I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
rm
a
t
ics a
nd
Co
m
m
un
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
20
2
6
,
p
p
.
665
~
6
7
3
I
SS
N:
2252
-
8
7
7
6
,
DOI
:
1
0
.
1
1
5
9
1
/iji
ct
.
v
1
5
i
2
.
pp
665
-
6
7
3
665
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ict.
ia
esco
r
e.
co
m
Efficien
t
ema
il
cl
a
ss
ificatio
n t
ech
ni
que:
a
com
pa
ra
ti
v
e study o
f
hea
der
-
o
nly
and
f
ull
-
co
nten
t
a
ppro
a
ches
Wo
ra
wit
K
it
ik
us
o
un
,
Na
wa
po
rn
Wis
it
po
ng
p
ha
n
D
e
p
a
r
t
me
n
t
o
f
D
i
g
i
t
a
l
N
e
t
w
o
r
k
a
n
d
I
n
f
o
r
ma
t
i
o
n
S
e
c
u
r
i
t
y
M
a
n
a
g
e
me
n
t
,
F
a
c
u
l
t
y
o
f
I
n
f
o
r
ma
t
i
o
n
Te
c
h
n
o
l
o
g
y
a
n
d
D
i
g
i
t
a
l
I
n
n
o
v
a
t
i
o
n
,
K
i
n
g
M
o
n
g
k
u
t
's U
n
i
v
e
r
si
t
y
o
f
Te
c
h
n
o
l
o
g
y
N
o
r
t
h
B
a
n
g
k
o
k
,
B
a
n
g
k
o
k
,
Th
a
i
l
a
n
d
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
r
1
8
,
2
0
2
5
R
ev
is
ed
No
v
1
5
,
2
0
2
5
Acc
ep
ted
Dec
1
4
,
2
0
2
5
Th
e
p
u
r
p
o
se
o
f
t
h
is
re
se
a
rc
h
is
t
o
e
x
p
l
o
re
e
fficie
n
t
tec
h
n
iq
u
e
s
a
n
d
su
fficie
n
t
fe
a
tu
re
s
fo
r
o
rg
a
n
iza
ti
o
n
a
l
e
m
a
il
c
las
sifica
ti
o
n
,
wi
th
a
fo
c
u
s
o
n
i
d
e
n
ti
f
y
in
g
e
m
a
il
s
th
a
t
a
re
n
o
t
b
e
n
e
ficia
l
f
o
r
wo
rk
to
re
d
u
c
e
th
e
b
u
rd
e
n
o
f
e
m
a
il
m
a
n
a
g
e
m
e
n
t.
Th
is
st
u
d
y
p
ro
p
o
se
s
a
n
o
v
e
l
a
p
p
r
o
a
c
h
b
y
c
o
m
p
a
rin
g
t
h
e
p
e
rfo
rm
a
n
c
e
o
f
u
sin
g
e
m
a
il
h
e
a
d
e
r
fe
a
tu
re
s
(He
a
d
e
r
-
On
ly
)
v
e
rsu
s
fu
ll
e
m
a
il
d
a
ta
(He
a
d
e
r
+
Bo
d
y
),
a
imin
g
to
e
v
a
lu
a
te
th
e
a
c
c
u
ra
c
y
a
n
d
p
r
o
c
e
ss
in
g
ti
m
e
o
f
wid
e
ly
u
se
d
m
a
c
h
in
e
lea
rn
in
g
a
lg
o
rit
h
m
s,
in
c
lu
d
in
g
Ra
n
d
o
m
F
o
re
st,
S
VM,
KN
N,
XG
Bo
o
st,
a
n
d
AN
N.
Th
e
e
x
p
e
rime
n
t
wa
s
c
o
n
d
u
c
te
d
u
si
n
g
t
h
e
En
ro
n
d
a
tas
e
t,
with
k
e
y
fe
a
tu
re
s
e
x
trac
ted
fro
m
e
m
a
il
h
e
a
d
e
rs
su
c
h
a
s
se
n
d
e
r
a
n
d
re
c
ip
ien
t
a
d
d
re
ss
e
s
a
n
d
fr
o
m
th
e
b
o
d
y
c
o
n
ten
t.
T
h
e
re
su
lt
s
sh
o
w
t
h
a
t
u
sin
g
o
n
l
y
h
e
a
d
e
r
in
f
o
rm
a
ti
o
n
p
r
o
v
i
d
e
s
c
las
sifica
ti
o
n
p
e
rfo
r
m
a
n
c
e
c
o
m
p
a
ra
b
le
to
u
si
n
g
f
u
ll
e
m
a
il
c
o
n
ten
t.
I
n
p
a
rti
c
u
lar,
m
o
d
e
ls
su
c
h
a
s
Ra
n
d
o
m
F
o
re
st,
XG
Bo
o
st,
a
n
d
Li
g
h
tG
BM
a
c
h
iev
e
d
a
c
c
u
ra
c
y
e
x
c
e
e
d
in
g
9
5
%
,
w
h
il
e
re
d
u
c
in
g
p
ro
c
e
ss
in
g
t
ime
b
y
u
p
t
o
2
1
.
6
6
%
in
th
e
Ra
n
d
o
m
F
o
re
s
t
m
o
d
e
l.
It
is
e
v
id
e
n
t
th
a
t
c
las
sify
i
n
g
e
m
a
i
ls
u
sin
g
h
e
a
d
e
r
-
o
n
ly
fe
a
tu
r
e
s
is
b
o
th
h
ig
h
ly
a
c
c
u
ra
te
a
n
d
re
so
u
rc
e
-
e
fficie
n
t.
T
h
is
re
se
a
rc
h
o
ffe
rs
p
ra
c
ti
c
a
l
g
u
i
d
a
n
c
e
fo
r
o
rg
a
n
iza
ti
o
n
s
i
n
d
e
v
e
lo
p
in
g
e
ffe
c
ti
v
e
e
m
a
il
fil
terin
g
sy
ste
m
s
with
o
u
t
c
o
m
p
r
o
m
isin
g
c
las
sifica
ti
o
n
q
u
a
li
t
y
.
K
ey
w
o
r
d
s
:
E
m
ail
class
if
ica
tio
n
E
m
ail
h
ea
d
er
E
m
ail
s
ec
u
r
ity
Ma
ch
in
e
lear
n
in
g
Sp
am
em
ail
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
:
Naw
ap
o
r
n
W
is
itp
o
n
g
p
h
a
n
Dep
ar
tm
en
t o
f
Dig
ital N
etwo
r
k
an
d
I
n
f
o
r
m
atio
n
Secu
r
ity
M
an
ag
em
en
t
Facu
lty
o
f
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
a
n
d
Dig
ital I
n
n
o
v
atio
n
Kin
g
Mo
n
g
k
u
t'
s
Un
iv
er
s
ity
o
f
T
ec
h
n
o
lo
g
y
No
r
th
B
an
g
k
o
k
B
an
g
k
o
k
,
T
h
ailan
d
E
m
ail:
n
awa
p
o
r
n
.
w@
itd
.
k
m
u
tn
b
.
ac
.
th
1.
I
NT
RO
D
UCT
I
O
N
E
m
ail
class
if
icatio
n
o
r
ca
teg
o
r
izatio
n
h
as
b
ec
o
m
e
i
n
cr
ea
s
in
g
ly
im
p
o
r
tan
t
as
th
e
n
u
m
b
er
o
f
em
ails
in
d
iv
id
u
als
r
ec
eiv
e
d
aily
co
n
tin
u
es
to
r
is
e.
