla
s
s
e
s
an
d
t
w
en
t
y
-
t
h
r
ee
class
e
s
T
ab
le
5
.
Sen
s
itiv
it
y
,
s
p
ec
if
ici
t
y
,
A
UC
a
n
d
FP
R
o
f
t
h
e
class
i
f
icatio
n
o
f
t
w
o
clas
s
es,
f
iv
e
cla
s
s
es a
n
d
t
w
en
t
y
-
th
r
ee
class
e
s
C
l
a
ssi
f
i
e
r
D
i
st
r
i
b
u
t
i
o
n
S
e
n
si
t
i
v
i
t
y
(
%)
S
p
e
c
i
f
i
c
i
t
y
(
%)
A
U
C
(
%)
F
P
R
(
%)
K
N
N
2
c
l
a
sse
s
1
0
0
1
0
0
1
0
0
0
5
c
l
a
sse
s
9
2
,
4
1
0
0
9
6
,
2
0
2
3
c
l
a
sse
s
7
7
,
7
1
0
0
8
8
,
7
8
0
S
V
M
2
c
l
a
sse
s
9
8
,
5
9
8
,
5
1
0
0
1
,
5
5
c
l
a
sse
s
8
8
,
6
9
9
,
8
9
9
,
8
0
,
2
2
3
c
l
a
sse
s
6
6
,
3
1
0
0
9
6
,
8
7
0
T
R
EE
2
c
l
a
sse
s
0
0
0
1
0
0
5
c
l
a
sse
s
85
1
0
0
9
9
,
2
0
2
3
c
l
a
sse
s
4
7
,
2
2
9
9
,
9
6
9
7
,
3
9
0
,
0
4
As
illu
s
tr
ated
b
y
F
ig
u
r
e
2
,
th
e
d
if
f
er
en
t
v
alu
e
s
r
ea
ch
ed
o
f
ac
cu
r
ac
y
f
o
r
all
m
ac
h
in
e
lear
n
in
g
class
i
f
ier
s
ar
e
g
e
n
er
all
y
v
er
y
h
ig
h
,
t
h
e
KNN
al
g
o
r
ith
m
is
v
er
y
e
f
f
icie
n
t
w
it
h
ac
cu
r
ac
y
v
a
lu
es
t
h
at
r
ea
ch
e
s
u
p
to
9
9
.
9
%
f
o
r
th
e
clas
s
if
icatio
n
o
f
t
w
o
clas
s
es,
9
9
.
9
%
f
o
r
th
e
class
if
ica
tio
n
o
f
f
i
v
e
cla
s
s
e
s
,
an
d
9
9
.
8
%
f
o
r
t
h
e
class
i
f
icatio
n
o
f
th
e
t
w
e
n
t
y
t
h
r
ee
class
es,
w
h
ic
h
m
ea
n
s
t
h
at
KNN
is
v
er
y
p
o
w
er
f
u
l
t
h
an
SVM
an
d
d
ec
is
io
n
tr
ee
f
o
r
id
en
ti
f
y
i
n
g
ea
ch
t
y
p
e
o
f
d
ata
w
h
ate
v
er
th
e
d
ata
d
is
tr
ib
u
tio
n
.
Fo
r
th
e
d
i
s
tr
ib
u
tio
n
in
t
w
o
clas
s
es
a
n
d
f
i
v
e
class
e
s
,
th
e
d
ata
i
s
n
o
t
d
i
s
tr
ib
u
ted
in
s
ev
er
al
cla
s
s
e
s
;
t
h
e
d
ec
is
io
n
tr
ee
is
m
o
r
e
ac
c
u
r
ate
th
a
n
SVM
w
i
th
ac
cu
r
ac
y
p
er
ce
n
tag
e
s
o
f
9
9
.
8
%
an
d
9
9
.
6
%
f
o
r
t
h
e
d
et
ec
tio
n
o
f
r
esp
ec
ti
v
el
y
t
w
o
an
d
f
iv
e
clas
s
es,
w
h
ic
h
m
ea
n
s
th
a
t
d
ec
is
io
n
tr
e
e
is
v
er
y
ef
f
icie
n
t
w
it
h
less
d
is
tr
ib
u
ted
d
ata.