S
p
am
em
ails
ar
e
also
ca
teg
o
r
ized
i
n
to
v
a
r
io
u
s
ty
p
es,
s
u
ch
as
p
r
o
m
o
tio
n
s
,
s
o
cial,
u
p
d
ates,
a
n
d
f
o
r
u
m
s
,
as
d
ef
in
ed
b
y
Gm
ail.
E
m
ails
in
th
ese
ca
teg
o
r
ie
s
ar
e
o
f
te
n
d
elete
d
with
o
u
t
b
ein
g
o
p
en
e
d
,
leav
in
g
u
s
er
s
f
ee
lin
g
o
v
er
wh
elm
ed
an
d
lead
in
g
th
em
to
c
h
ec
k
th
eir
in
b
o
x
es
less
f
r
eq
u
e
n
tly
,
p
o
ten
tially
m
is
s
in
g
im
p
o
r
tan
t e
m
ails
h
id
d
en
am
o
n
g
less
im
p
o
r
tan
t o
n
es.
T
o
ad
d
r
ess
th
is
is
s
u
e,
em
ail
p
r
o
v
id
er
s
h
av
e
in
tr
o
d
u
ce
d
v
a
r
io
u
s
f
ea
tu
r
es
to
h
elp
u
s
er
s
m
a
n
ag
e
th
ei
r
in
b
o
x
es
m
o
r
e
ef
f
ec
tiv
ely
.
T
wo
o
f
th
e
m
o
s
t
co
m
m
o
n
f
ea
tu
r
es
ar
e
em
ail
b
lo
c
k
in
g
a
n
d
s
p
am
/ju
n
k
f
o
ld
er
s
,
wh
er
e
u
s
er
s
ca
n
b
lo
c
k
em
ails
f
r
o
m
s
p
ec
if
ic
ad
d
r
ess
es
o
r
e
m
ails
th
at
m
atch
p
r
ed
ef
in
e
d
cr
it
er
ia.
Ad
d
itio
n
ally
,
em
ail
p
r
o
v
id
er
s
lik
e
Gm
ail
h
av
e
in
t
r
o
d
u
ce
d
f
u
r
t
h
er
clas
s
if
icatio
n
f
ea
tu
r
es
t
o
s
ep
ar
at
e
em
ails
r
elate
d
to
p
r
o
m
o
tio
n
s
,
u
p
d
ates
,
an
d
s
o
ci
al
ca
teg
o
r
ies
in
to
d
esig
n
ated
f
o
ld
er
s
.
T
h
is
allo
ws
u
s
er
s
to
m
an
ag
e
th
e
ir
em
ails
m
o
r
e
ea
s
ily
with
o
u
t
n
ee
d
in
g
to
o
p
en
th
em
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J
I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
20
2
6
:
6
6
5
-
6
7
3
666
E
m
ail
class
if
icatio
n
h
as
lo
n
g
b
ee
n
a
f
o
u
n
d
atio
n
al
asp
ec
t
o
f
em
ail
m
an
ag
em
e
n
t.
Ma
n
y
r
e
s
ea
r
ch
er
s
aim
ed
to
r
ed
u
ce
th
e
b
u
r
d
en
o
f
em
ail
m
an
ag
e
m
en
t
o
n
u
s
er
s
b
y
p
r
ed
ictin
g
em
ail
ca
teg
o
r
ies,
h
elp
in
g
u
s
er
s
d
ec
id
e
wh
eth
er
an
em
ail
is
i
m
p
o
r
tan
t.
T
h
is
ca
n
h
elp
im
p
r
o
v
e
tim
e
m
an
ag
em
en
t,
in
cr
ea
s
e
p
r
o
d
u
ctiv
ity
,
an
d
ef
f
ec
tiv
ely
f
ilter
h
a
r
m
f
u
l e
m
ails
,
s
u
ch
as th
o
s
e
co
n
tain
in
g
v
ir
u
s
-
in
f
ec
ted
attac
h
m
e
n
ts
o
r
cy
b
er
-
attac
k
co
n
te
n
t.
T
h
e
m
ain
ch
allen
g
e
in
m
an
ag
in
g
e
m
ails
to
d
ay
lie
s
in
t
h
e
o
v
er
wh
elm
in
g
v
o
lu
m
e
o
f
in
co
m
in
g
m
ess
ag
es,
wh
ich
o
f
te
n
ca
u
s
es
u
s
er
s
to
o
v
er
l
o
o
k
im
p
o
r
tan
t
em
ails
h
id
d
en
am
o
n
g
ir
r
elev
a
n
t
o
n
es.
T
h
is
s
tu
d
y
f
o
cu
s
es
o
n
ev
alu
atin
g
th
e
p
er
f
o
r
m
an
ce
o
f
p
o
p
u
lar
m
ac
h
i
n
e
lear
n
in
g
tech
n
iq
u
es
in
ter
m
s
o
f
b
o
th
ir
r
elev
an
t
em
ail
class
if
icatio
n
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
s
p
ee
d
,
u
s
in
g
f
ea
tu
r
es
d
er
iv
ed
f
r
o
m
em
ail
h
e
ad
er
-
o
n
ly
as
well
as
f
r
o
m
th
e
em
ail
h
ea
d
e
r
+b
o
d
y
.
I
n
co
n
tr
ast
to
ex
is
tin
g
s
tu
d
ies
th
at
p
r
im
ar
ily
f
o
c
u
s
o
n
class
if
y
in
g
s
p
am
o
r
m
alicio
u
s
em
ail
s
,
th
e
m
ain
f
o
cu
s
o
f
th
is
s
tu
d
y
i
s
o
n
class
if
y
in
g
ir
r
elev
an
t,
n
o
n
-
wo
r
k
-
r
el
ated
em
ails
.
Mo
r
eo
v
er
,
wh
ile
m
o
s
t
p
r
io
r
r
esear
ch
em
p
h
asizes
class
if
icatio
n
ac
cu
r
ac
y
,
th
ey
o
f
ten
o
v
e
r
lo
o
k
th
e
tr
ad
e
-
o
f
f
s
ass
o
ciate
d
with
p
r
o
ce
s
s
in
g
tim
e.
T
h
e
n
o
v
elty
o
f
th
is
r
esear
ch
lies
in
its
co
m
p
ar
ativ
e
a
n
aly
s
is
o
f
v
a
r
io
u
s
m
ac
h
i
n
e
lear
n
in
g
m
o
d
els
u
s
in
g
th
e
s
tan
d
ar
d
E
n
r
o
n
d
ataset,
aim
i
n
g
to
p
r
o
p
o
s
e
a
n
o
p
tim
al
an
d
e
f
f
icien
t
ap
p
r
o
ac
h
to
em
ail
cla
s
s
if
icatio
n
th
at
ca
n
b
e
ef
f
ec
tiv
ely
a
p
p
lied
in
r
ea
l
-
wo
r
ld
o
r
g
an
izatio
n
al
e
n
v
ir
o
n
m
en
ts
.
2.