Fo
r
th
e
d
is
tr
ib
u
tio
n
i
n
t
w
e
n
t
y
-
th
r
ee
cla
s
s
es,
t
h
e
d
at
aset
is
m
o
r
e
d
is
tr
ib
u
ted
;
th
e
r
ea
ch
ed
ac
cu
r
ac
y
b
y
SV
M
m
et
h
o
d
is
9
9
.
4
%,
w
h
ic
h
ex
p
lai
n
s
t
h
at
SV
M
is
m
o
r
e
p
r
ec
is
e
f
o
r
d
ata
w
it
h
a
h
i
g
h
d
is
tr
ib
u
tio
n
.
A
l
s
o
,
as
s
h
o
w
n
b
y
T
ab
le
5
,
th
e
h
i
g
h
est
Sen
s
iti
v
it
y
v
al
u
es
ar
e
t
h
o
s
e
o
f
t
h
e
KNN
alg
o
r
ith
m
,
1
0
0
%
f
o
r
th
e
d
etec
tio
n
o
f
t
w
o
class
es,
9
2
.
4
%
f
o
r
t
h
e
d
etec
ti
o
n
o
f
f
iv
e
c
lass
e
s
a
n
d
7
7
.
7
%
f
o
r
th
e
d
etec
tio
n
o
f
t
w
e
n
t
y
t
h
r
ee
cla
s
s
e
s
,
w
h
ich
m
ea
n
s
th
at
KNN
ca
n
co
r
r
ec
tl
y
i
d
en
tify
th
e
n
at
u
r
e
o
f
t
h
e
d
ata
m
o
r
e
th
a
n
th
e
b
o
th
m
et
h
o
d
s
SVM
an
d
d
ec
is
io
n
t
r
ee
.
T
h
e
v
alu
e
s
o
f
Sen
s
iti
v
it
y
w
h
ich
ar
e
in
s
ec
o
n
d
p
o
s
itio
n
ar
e
th
o
s
e
o
f
SVM
w
it
h
p
er
ce
n
tag
e
s
o
f
9
8
.
5
%
f
o
r
th
e
id
e
n
ti
f
icatio
n
o
f
t
w
o
cla
s
s
es,
8
8
.
6
%
f
o
r
th
e
id
en
ti
f
icat
io
n
o
f
f
i
v
e
cla
s
s
e
s
,
97
97,
5
98
98,
5
99
99,
5
100
T
r
e
e
S
V
M
K
NN
99,
8
98,
2
99,
9
99,
6
98,
8
99,
9
98,
4
99,
4
99,
8
A
c
c
ur
a
c
y
(%
)
2 c
l
as
se
s
5 c
l
as
se
s
23 c
l
as
se
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
P
a
r
a
llel
p
r
o
ce
s
s
in
g
u
s
in
g
b
ig
d
a
ta
a
n
d
ma
c
h
in
e
lea
r
n
in
g
tec
h
n
iq
u
es fo
r
... (
A
la
ed
d
i
n
e
B
o
u
k
h
a
lfa
)
559
an
d
6
6
.
3
%
f
o
r
t
h
e
id
e
n
ti
f
icati
o
n
o
f
t
w
e
n
t
y
t
h
r
ee
clas
s
es,
w
h
ich
p
r
o
v
es
t
h
at
S
VM
ca
n
p
er
f
ec
tl
y
d
etec
t
th
e
t
y
p
e
o
f
d
ata,
m
o
r
e
th
an
t
h
e
d
ec
is
io
n
tr
ee
m
e
th
o
d
.
E
f
f
ec
t
iv
el
y
,
th
e
p
er
ce
n
tag
e
o
f
th
e
f
al
s
e
p
o
s
itiv
e
r
ate
(
FP
R
)
n
o
ted
b
y
KNN
is
n
u
ll
f
o
r
an
y
t
y
p
e
o
f
clas
s
i
f
icatio
n
,
t
h
is
ex
p
lai
n
s
th
at
KNN
ca
n
d
etec
t
w
i
th
o
u
t
er
r
o
r
co
m
p
ar
ed
to
th
e
o
th
er
cla
s
s
i
f
ier
s
.
An
d
al
s
o
,
th
e
v
alu
e
s
ac
h
iev
ed
o
f
t
h
e
f
al
s
e
p
o
s
itiv
e
r
ate
(
FP
R
)
b
y
SV
M
ar
e
o
n
l
y
1
.
5
%
f
o
r
th
e
d
etec
tio
n
o
f
t
w
o
clas
s
es,
0
.
2
%
f
o
r
th
e
d
etec
tio
n
o
f
f
i
v
e
class
es
an
d
n
u
ll
f
o
r
th
e
d
etec
t
io
n
o
f
t
w
en
t
y
-
t
h
r
ee
class
es,
w
h
ic
h
m
ea
n
s
t
h
at
S
VM
d
etec
ts
w
it
h
le
s
s
er
r
o
r
th
an
d
ec
i
s
io
n
tr
ee
.
W
e
also
n
o
tice
th
at
s
p
ec
i
f
icit
y
v
alu
e
s
o
f
KN
N
r
ea
ch
u
p
to
1
0
0
%
f
o
r
r
ec
o
g
n
it
io
n
o
f
t
w
o
class
e
s
,
f
i
v
e
cla
s
s
es
an
d
t
wen
t
y
-
t
h
r
ee
cla
s
s
e
s
,
w
h
ic
h
m
ea
n
s
t
h
at
KN
N
ca
n
p
e
r
f
ec
tl
y
d
etec
t
n
eg
a
tiv
e
i
n
s
tan
c
es
m
o
r
e
th
a
n
th
e
o
t
h
er
m
et
h
o
d
s
.
As
r
ep
r
esen
ted
b
y
t
h
e
Fig
u
r
e
3
(
a
-
c
)
,
th
e
ti
m
e
p
r
o
ce
s
s
in
g
o
f
th
e
alg
o
r
it
h
m
s
d
ec
r
ea
s
es
as
th
e
n
u
m
b
e
r
o
f
n
o
d
es
in
th
e
clu
s
ter
in
cr
ea
s
es.
I
n
th
e
ca
s
e
o
f
a
clu
s
ter
w
it
h
a
s
in
g
le
n
o
d
e,
th
e
tr
ain
in
g
a
n
d
v
alid
atio
n
ti
m
e
tak
en
b
y
K
NN
is
1
8
2
6
s
f
o
r
t
h
e
class
i
f
ica
tio
n
o
f
t
w
o
clas
s
e
s
,
1
8
5
6
.
5
s
f
o
r
th
e
class
if
icat
i
o
n
o
f
f
i
v
e
class
e
s
,
an
d
1
7
9
2
.
8
s
f
o
r
th
e
clas
s
i
f
ic
atio
n
o
f
t
w
en
t
y
t
h
r
ee
cla
s
s
es,
th
ese
v
al
u
es
d
ec
r
ea
s
e
as
lo
n
g
as
t
h
e
n
u
m
b
er
o
f
n
o
d
es
o
f
t
h
e
cl
u
s
ter
i
n
cr
ea
s
e
s
,
u
n
til
r
ea
c
h
i
n
g
o
n
l
y
in
th
e
ca
s
e
o
f
a
cl
u
s
ter
w
it
h
a
f
i
v
e
n
o
d
e,
1
6
6
7
.
6
s
f
o
r
th
e
class
if
ica
tio
n
o
f
t
wo
class
es,
1
6
1
1
.
9
s
f
o
r
th
e
class
i
f
icatio
n
o
f
f
i
v
e
class
es,
an
d
1
6
5
9
.
4
s
f
o
r
th
e
class
i
f
icat
io
n
o
f
t
wen
t
y
t
h
r
ee
class
e
s
.
W
e
clea
r
l
y
n
o
tice
th
e
s
a
m
e
ev
o
l
u
tio
n
f
o
r
th
e
o
th
er
t
w
o
m
eth
o
d
s
SV
M
an
d
d
ec
is
io
n
tr
ee
.