E
M
A
I
L
F
E
A
T
UR
E
S
T
h
e
s
tr
u
ctu
r
e
o
f
an
e
m
ail,
s
h
o
wn
in
Fig
u
r
e
1
,
in
clu
d
e
s
b
o
th
h
ea
d
e
r
an
d
b
o
d
y
s
e
ctio
n
s
[
1
]
.
T
h
e
h
ea
d
e
r
in
clu
d
es
in
f
o
r
m
ati
o
n
s
u
ch
as
th
e
s
en
d
er
'
s
n
am
e,
r
ec
ip
ien
t'
s
n
am
e,
d
ate
an
d
tim
e,
I
P
s
er
v
er
s
en
d
e
r
,
an
d
I
P r
ec
eiv
e
r
,
wh
ile
th
e
b
o
d
y
co
n
tain
s
th
e
c
o
n
ten
t o
f
th
e
e
m
ail
[
2
]
.
C
o
m
m
o
n
em
ail
f
ea
tu
r
es
ca
n
b
e
g
r
o
u
ted
in
t
o
th
r
ee
m
ain
ca
te
g
o
r
ies:
h
ea
d
e
r
f
ea
tu
r
es,
co
n
te
n
t
f
ea
tu
r
es,
an
d
b
eh
a
v
io
r
al
f
ea
t
u
r
es.
E
ac
h
ca
teg
o
r
y
c
o
n
tr
ib
u
tes
to
e
v
alu
a
tin
g
th
e
s
ec
u
r
ity
an
d
tr
u
s
two
r
t
h
in
ess
o
f
an
em
ail.
Hea
d
er
f
ea
tu
r
es
an
aly
ze
th
e
e
m
ail
h
ea
d
er
,
s
u
ch
as
th
e
s
en
d
e
r
'
s
ad
d
r
ess
(
f
r
o
m
)
,
r
ec
ip
ie
n
t'
s
ad
d
r
ess
(
to
)
,
an
d
th
e
s
en
d
er
'
s
I
P
ad
d
r
ess
.
T
h
is
h
elp
s
v
er
if
y
th
e
em
ail'
s
au
th
en
ticity
an
d
p
r
ev
e
n
ts
s
en
d
er
s
p
o
o
f
in
g
.
C
o
n
ten
t
f
ea
tu
r
es
an
aly
ze
th
e
co
n
ten
t
o
f
th
e
e
m
ail,
s
u
ch
as
th
e
p
r
esen
ce
o
f
attac
h
m
en
ts
(
attac
h
m
en
t
p
r
esen
ce
)
,
k
e
y
wo
r
d
f
r
eq
u
e
n
cy
(
T
F
-
I
DF)
,
a
n
d
e
m
ail
co
n
te
n
t
a
n
aly
s
is
(
s
en
tim
en
t
an
aly
s
is
)
to
id
en
tify
p
o
ten
tial
r
is
k
s
r
elate
d
to
p
h
is
h
in
g
o
r
m
alwa
r
e.
B
eh
av
i
o
r
al
f
ea
tu
r
es
an
aly
ze
th
e
b
eh
av
io
r
o
f
th
e
em
ail,
s
u
ch
as
f
o
r
war
d
in
g
p
atter
n
s
,
wh
ich
m
ay
in
d
icate
in
ter
n
al
s
e
cu
r
ity
r
is
k
s
with
in
th
e
o
r
g
an
iz
atio
n
.
Fig
u
r
e
1
.
E
m
ail
s
tr
u
ctu
r
e
[
3
]
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
E
fficien
t e
ma
il c
la
s
s
ifica
tio
n
t
ec
h
n
iq
u
e:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f h
ea
d
er
-
o
n
ly
a
n
d
…
(
Wo
r
a
w
it K
i
tiku
s
o
u
n
)
667
I
n
r
esear
ch
o
n
e
m
ail
class
if
icatio
n
,
b
esid
es
d
etec
tin
g
s
p
am
a
n
d
p
h
is
h
in
g
,
th
e
class
if
icatio
n
o
f
"h
a
m
"
em
ails
––
non
-
s
p
a
m
m
ess
ag
es
th
at
ar
e
im
p
o
r
ta
n
t
to
th
e
r
ec
i
p
ien
t
––
is
also
cr
u
cial.
Ham
class
if
icatio
n
o
f
ten
r
elies
o
n
f
ea
tu
r
es
th
at
em
p
h
asi
ze
th
e
r
elev
an
ce
an
d
im
p
o
r
ta
n
ce
o
f
th
e
c
o
n
ten
t,
s
u
c
h
as
an
al
y
zin
g
k
e
y
ter
m
s
in
th
e
s
u
b
ject
an
d
b
o
d
y
,
an
d
th
e
f
r
eq
u
en
cy
o
f
co
n
tact
f
r
o
m
t
h
e
s
en
d
er
.
R
esear
ch
er
s
ca
n
u
s
e
f
ea
tu
r
es
d
er
iv
e
d
s
o
lely
f
r
o
m
th
e
h
ea
d
er
o
r
u
s
e
b
o
th
th
e
h
ea
d
er
an
d
b
o
d
y
co
n
ten
t
f
o
r
class
if
icatio
n
.
W
h
en
co
m
p
ar
in
g
cla
s
s
if
icatio
n
b
etwe
en
u
s
in
g
o
n
ly
th
e
h
ea
d
er
an
d
u
s
in
g
b
o
th
th
e
h
ea
d
er
an
d
f
u
ll c
o
n
ten
t,
it h
as b
ee
n
f
o
u
n
d
th
at
u
s
in
g
o
n
ly
th
e
h
ea
d
er
f
o
r
h
a
m
class
if
icatio
n
ca
n
p
r
o
d
u
ce
s
atis
f
ac
to
r
y
r
esu
lts
in
s
o
m
e
ca
s
es,
esp
ec
ially
wh
en
th
e
em
ail
h
as
a
clea
r
f
o
r
m
at
an
d
co
m
es
f
r
o
m
a
tr
u
s
ted
s
o
u
r
ce
[
4
]
.
Usi
n
g
b
o
th
h
ea
d
e
r
an
d
co
n
te
n
t
d
ata
g
en
er
ally
lea
d
s
to
h
i
g
h
er
ac
cu
r
ac
y
i
n
class
if
y
in
g
h
am
f
o
r
ce
r
tain
ca
s
es
,
esp
ec
ially
w
h
en
em
ail
c
o
n
tain
s
co
m
p
lex
co
n
ten
t
s
u
ch
as
im
a
g
es.
I
n
s
u
ch
ca
s
es
tex
t
v
is
ib
le
to
u
s
er
s
with
in
a
n
im
ag
e
ca
n
b
e
e
x
tr
ac
ted
u
s
in
g
o
p
tical
ch
ar
ac
ter
r
ec
o
g
n
itio
n
(
OC
R
)
tech
n
iq
u
es.
T
h
is
ex
tr
a
in
f
o
r
m
atio
n
ca
n
f
u
r
th
e
r
en
h
a
n
c
e
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
m
o
d
el
f
o
r
d
is
tin
g
u
is
h
in
g
b
etwe
en
'
h
am
'
an
d
'
s
p
am
'
[
1
]
.
Featu
r
e
ex
tr
ac
tio
n
is
a
cr
u
cial
s
tep
in
s
elec
tin
g
an
d
ex
tr
ac
ti
n
g
k
ey
c
h
ar
ac
ter
is
tics
f
r
o
m
th
e
d
ata.