W
h
ich
j
u
s
tifie
s
t
h
at
p
ar
allel
an
d
d
is
tr
ib
u
ted
p
r
o
ce
s
s
in
g
r
ed
u
ce
s
ef
f
ec
tiv
e
l
y
ti
m
e
co
n
s
u
m
p
tio
n
.
N
u
mb
e
r
o
f
n
o
d
e
s
2
c
l
a
sse
s
5
c
l
a
sse
s
2
3
c
l
a
sse
s
1
1
8
2
6
1
8
5
6
,
5
1
7
9
2
,
8
2
1
8
2
3
,
1
1
7
9
7
,
6
1
7
2
1
,
9
3
1
8
2
0
,
6
1
7
8
4
,
1
1
7
0
4
,
1
4
1
6
7
2
,
6
1
6
6
0
,
1
1
6
7
0
,
9
5
1
6
6
7
,
6
1
6
1
1
,
9
1
6
5
9
,
4
(
a)
N
u
mb
e
r
o
f
n
o
d
e
s
2
c
l
a
sse
s
5
c
l
a
sse
s
2
3
c
l
a
sse
s
1
7
4
2
,
0
6
6
3
7
,
1
3
4
9
9
,
7
7
2
7
1
6
,
6
9
5
5
5
,
5
7
4
9
0
,
6
2
3
6
9
1
,
5
6
5
3
8
,
5
9
4
6
7
,
9
8
4
5
3
6
,
5
7
4
8
8
,
4
9
4
5
1
,
2
4
5
5
1
6
,
5
2
4
4
7
,
6
3
4
4
6
,
3
9
(
b
)
N
u
mb
e
r
o
f
n
o
d
e
s
2
c
l
a
sse
s
5
c
l
a
sse
s
2
3
c
l
a
sse
s
1
5
7
,
5
3
2
6
0
,
0
4
8
1
0
4
,
8
6
2
5
5
,
9
3
5
4
5
,
5
7
5
9
,
1
7
7
3
4
7
,
5
4
5
4
2
,
6
8
7
5
6
,
0
9
6
4
4
6
,
0
9
2
4
2
,
1
1
7
5
1
,
9
0
6
5
3
8
,
0
4
2
4
1
,
0
5
1
3
4
,
2
0
9
(
c)
Fig
u
r
e
3
.
T
h
e
ev
o
lu
tio
n
o
f
tr
ai
n
in
g
a
n
d
v
alid
atio
n
t
i
m
e
ac
co
r
d
in
g
to
th
e
n
u
m
b
er
o
f
n
o
d
es in
th
e
clu
s
ter
(
a)
KNN
,
(
b
)
SVM
,
(
c
)
T
r
ee
1
5
5
0
1
6
0
0
1
6
5
0
1
7
0
0
1
7
5
0
1
8
0
0
1
8
5
0
1
9
0
0
0
1
2
3
4
5
6
T
r
a
i
ni
ng
a
nd
V
a
l
i
da
ti
o
n
ti
me
(s
)
Nu
mb
e
r
o
f
n
o
d
e
s
K
NN
2
c
l
a
s
s
e
s
5
c
l
a
s
s
e
s
2
3
c
l
a
s
s
es
4
0
0
5
0
0
6
0
0
7
0
0
8
0
0
0
1
2
3
4
5
6
T
r
a
i
ni
ng
a
nd
V
a
l
i
da
ti
o
n
ti
me
(s
)
N
um
be
r
o
f
no
de
s
SV
M
2
c
l
a
s
s
e
s
5
c
l
a
s
s
e
s
2
3
c
l
a
s
s
es
20
40
60
80
1
0
0
1
2
0
0
1
2
3
4
5
6
T
r
a
i
ni
ng
a
nd
V
a
l
i
da
ti
o
n
ti
me
(s
)
N
u
m
b
e
r
o
f
n
o
d
e
s
T
ree
2
c
l
a
s
s
e
s
5
c
l
a
s
s
e
s
2
3
c
l
a
s
s
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
5
53
–
56
0
560
T
h
e
ex
p
er
i
m
en
ts
h
a
v
e
p
r
o
v
en
th
at
m
ac
h
i
n
e
lear
n
in
g
al
g
o
r
it
h
m
s
ar
e
v
er
y
e
f
f
ec
ti
v
e
at
d
ete
ctin
g
n
e
w
h
id
d
en
attac
k
s
an
d
in
tr
u
s
io
n
s
,
an
d
ap
p
ly
i
n
g
t
h
e
m
i
n
a
p
ar
allel
w
a
y
in
a
d
is
tr
ib
u
ted
en
v
i
r
o
n
m
e
n
t
i
m
p
r
o
v
e
s
s
ig
n
i
f
ica
n
tl
y
ti
m
e
co
n
s
u
m
p
t
io
n
.
6.
CO
NCLU
SI
O
N
AND
F
U
T
U
RE
WO
RK
I
n
th
i
s
s
t
u
d
y
,
w
e
s
u
g
g
ested
a
n
e
w
ap
p
r
o
ac
h
estab
lis
h
ed
o
n
th
e
s
to
r
ag
e
o
f
t
h
e
lar
g
e
v
o
lu
m
e
a
n
d
v
ar
iet
y
o
f
n
et
w
o
r
k
tr
a
f
f
ic
d
a
ta
u
s
in
g
b
ig
d
ata
tech
n
iq
u
e
s
,
an
d
th
e
a
n
al
y
s
is
o
f
th
e
s
e
d
ata
u
s
i
n
g
m
ac
h
in
e
lear
n
in
g
al
g
o
r
ith
m
s
in
a
d
is
tr
ib
u
ted
an
d
p
ar
allel
w
a
y
,
i
n
o
r
d
er
to
d
etec
t
n
e
w
h
id
d
en
i
n
tr
u
s
io
n
s
w
it
h
le
s
s
ti
m
e
co
n
s
u
m
p
tio
n
.
T
o
p
r
o
v
e
th
e
v
al
id
it
y
o
f
o
u
r
ap
p
r
o
ac
h
,
a
b
ig
d
ata
clu
s
ter
h
a
s
b
ee
n
s
et
u
p
,
t
h
e
p
o
p
u
lar
NSL
K
DD
w
a
s
c
h
o
s
en
as
d
ata
s
et
f
o
r
th
e
e
v
al
u
atio
n
.
T
h
e
a
s
s
e
s
s
m
en
t
w
a
s
ca
r
r
ied
o
u
t
f
o
llo
w
i
n
g
s
e
v
er
al
s
tep
s
,
at
ea
ch
s
tep
,
th
e
n
u
m
b
er
o
f
n
o
d
es
i
n
t
h
e
b
ig
d
ata
cl
u
s
t
er
is
in
cr
ea
s
ed
,
th
e
N
S
L
K
DD
is
s
to
r
ed
in
t
h
e
b
ig
d
ata
clu
s
ter
,
th
e
m
ac
h
i
n
e
l
ea
r
n
in
g
al
g
o
r
ith
m
s
ar
e
ap
p
lied
f
o
r
th
e
an
a
l
y
s
is
,
a
n
d
th
e
n
th
e
ev
alu
a
tio
n
m
etr
ic
s
ar
e
ca
lcu
lated
.
T
o
s
u
p
p
o
r
t
th
e
v
alid
it
y
o
f
o
u
r
p
r
o
p
o
s
al,
th
e
e
x
p
er
i
m
e
n
ts
r
es
u
lts
s
h
o
wn
t
h
at
th
e
m
ac
h
i
n
e
lear
n
in
g
m
et
h
o
d
s
ar
e
v
er
y
e
f
f
e
ctiv
e
to
s
e
n
s
i
n
g
i
n
tr
u
s
io
n
a
n
d
th
eir
ap
p
licatio
n
i
n
a
p
ar
allel
a
n
d
d
is
tr
ib
u
ted
w
a
y
r
ed
u
ce
s
co
n
s
id
er
ab
l
y
ti
m
e
c
o
n
s
u
m
p
tio
n
.