Fo
r
ex
am
p
le,
ter
m
f
r
e
q
u
en
c
y
-
in
v
e
r
s
e
d
o
cu
m
e
n
t
f
r
e
q
u
e
n
c
y
(
TF
-
I
DF
)
is
a
s
tatis
tical
m
ea
s
u
r
e
u
s
ed
to
ev
alu
ate
th
e
im
p
o
r
tan
ce
o
f
a
wo
r
d
in
a
d
o
cu
m
en
t
co
m
p
ar
ed
to
th
e
en
tir
e
d
o
cu
m
en
t
s
et.
Ad
d
itio
n
ally
,
wo
r
d
e
m
b
ed
d
in
g
s
,
s
u
ch
as
W
o
r
d
2
Vec
[
5
]
an
d
Gl
o
Ve
[
6
]
,
ar
e
u
s
ed
to
r
e
p
r
esen
t
wo
r
d
s
in
a
co
n
tin
u
o
u
s
v
ec
to
r
s
p
ac
e,
wh
er
e
wo
r
d
s
with
s
im
ilar
m
ea
n
in
g
s
h
av
e
s
i
m
ilar
r
ep
r
esen
tatio
n
s
.
C
o
u
n
tin
g
th
e
n
u
m
b
er
o
f
wo
r
d
s
(
co
u
n
t
in
g
wo
r
d
s
)
is
also
p
ar
t
o
f
f
ea
tu
r
e
e
x
tr
ac
tio
n
,
wh
e
r
e
th
e
o
cc
u
r
r
e
n
ce
o
f
wo
r
d
s
in
a
d
o
cu
m
en
t
is
co
u
n
te
d
[
7
]
.
c
h
o
o
s
in
g
f
ea
tu
r
es
lik
e
s
en
d
er
(
f
r
o
m
)
,
r
ec
ip
ie
n
t
(
to
)
,
cc
,
b
cc
,
s
u
b
ject
(
s
u
b
ject)
,
b
o
d
y
(
b
o
d
y
)
,
a
n
d
s
en
d
er
ty
p
e
(
s
en
d
er
-
ty
p
e)
is
ess
en
tial,
esp
ec
ially
in
an
aly
zin
g
th
e
r
is
k
o
f
m
alicio
u
s
e
m
ails
,
s
u
ch
as
p
h
is
h
in
g
o
r
s
p
am
.
Ad
d
itio
n
ally
,
ch
ec
k
in
g
wh
eth
er
an
em
ail
h
as
b
ee
n
f
o
r
war
d
ed
o
r
r
e
p
lied
to
(
co
n
tain
s
-
r
ep
ly
-
f
o
r
war
d
s
)
is
an
o
t
h
er
f
ea
tu
r
e
ass
o
ciate
d
with
d
etec
tin
g
p
o
te
n
tially
h
ar
m
f
u
l e
m
ails
[
8
]
.
I
n
th
is
s
tu
d
y
,
we
ex
p
lo
r
e
th
e
u
s
e
o
f
f
ea
tu
r
es
ex
tr
ac
ted
f
r
o
m
h
ea
d
er
-
o
n
ly
an
d
h
ea
d
er
+b
o
d
y
o
f
th
e
em
ail
to
g
et
a
b
etter
u
n
d
er
s
tan
d
in
g
o
f
th
e
a
d
v
an
tag
es/d
is
ad
v
an
tag
es
o
f
th
e
two
ty
p
es
o
f
f
e
atu
r
es.
T
h
e
s
p
ec
if
ic
f
ea
tu
r
es e
x
tr
ac
ted
f
o
r
class
if
icatio
n
f
r
o
m
b
o
t
h
h
ea
d
e
r
s
an
d
f
u
ll c
o
n
ten
t
u
s
ed
i
n
th
is
s
tu
d
y
a
r
e
lis
ted
in
T
ab
le
1
.
T
ab
le
1
.
Featu
r
es e
x
tr
ac
te
d
f
r
o
m
em
ail
F
e
a
t
u
r
e
D
e
scri
p
t
i
o
n
H
e
a
d
e
r
-
o
n
l
y
H
e
a
d
e
r
+
b
ody
F
r
o
m
Th
e
e
m
a
i
l
se
n
d
e
r
's
a
d
d
r
e
ss
X
X
To
Th
e
r
e
c
i
p
i
e
n
t
's
e
m
a
i
l
a
d
d
r
e
ss
X
X
CC
Emai
l
a
d
d
r
e
ss
o
f
a
n
y
C
C
r
e
c
i
p
i
e
n
t
s
X
X
B
C
C
Emai
l
a
d
d
r
e
ss
o
f
a
n
y
B
C
C
r
e
c
i
p
i
e
n
t
s
X
X
I
P
a
d
d
r
e
ss
S
h
o
w
s
t
h
e
s
e
n
d
e
r
’
s I
P
a
d
d
r
e
ss
X
X
A
t
t
a
c
h
m
e
n
t
p
r
e
s
e
n
c
e
C
h
e
c
k
s w
h
e
t
h
e
r
t
h
e
e
ma
i
l
c
o
n
t
a
i
n
s a
t
t
a
c
h
me
n
t
s
,
w
h
i
c
h
m
a
y
c
o
n
t
a
i
n
m
a
l
i
c
i
o
u
s fi
l
e
s
X
X
S
e
n
d
e
r
t
y
p
e
e
.
g
.
,
o
r
g
a
n
i
z
a
t
i
o
n
o
r
i
n
d
i
v
i
d
u
a
l
X
X
C
o
n
t
a
i
n
s
r
e
p
l
y
/
f
o
r
w
a
r
d
s
D
e
t
e
c
t
s r
e
p
l
i
e
s
o
r
f
o
r
w
a
r
d
e
d
e
ma
i
l
s
,
w
h
i
c
h
ma
y
i
n
d
i
c
a
t
e
su
sp
i
c
i
o
u
s
f
o
r
w
a
r
d
i
n
g
p
r
a
c
t
i
c
e
s
X
X
D
K
I
M
/
S
P
F
a
u
t
h
e
n
t
i
c
a
t
i
o
n
U
se
o
f
D
o
m
a
i
n
K
e
y
s I
d
e
n
t
i
f
i
e
d
M
a
i
l
o
r
S
e
n
d
e
r
P
o
l
i
c
y
F
r
a
mew
o
r
k
X
X
K
e
y
w
o
r
d
f
r
e
q
u
e
n
c
y
(
TF
-
I
D
F
)
F
r
e
q
u
e
n
c
y
o
f
i
m
p
o
r
t
a
n
t
k
e
y
w
o
r
d
s t
h
e
e
mai
l
b
o
d
y
f
o
r
d
e
t
e
c
t
i
n
g
si
g
n
i
f
i
c
a
n
t
w
o
r
d
p
a
t
t
e
r
n
s
X
S
e
n
t
i
me
n
t
a
n
a
l
y
si
s
Th
e
t
o
n
e
o
f
t
h
e
e
m
a
i
l
(
p
o
s
i
t
i
v
e
,
n
e
g
a
t
i
v
e
,
o
r
n
e
u
t
r
a
l
)
X
C
o
n
t
e
n
t
l
e
n
g
t
h
Le
n
g
t
h
o
f
t
h
e
e
ma
i
l
c
o
n
t
e
n
t
X
A
t
t
a
c
h
m
e
n
t
s
a
n
a
l
y
s
i
s
S
c
a
n
r
e
s
u
l
t
o
f
t
h
e
a
t
t
a
c
h
e
d
f
i
l
e
s f
o
r
p
o
t
e
n
t
i
a
l
t
h
r
e
a
t
s
o
r
m
a
l
w
a
r
e
X
H
y
p
e
r
l
i
n
k
p
r
e
se
n
c
e
I
d
e
n
t
i
f
i
e
s
a
n
y
h
y
p
e
r
l
i
n
k
s
i
n
t
h
e
e
m
a
i
l
X
T
h
e
s
elec
tio
n
o
f
f
ea
tu
r
es
f
o
r
em
ail
class
if
icatio
n
an
aly
s
is
in
th
is
tab
le
is
d
iv
i
d
ed
in
t
o
two
p
ar
ts
:
h
ea
d
er
-
o
n
ly
an
d
h
ea
d
er
+
b
o
d
y
.