I
n
t
h
e
f
u
t
u
r
e,
we
w
ill
tr
y
to
i
m
p
le
m
en
t
r
ea
l
l
y
a
n
e
w
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
te
m
(
I
DS)
u
s
in
g
o
u
r
n
e
w
d
is
tr
ib
u
ted
ap
p
r
o
ac
h
.
RE
F
E
R
E
NC
E
S
[1
]
J.
Ca
m
a
c
h
o
,
e
t
a
l.
,
“
M
u
lt
iv
a
riate
Big
Da
ta
A
n
a
l
y
sis
f
o
r
in
tru
sio
n
d
e
tec
ti
o
n
:
5
ste
p
s
f
ro
m
th
e
h
a
y
st
a
c
k
to
th
e
n
e
e
d
le
”
,
Co
mp
u
t.
S
e
c
u
r.
,
v
o
l.
8
7
,
p
.
1
0
1
6
0
3
,
n
o
v
.
2
0
1
9
.
[2
]
L
.
W
a
n
g
,
e
t
a
l.
,
“
Big
Da
ta A
n
a
l
y
t
ics
in
Cy
b
e
r
S
e
c
u
rit
y
:
Ne
t
w
o
rk
T
r
a
ff
ic an
d
A
tt
a
c
k
s
”
,
J
.
C.
I.
S
.
,
p
.
1
‑
8
,
2
0
2
0
.
[3
]
“
KD
D Cu
p
1
9
9
9
Da
ta
”
.
h
tt
p
:/
/k
d
d
.
ics
.
u
c
i.
e
d
u
/
d
a
tab
a
se
s/k
d
d
c
u
p
9
9
/
k
d
d
c
u
p
9
9
.
h
tm
l
[4
]
“
NSL
-
KD
D D
a
tas
e
t
”
.
h
tt
p
s://
g
it
h
u
b
.
c
o
m
/d
e
f
c
o
m
1
7
/NS
L
_
KD
D
[5
]
V
.
K.
V
a
sa
n
t
h
a
m
,
e
t
a
l.
,
“
Us
e
r
-
An
o
m
a
l
y
De
te
c
ti
o
n
i
n
T
e
lec
o
m
m
u
n
ica
ti
o
n
Us
in
g
Big
Da
ta
A
n
a
l
y
ti
c
s
”
,
v
o
l.
7
,
n
o
5
,
p
.
4
,
2
0
1
9
.
[6
]
M
.
T
.
T
u
n
,
e
t
a
l.
,
“
P
e
rf
o
rm
a
n
c
e
E
v
a
lu
a
ti
o
n
o
f
In
tru
si
o
n
De
tec
ti
o
n
S
trea
m
in
g
T
ra
n
sa
c
ti
o
n
s Us
in
g
A
p
a
c
h
e
Ka
f
k
a
a
n
d
S
p
a
rk
S
trea
m
in
g
”
,
in
ICAIT
2
0
1
9
,
Ya
n
g
o
n
,
M
y
a
n
m
a
r,
No
v
.
2
0
1
9
,
p
.
2
5
‑
30.
[7
]
“
A
p
a
c
h
e
Ka
f
k
a
”
,
Ap
a
c
h
e
K
a
fka
.
h
tt
p
:
//
k
a
f
k
a
.
a
p
a
c
h
e
.
o
rg
/.
[8
]
“
A
p
a
c
h
e
S
p
a
rk
”
.
h
tt
p
s:
//
sp
a
rk
.
a
p
a
c
h
e
.
o
rg
/.
[9
]
“
S
p
a
rk
S
trea
m
in
g
”
.
h
tt
p
s://
s
p
a
rk
.
a
p
a
c
h
e
.
o
rg
/d
o
c
s/late
st/strea
m
in
g
-
p
ro
g
ra
m
m
in
g
-
g
u
id
e
.
h
tm
l
[1
0
]
N.
M
o
u
sta
f
a
,
e
t
a
l.