T
h
e
h
ea
d
er
-
o
n
ly
a
p
p
r
o
ac
h
f
o
cu
s
es
o
n
d
ata
th
at
ca
n
id
en
ti
f
y
th
e
s
o
u
r
ce
o
f
th
e
em
ail,
s
u
ch
as
f
r
o
m
,
to
,
s
u
b
je
ct
,
I
P
a
d
d
r
ess
,
an
d
attac
h
m
en
t
p
r
esen
ce
.
T
h
is
in
f
o
r
m
atio
n
c
an
ef
f
ec
tiv
ely
ass
is
t
in
class
if
y
in
g
em
ails
,
esp
ec
ial
ly
in
ca
s
es
o
f
d
etec
tin
g
s
p
am
an
d
p
h
is
h
in
g
em
ails
wit
h
h
ig
h
ac
cu
r
ac
y
.
T
h
e
I
P
a
d
d
r
ess
is
a
cr
u
cial
p
iece
o
f
in
f
o
r
m
ati
o
n
u
s
ed
to
tr
ac
e
th
e
em
ail’
s
tr
an
s
m
is
s
io
n
p
ath
,
allo
win
g
f
o
r
th
e
id
en
tific
atio
n
o
f
s
u
s
p
icio
u
s
s
o
u
r
ce
s
.
Similar
ly
,
th
e
attac
h
m
en
t
p
r
esen
ce
h
elp
s
d
eter
m
in
e
wh
eth
er
th
e
em
ail
co
n
tain
s
p
o
te
n
tially
h
ar
m
f
u
l
at
tach
m
en
ts
,
an
d
th
e
s
en
d
er
ty
p
e
d
is
tin
g
u
is
h
es
wh
eth
er
th
e
e
m
ail
o
r
ig
in
ates
f
r
o
m
a
n
i
n
t
e
r
n
a
l
o
r
e
x
t
e
r
n
a
l
s
o
u
r
c
e
,
m
a
k
i
n
g
i
t
e
f
f
e
c
t
i
v
e
f
o
r
d
e
t
e
c
t
i
n
g
e
x
t
e
r
n
a
l
a
t
t
a
c
k
s
.
T
h
e
u
s
e
o
f
D
K
I
M
[
9
]
o
r
S
P
F
[
1
0
]
au
th
en
ticatio
n
h
elp
s
v
er
if
y
th
e
leg
itima
cy
o
f
th
e
s
en
d
er
’
s
d
o
m
ain
b
y
en
s
u
r
in
g
th
at
a
r
el
iab
le
au
th
en
ticatio
n
m
eth
o
d
h
as
b
ee
n
u
s
ed
to
p
r
e
v
en
t
em
a
il
s
p
o
o
f
i
n
g
.
W
h
en
c
o
n
s
id
er
in
g
th
e
h
ea
d
er
+
b
o
d
y
a
p
p
r
o
ac
h
,
t
h
e
em
ail
s
tr
u
ctu
r
e
ca
n
b
ec
o
m
e
q
u
ite
c
o
m
p
lex
[1
1
]
.
Ad
d
itio
n
al
f
ea
t
u
r
e
s
r
elate
d
to
th
e
e
m
ail
co
n
ten
t
ar
e
o
f
te
n
co
n
v
er
te
d
in
to
n
u
m
er
ical
o
r
v
ec
to
r
ized
f
o
r
m
ats
f
o
r
f
u
r
th
er
p
r
o
ce
s
s
in
g
.
E
x
am
p
les
o
f
s
u
c
h
f
ea
tu
r
es
i
n
c
l
u
d
e
k
e
y
w
o
r
d
f
r
e
q
u
e
n
c
y
(
e
.
g
.
,
TF
-
I
D
F
)
,
s
e
n
t
i
m
e
n
t
s
c
o
r
e
s
[1
2
]
c
o
n
t
e
n
t
l
e
n
g
t
h
a
n
d
p
r
e
s
e
n
c
e
o
f
h
y
p
e
r
l
i
n
k
s
[
2
]
.
Ho
wev
er
,
s
o
m
e
f
ea
tu
r
es,
s
u
ch
as M
I
ME
-
Ver
s
io
n
[1
3
]
,
wh
ich
in
d
icate
s
th
e
v
er
s
io
n
o
f
m
u
ltime
d
ia
co
n
ten
t,
co
n
ten
t
-
ty
p
e
,
wh
ich
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J
I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
20
2
6
:
6
6
5
-
6
7
3
668
s
p
ec
if
ies
th
e
ty
p
e
o
f
co
n
ten
t
i
n
th
e
em
ail,
a
n
d
X
-
Prio
r
ity
,
wh
ich
s
ig
n
if
ies
th
e
p
r
i
o
r
ity
lev
el
o
f
th
e
em
ail,
a
r
e
n
o
t u
s
ed
b
ec
a
u
s
e
th
ey
d
o
n
o
t c
o
n
tr
ib
u
te
to
im
p
r
o
v
in
g
th
e
ac
c
u
r
ac
y
o
f
em
ail
class
if
icatio
n
.
R
elate
d
wo
r
k
:
m
ac
h
in
e
lear
n
in
g
(
ML
)
tec
h
n
i
q
u
es
h
av
e
b
ee
n
ex
ten
s
iv
ely
em
p
lo
y
e
d
in
em
ail
class
if
icatio
n
,
u
tili
zin
g
v
ar
io
u
s
f
ea
tu
r
es
f
r
o
m
e
m
ail
h
ea
d
e
r
s
,
b
o
d
ies,
o
r
th
eir
co
m
b
i
n
atio
n
(
Hea
d
er
+
B
o
d
y
)
.
T
h
e
ch
o
ice
o
f
f
ea
tu
r
es
a
n
d
d
a
tasets
s
ig
n
if
ican
tly
in
f
lu
en
ce
s
th
e
p
er
f
o
r
m
an
ce
o
f
ML
class
if
ier
s
.
T
h
i
s
s
ec
tio
n
r
ev
iews
n
o
tab
le
s
tu
d
ies
b
y
ca
t
eg
o
r
izin
g
th
em
b
ased
o
n
ML
tech
n
iq
u
es,
f
ea
tu
r
es
u
s
ed
,
d
at
asets
,
an
d
r
e
p
o
r
ted
p
er
f
o
r
m
an
ce
s
.
Hea
d
er
-
o
n
l
y
em
ail
class
if
icat
i
o
n
r
elies
o
n
m
etad
ata
s
u
ch
as
s
en
d
er
in
f
o
r
m
atio
n
,
em
ail
s
u
b
ject,
an
d
r
o
u
tin
g
p
ath
s
.
T
h
e
lig
h
twe
ig
h
t
an
d
s
im
p
le
n
atu
r
e
o
f
h
ea
d
er
f
ea
tu
r
es
allo
ws
f
o
r
e
f
f
icien
t
s
p
am
d
etec
tio
n
.
T
h
e
s
tu
d
ies
co
n
d
u
cted
b
y
B
ea
m
a
n
an
d
I
s
ah
[1
4
]
,
a
n
d
Al
-
J
ar
r
a
h
et
a
l
.
[1
5
]
,
s
h
o
wed
t
h
at
r
a
n
d
o
m
f
o
r
est
m
o
d
el
p
er
f
o
r
m
s
b
est
in
class
if
y
i
n
g
em
ails
b
ased
s
o
lely
o
n
h
ea
d
er
-
o
n
ly
f
ea
tu
r
es.
Similar
ly
,
Fan
g
et
a
l.
[1
1
]
d
em
o
n
s
tr
ated
th
at
co
m
b
in
in
g
h
ea
d
er
f
ea
tu
r
es
with
B
i
-
L
S
T
M
an
d
atten
tio
n
m
ec
h
an
is
m
s
in
th
e
T
HE
MI
S
m
o
d
el
ac
h
ie
v
ed
a
s
p
am
cla
s
s
if
icatio
n
ac
cu
r
ac
y
o
f
u
p
t
o
9
9
.
8
4
8
%
o
n
t
h
e
I
W
SP
A
-
AP
2
0
1
8
d
ataset.
Ad
d
itio
n
ally
,
Fer
n
a
n
d
ez
et
a
l.
[
4
]
p
r
o
p
o
s
ed
lev
e
r
ag
in
g
em
ail
-
r
elate
d
d
ata
b
y
in
te
g
r
atin
g
p
ass
iv
e
DNS
in
f
o
r
m
atio
n
an
d
d
o
m
ai
n
SP
F r
u
les,
en
ab
lin
g
th
e
r
ap
id
d
etec
tio
n
o
f
s
p
am
d
o
m
ain
s
.
On
th
e
o
th
er
h
an
d
,
f
u
ll
-
co
n
te
n
t
em
ail
class
if
icatio
n
in
co
r
p
o
r
ates
in
f
o
r
m
atio
n
f
r
o
m
b
o
t
h
th
e
h
ea
d
er
an
d
b
o
d
y
o
f
em
ails
,
en
ab
lin
g
m
o
r
e
co
m
p
r
eh
en
s
iv
e
s
em
a
n
tic
an
d
s
tatis
tical
an
aly
s
is
.
T
h
is
ap
p
r
o
ac
h
is
p
ar
ticu
lar
ly
ef
f
ec
tiv
e
in
d
etec
t
in
g
s
o
p
h
is
ticated
p
h
is
h
in
g
tech
n
iq
u
es,
s
u
ch
as
m
alicio
u
s
p
ay
lo
ad
s
em
b
ed
d
e
d
in
em
ail
b
o
d
ies.
Ma
jo
r
ity
o
f
ex
is
tin
g
s
tu
d
ies ex
p
lo
r
ed
h
y
b
r
id
m
o
d
els th
at
co
m
b
in
ed
at
least t
wo
well
-
estab
lis
h
ed
m
ac
h
in
e
lear
n
in
g
m
o
d
els.
N
o
tab
le
ex
am
p
les
o
f
s
u
ch
r
es
ea
r
ch
in
clu
d
e
th
e
f
o
llo
win
g
:
Gib
s
o
n
et
a
l.
[1
6
]
ap
p
lied
b
io
-
i
n
s
p
ir
ed
o
p
tim
izat
io
n
tech
n
iq
u
es,
p
a
r
ticle
s
war
m
o
p
tim
izatio
n
(
PS
O)
an
d
g
en
et
ic
alg
o
r
ith
m
(
GA)
,
to
f
iv
e
m
ac
h
in
e
lear
n
in
g
m
o
d
els
m
u
ltin
o
m
ial
Naïv
e
B
ay
es
(
MN
B
)
,
s
to
ch
asti
c
g
r
ad
ien
t
d
e
s
ce
n
t
(
SGD)
,
d
ec
is
io
n
tr
ee
(
DT
)
,
r
a
n
d
o
m
f
o
r
est
(
R
F),
an
d
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
(
ML
P)
an
d
d
em
o
n
s
tr
ated
th
at
s
u
ch
tech
n
iq
u
es
ca
n
im
p
r
o
v
e
th
e
cl
ass
if
icatio
n
ac
cu
r
ac
y
o
f
all
th
e
m
o
d
els
co
n
s
id
er
e
d
.
I
n
p
ar
tic
u
lar
,
PS
O
to
g
eth
er
with
MN
B
ca
n
ac
h
iev
e
u
p
t
o
1
0
0
%
ac
cu
r
ac
y
o
n
th
e
Sp
am
Ass
as
s
in
d
ataset.
L
ik
ewise,
SeFAC
E
D
[1
7
]
,
a
f
r
am
ewo
r
k
u
s
in
g
l
o
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
b
a
s
ed
g
ated
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
(
GR
U)
f
o
r
s
em
an
tic
an
aly
s
is
,
ac
h
iev
ed
9
5
%
ac
cu
r
ac
y
with
a
b
alan
ce
d
p
r
ec
is
io
n
an
d
r
ec
all
b
y
u
tili
zin
g
b
o
t
h
h
ea
d
e
r
an
d
b
o
d
y
f
ea
tu
r
es
f
r
o
m
th
e
E
n
r
o
n
d
ataset
an
d
C
L
AI
R
co
llectio
n
o
f
f
r
a
u
d
em
ails
d
ataset.
Similar
ly
,
h
y
b
r
id
d
ee
p
lear
n
in
g
m
o
d
els
s
u
ch
as
d
ee
p
n
eu
r
al
n
etwo
r
k
with
b
id
ir
ec
t
io
n
al
lo
n
g
s
h
o
r
t
-
ter
m
m
o
m
o
r
y
(
DNN
-
B
iLST
M)
co
m
b
in
ed
f
ea
tu
r
es
f
r
o
m
b
o
t
h
h
ea
d
er
s
an
d
b
o
d
ies
to
class
if
y
em
ails
an
d
d
etec
t
em
o
ti
o
n
al
to
n
e,
ac
h
iev
in
g
9
6
.
3
9
% a
cc
u
r
ac
y
o
n
th
e
E
n
r
o
n
d
atasets
[1
2
]
.
Gh
aleb
et
a
l.
[1
8
]
em
p
lo
y
e
d
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
with
en
h
an
ce
d
g
r
ass
h
o
p
p
e
r
o
p
t
im
izatio
n
a
lg
o
r
ith
m
s
(
E
GOA)
to
o
p
ti
m
ize
f
ea
tu
r
e
s
elec
tio
n
f
r
o
m
b
o
th
h
ea
d
er
s
an
d
b
o
d
ies,
r
esu
ltin
g
in
98
.1
%
class
if
icatio
n
ac
cu
r
ac
ies
o
n
Sp
am
Ass
a
s
s
in
d
ataset
with
9
7
.