,
“
Big
Da
ta
An
a
ly
ti
c
s
f
o
r
In
tru
sio
n
De
tec
ti
o
n
S
y
ste
m
:
S
tatisti
c
a
l
De
c
isio
n
-
M
a
k
in
g
Us
in
g
F
in
it
e
Dirich
let
M
ix
tu
re
M
o
d
e
ls
”
,
in
DA
DS
C
,
S
p
rin
g
e
r
In
ter
n
a
ti
o
n
a
l
P
u
b
li
sh
in
g
,
2
0
1
7
,
p
.
1
2
7
‑
1
5
6
.
[1
1
]
“
UN
S
W
-
NB1
5
”
.
ww
w
.
u
n
sw
.
a
d
fa
.
e
d
u
.
a
u
/
u
n
sw
-
c
a
n
b
e
rra
-
c
y
b
e
r/c
y
b
e
rse
c
u
rit
y
/
AD
F
A
-
NB1
5
-
Da
tas
e
ts/
[1
2
]
L
.
U.
L
a
b
o
sh
in
,
e
t
a
l
.
,
“
T
h
e
B
ig
Da
ta
A
p
p
ro
a
c
h
to
Co
ll
e
c
ti
n
g
a
n
d
A
n
a
l
y
z
in
g
T
ra
ff
i
c
Da
ta
in
L
a
rg
e
S
c
a
le
Ne
tw
o
rk
s
”
,
Pro
c
e
d
ia
Co
m
p
u
t
.
S
c
i
.
,
v
o
l.
1
0
3
,
p
.
5
3
6
‑
5
4
2
,
2
0
1
7.
[1
3
]
S
.
K.
Ba
n
sa
l,
e
t
a
l
.
,
“
In
teg
ra
ti
n
g
Big
Da
ta:
A
S
e
m
a
n
ti
c
Ex
trac
t
-
T
r
a
n
sf
o
r
m
-
L
o
a
d
F
ra
m
e
w
o
r
k
”
,
Co
mp
u
ter
,
v
o
l.
4
8
,
n
o
3
,
p
.
4
2
‑
5
0
,
m
a
rs 2
0
1
5
.
[1
4
]
“
A
p
a
c
h
e
Ha
d
o
o
p
”
.
h
tt
p
s://
h
a
d
o
o
p
.
a
p
a
c
h
e
.
o
rg
/.
[1
5
]
A
.
Bo
u
k
h
a
lfa,
e
t
a
l.
,
“
A
Ho
n
e
y
Ne
t,
Big
Da
ta
a
n
d
RNN
A
rc
h
it
e
c
tu
re
f
o
r
A
u
to
m
a
ti
c
S
e
c
u
rit
y
M
o
n
i
to
ri
n
g
o
f
In
f
o
rm
a
ti
o
n
S
y
ste
m
”
,
in
(
AI2
S
D’2
0
1
8
)
,
v
o
l.
9
1
5
,
S
p
ri
n
g
e
r
In
tern
a
t
io
n
a
l
P
u
b
li
sh
i
n
g
,
2
0
1
9
,
p
.
8
0
0
‑
8
0
8
.
[1
6
]
I.
M
a
v
rid
is,
e
t
a
l.
,
“
P
e
rf
o
rm
a
n
c
e
e
v
a
lu
a
ti
o
n
o
f
c
lo
u
d
-
b
a
se
d
lo
g
f
il
e
a
n
a
ly
sis
w
it
h
A
p
a
c
h
e
Ha
d
o
o
p
a
n
d
A
p
a
c
h
e
S
p
a
rk
”
,
J
.
S
y
s
t.
S
o
ft
w
.
,
v
o
l
.
1
2
5
,
p
.
1
3
3
‑
1
5
1
,
m
a
rs 2
0
1
7
.
[1
7
]
A
.
Bo
u
k
h
a
lf
a
,
e
t
a
l
,
“
Ne
t
w
o
rk
T
r
a
ff
ic
A
n
a
l
y
sis
u
sin
g
Big
Da
ta
a
n
d
De
e
p
L
e
a
rn
in
g
T
e
c
h
n
iq
u
e
s
”
,
i
n
(
ICOA
2
0
2
0
)
,
2
0
2
0
,
p
.
1
‑
4.
[1
8
]
L
.