8
%
d
etec
tio
n
r
ates
.
Hy
b
r
id
m
o
d
els
u
s
in
g
h
ier
ar
ch
ical
atten
tio
n
m
ec
h
an
i
s
m
s
,
as
d
em
o
n
s
tr
ated
b
y
Z
av
r
ak
a
n
d
Yilm
az
[1
9
]
,
co
m
b
in
e
d
C
NNs
an
d
GR
U
s
to
cla
s
s
if
y
s
p
am
s
u
s
in
g
f
iv
e
wid
ely
u
s
ed
d
atasets
(
T
R
E
C
2
0
0
7
,
Gen
Sp
am
,
Sp
am
Ass
ass
in
,
E
n
r
o
n
,
an
d
L
in
g
Sp
am
)
,
ac
h
iev
e
AUC
o
f
u
p
t
o
0
.
9
5
7
with
an
av
er
a
g
e
o
f
0
.
8
0
6
in
c
r
o
s
s
-
d
ataset
ev
alu
atio
n
s
.
E
v
en
th
e
b
asic
ap
p
r
o
ac
h
o
f
in
teg
r
atin
g
two
co
n
v
en
tio
n
al
m
o
d
els,
Naïv
e
B
a
y
es a
n
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
(
ANN)
,
h
as b
ee
n
s
h
o
wn
to
s
lig
h
tly
o
u
tp
e
r
f
o
r
m
s
in
d
iv
id
u
al
m
o
d
els
[
20
]
.
Sp
ec
if
ically
,
th
e
h
y
b
r
id
NB
-
ANN
m
o
d
el
ac
h
iev
e
d
an
ac
cu
r
ac
y
o
f
9
9
.
0
1
%,
o
u
tp
e
r
f
o
r
m
in
g
t
h
e
s
tan
d
alo
n
e
ANN
an
d
NB
m
o
d
els,
wh
ic
h
attain
ed
ac
cu
r
ac
ies
o
f
9
8
.
5
7
% a
n
d
9
8
.
1
2
%,
r
esp
ec
tiv
ely
.
A
s
y
s
tem
atic
r
ev
iew
o
f
h
o
w
co
m
b
in
atio
n
o
f
m
ac
h
in
e
lear
n
in
g
tech
n
iq
u
es
ca
n
b
e
u
s
ed
to
im
p
r
o
v
e
s
p
am
/h
am
em
ail
class
if
icatio
n
is
f
u
r
t
h
er
s
u
p
p
o
r
te
d
b
y
s
tu
d
ies
in
[
2
1]
.
Said
an
d
Allan
s
tu
d
ie
d
3
7
r
esear
ch
wo
r
k
r
elate
d
to
p
h
is
h
in
g
web
s
ite,
em
ail,
an
d
SMS
attac
k
s
p
u
b
lis
h
ed
b
etwe
en
2
0
1
9
a
n
d
2
0
2
3
a
n
d
f
o
u
n
d
t
h
at
Stack
in
g
an
d
Ad
ab
o
o
s
t
en
s
em
b
le
m
eth
o
d
s
ar
e
p
o
p
u
lar
f
o
r
d
ev
elo
p
in
g
p
h
is
h
i
n
g
em
ail
clas
s
if
icatio
n
m
o
d
el.
I
n
ad
d
itio
n
,
p
o
p
u
lar
m
ac
h
in
e
lea
r
n
in
g
m
o
d
els
t
o
b
e
u
s
ed
in
e
n
s
em
b
le
m
eth
o
d
s
ar
e
m
u
ltin
o
m
ial
Naïv
e
B
ay
es,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e,
an
d
r
an
d
o
m
f
o
r
est.
Ho
wev
er
,
th
e
f
in
d
in
g
s
o
f
[
2
]
in
d
icate
th
at
n
o
s
in
g
le
class
if
ier
co
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
s
o
th
e
r
s
ac
r
o
s
s
all
s
ce
n
ar
io
s
,
d
u
e
to
th
e
v
ar
iab
ilit
y
o
f
d
ep
lo
y
m
en
t
en
v
ir
o
n
m
en
t
s
an
d
th
e
ev
o
l
v
in
g
n
atu
r
e
o
f
attac
k
t
ec
h
n
iq
u
es.
C
o
n
s
eq
u
en
tly
,
p
er
i
o
d
ic
r
etr
ai
n
in
g
o
f
m
o
d
els
is
es
s
en
tial
to
m
ain
tain
th
eir
ef
f
ec
tiv
en
ess
.
E
x
is
tin
g
s
tu
d
ies
s
u
g
g
ested
th
at
em
ail
h
ea
d
er
s
alo
n
e
ca
n
ef
f
ec
tiv
ely
class
if
y
em
ails
with
o
u
t
r
ely
in
g
on
th
e
em
ail
b
o
d
y
c
o
n
ten
t.
I
n
th
is
s
tu
d
y
,
we
f
u
r
th
er
ex
p
lo
r
e
th
e
tr
ad
e
-
o
f
f
in
p
e
r
f
o
r
m
an
ce
a
n
d
p
r
o
ce
s
s
in
g
tim
e
b
etwe
en
u
s
in
g
f
ea
t
u
r
es
d
er
iv
ed
f
r
o
m
em
ail
h
ea
d
er
s
alo
n
e
an
d
th
o
s
e
d
er
iv
e
d
f
r
o
m
b
o
th
em
ail
h
ea
d
er
s
an
d
b
o
d
ies
f
o
r
t
h
e
class
if
icatio
n
o
f
s
p
am
an
d
h
am
em
ails
.
T
a
b
le
2
p
r
esen
ts
a
co
m
p
ar
ativ
e
s
u
m
m
ar
y
o
f
p
r
e
v
io
u
s
s
tu
d
ies
o
n
em
ail
class
if
icatio
n
,
h
ig
h
lig
h
tin
g
th
e
ty
p
es
o
f
f
ea
tu
r
es
u
s
ed
,
th
e
m
ac
h
in
e
l
ea
r
n
in
g
tech
n
i
q
u
es
ap
p
lied
,
th
e
u
s
e
o
f
h
ea
d
e
r
in
f
o
r
m
atio
n
,
th
e
in
clu
s
io
n
o
f
p
r
o
ce
s
s
in
g
tim
e
as
an
ev
alu
atio
n
m
etr
ic,
an
d
th
e
k
ey
d
if
f
er
en
ce
s
co
m
p
ar
ed
to
th
e
c
u
r
r
en
t stu
d
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
E
fficien
t e
ma
il c
la
s
s
ifica
tio
n
t
ec
h
n
iq
u
e:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f h
ea
d
er
-
o
n
ly
a
n
d
…
(
Wo
r
a
w
it K
i
tiku
s
o
u
n
)
669
T
ab
le
2
.