Zh
o
u
,
e
t
a
l.
,
“
M
a
c
h
in
e
lea
rn
in
g
o
n
b
ig
d
a
ta:
Op
p
o
rt
u
n
i
ti
e
s
a
n
d
c
h
a
ll
e
n
g
e
s
”
,
Ne
u
ro
c
o
m
p
u
t
in
g
,
v
o
l.
2
3
7
,
p
.
350
‑
3
6
1
,
m
a
i
2
0
1
7
.
[1
9
]
B.
M
.
A
sla
h
i
-
S
h
a
h
ri
e
t
a
l.
,
“
A
h
y
b
rid
m
e
th
o
d
c
o
n
sisti
n
g
o
f
GA
a
n
d
S
V
M
f
o
r
in
tr
u
sio
n
d
e
tec
ti
o
n
sy
ste
m
”
,
N
e
u
ra
l
Co
mp
u
t
.
A
p
p
l
.
,
v
o
l.
2
7
,
n
o
6
,
p
.
1
6
6
9
‑
1
6
7
6
,
a
o
û
t
2
0
1
6
.
[2
0
]
N.
S
a
m
e
e
ra
,
e
t
a
l.
,
“
En
c
o
d
i
n
g
Ap
p
r
o
a
c
h
f
o
r
In
tr
u
sio
n
De
tec
ti
o
n
Us
in
g
P
CA
a
n
d
KN
N
Clas
sif
ier
”
,
in
ICCII
2
0
2
0
,
S
in
g
a
p
o
re
,
2
0
2
0
,
p
.
1
8
7
‑
1
9
9
.
[2
1
]
J.
Es
m
a
il
y
,
e
t
a
l.
,
“
In
tru
sio
n
d
e
t
e
c
ti
o
n
sy
ste
m
b
a
se
d
o
n
M
u
lt
i
-
L
a
y
e
r
P
e
rc
e
p
tro
n
Ne
u
ra
l
Ne
t
w
o
rk
s
a
n
d
De
c
isio
n
T
re
e
”
,
in
2
0
1
5
7
th
C
o
n
fer
e
n
c
e
o
n
(
IKT
)
,
Ira
n
,
m
a
i
2
0
1
5
,
p
.
1
‑
5
.
[2
2
]
“
S
V
M
”
.
h
tt
p
s:
//
ww
w
.
sc
ien
c
e
d
irec
t.
c
o
m
/t
o
p
ics
/co
m
p
u
ter
-
sc
ien
c
e
/su
p
p
o
rt
-
v
e
c
to
r
-
m
a
c
h
in
e
[2
3
]
W
.
L
i,
e
t
a
l.
,
“
A
Ne
w
In
tru
sio
n
De
tec
ti
o
n
S
y
ste
m
Ba
s
e
d
o
n
K
NN
Clas
sif
i
c
a
ti
o
n
A
lg
o
rit
h
m
in
W
irele
ss
S
e
n
so
r
Ne
tw
o
rk
”
,
J
.
El
e
c
tr.
Co
mp
u
t.
E
n
g
.
,
v
o
l.
2
0
1
4
,
p
.
1
‑
8
,
2
0
1
4
.
[2
4
]
S
.
S
a
h
u
e
t
B.
M
.
M
e
h
tre,
“
Ne
tw
o
rk
in
tru
si
o
n
d
e
tec
ti
o
n
sy
ste
m
u
sin
g
J4
8
De
c
isio
n
T
re
e
”
,
in
2
0
1
5
(
ICACCI)
,
Ko
c
h
i
,
In
d
ia,
a
o
û
t
2
0
1
5
,
p
.
2
0
2
3
‑
2
0
2
6
.
[2
5
]
A
.
Bo
u
k
h
a
lf
a
,
e
t
a
l.
,
“
L
S
T
M
d
e
e
p
lea
rn
in
g
m
e
th
o
d
f
o
r
n
e
tw
o
rk
in
tru
sio
n
d
e
tec
ti
o
n
sy
ste
m
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
0
,
n
o
3
,
p
.
3
3
1
5
,
2
0
2
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.