C
o
m
p
a
r
is
o
n
o
f
r
ec
en
t
em
ail
class
if
icatio
n
ap
p
r
o
ac
h
es
R
e
f
e
r
e
n
c
e
D
a
t
a
s
o
u
r
c
e
H
e
a
d
e
r
o
n
l
y
F
u
l
l
c
o
n
t
e
n
t
M
L
t
e
c
h
n
i
q
u
e
s
R
u
n
-
t
i
me
a
n
a
l
y
si
s
[
4
]
C
o
r
p
o
r
a
t
e
l
o
g
r
u
l
e
b
a
se
Y
e
s
No
D
T,
R
F
No
[
1
1
]
En
r
o
n
D
a
t
a
set
,
S
p
a
mA
ss
a
ssi
n
No
Y
e
s
C
N
N
+
Bi
-
LS
TM
+
A
t
t
e
n
t
i
o
n
No
[
1
2
]
En
r
o
n
C
o
r
p
o
r
a
,
P
h
i
sh
e
d
,
O
f
f
e
n
si
v
e
d
a
t
a
s
e
t
No
Y
e
s
H
y
b
r
i
d
D
N
N
(
C
N
N
+
LST
M
)
,
Emo
t
i
o
n
D
e
t
e
c
t
i
o
n
No
[
1
4
]
TR
E
C
2
0
0
7
c
o
r
p
u
s
,
P
h
i
s
h
i
n
g
e
mai
l
s fr
o
m
2
0
1
7
–
2
0
2
0
Y
e
s
No
R
F
,
S
V
M
,
M
LP
,
K
N
N
No
[
1
5
]
C
EA
S
2
0
0
8
a
n
d
C
S
D
M
C
2
0
1
0
sp
a
m
d
a
t
a
set
s
Y
e
s
No
D
T,
S
V
M
,
M
P
,
N
B
,
B
N
,
R
F
No
[
1
6
]
Li
n
g
-
S
p
a
m,
P
U
1
-
3
,
P
U
A
S
p
a
mA
ss
a
ssi
n
,
En
r
o
n
No
Y
e
s
M
N
B
,
S
t
o
c
h
a
s
t
i
c
G
r
a
d
i
e
n
t
D
e
s
c
e
n
t
(
S
G
D
)
,
D
T,
R
F
,
M
LP +
P
S
O
a
n
d
GA
No
[
1
7
]
En
r
o
n
a
n
d
C
LA
I
R
c
o
l
l
e
c
t
i
o
n
o
f
f
r
a
u
d
e
m
a
i
l
No
Y
e
s
LSTM
+
G
R
U
(
S
e
F
A
C
ED
)
No
[
1
8
]
S
p
a
m
B
a
s
e
,
S
p
a
mA
ss
a
ssi
n
,
UK
-
2
0
1
1
No
Y
e
s
M
LP w
i
t
h
e
n
h
a
n
c
e
d
G
r
a
ssh
o
p
p
e
r
o
p
t
i
m
i
z
a
t
i
o
n
No
[
1
9
]
TR
E
C
2
0
0
7
,
G
e
n
S
p
a
m
,
S
p
a
mA
ss
a
ssi
n
,
En
r
o
n
,
L
i
n
g
S
p
a
m
No
Y
e
s
F
a
st
T
e
x
t
+
H
i
e
r
a
r
c
h
i
c
a
l
A
t
t
e
n
t
i
o
n
H
y
b
r
i
d
N
e
u
r
a
l
N
e
t
w
o
r
k
s
No
[
2
0
]
K
a
g
g
l
e
No
Y
e
s
H
y
b
r
i
d
N
B
-
ANN
No
[
2
1
]
En
r
o
n
,
U
C
I
,
H
ELPH
ED
,
S
p
a
mA
ss
a
ssi
n
No
Y
e
s
A
d
a
B
o
o
st
,
B
a
g
g
i
n
g
,
G
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
No
Th
i
s
S
t
u
d
y
En
r
o
n
D
a
t
a
set
Y
e
s
Y
e
s
1
1
M
o
d
e
l
s
Y
e
s
3.
RE
S
E
ARCH
M
E
T
H
O
DS
I
n
th
is
s
ec
tio
n
,
it
is
ex
p
lain
ed
th
e
r
esu
lts
o
f
r
esear
c
h
an
d
at
t
h
e
s
am
e
tim
e
is
g
iv
en
th
e
co
m
p
r
eh
e
n
s
iv
e
d
is
cu
s
s
io
n
.
R
esu
lts
ca
n
b
e
p
r
esen
ted
in
f
ig
u
r
es,
g
r
ap
h
s
,
tab
les
an
d
o
th
er
s
t
h
at
m
ak
e
th
e
r
ea
d
er
u
n
d
er
s
tan
d
ea
s
ily
[1
5
]
,
[
1
6
]
.
T
h
e
d
is
cu
s
s
io
n
ca
n
b
e
m
ad
e
in
s
ev
er
al
s
u
b
-
s
ec
tio
n
s
.
3
.
1
.
Cho
o
s
ing
e
m
a
il da
t
a
s
et
T
h
e
E
n
r
o
n
d
ataset
[
22
]
is
o
n
e
o
f
t
h
e
m
o
s
t
wid
ely
u
s
ed
d
a
tasets
f
o
r
em
ail
class
if
icatio
n
b
ec
au
s
e
it
o
r
ig
in
ates
f
r
o
m
r
ea
l
in
ter
n
al
co
m
m
u
n
icatio
n
s
with
in
th
e
E
n
r
o
n
o
r
g
an
izatio
n
.
T
h
e
d
ata
s
et
co
m
p
r
is
es
o
v
er
5
0
0
,
0
0
0
em
ails
f
r
o
m
m
o
r
e
th
a
n
1
5
0
em
p
lo
y
ee
s
.
T
h
is
d
ataset
h
as
b
ee
n
ex
ten
s
iv
e
ly
u
s
ed
in
r
esear
ch
d
u
e
to
t
h
e
d
iv
er
s
ity
o
f
em
ail
ty
p
es
it
co
n
tain
s
,
in
clu
d
in
g
r
eg
u
lar
em
a
ils
(
h
am
)
,
s
p
am
,
an
d
p
h
is
h
in
g
.
Ad
d
itio
n
ally
,
th
e
E
n
r
o
n
d
ataset
is
p
u
b
licly
ac
ce
s
s
ib
le,
m
ak
in
g
it
ea
s
y
to
co
m
p
ar
e
r
esear
ch
r
esu
lts
an
d
p
r
o
v
id
in
g
a
s
tan
d
ar
d
f
o
r
test
in
g
m
ac
h
in
e
lear
n
in
g
m
o
d
els
[
2
1]
,
[
23
]
,
[
24
]
.
B
esid
es
t
h
e
E
n
r
o
n
d
ataset,
o
th
er
d
atasets
h
av
e
also
b
ee
n
u
s
ed
ex
ten
s
iv
ely
in
r
esear
c
h
s
u
ch
as
th
e
Sp
am
Ass
a
s
s
in
Pu
b
lic
C
o
r
p
u
s
[
25
]
wh
ich
f
o
cu
s
es
o
n
s
p
am
class
if
icatio
n
an
d
th
e
L
in
g
-
S
p
am
d
ataset,
u
s
ed
in
s
p
ec
ialized
ac
ad
em
ic
f
ield
s
.
T
h
ese
d
atasets
ar
e
ch
o
s
en
b
ased
o
n
th
e
r
esear
ch
o
b
jecti
v
es,
s
u
ch
as
g
en
er
al
s
p
am
d
e
tectio
n
o
r
p
h
is
h
in
g
d
etec
tio
n
,
with
ea
ch
d
ataset
h
av
in
g
u
n
iq
u
e
ch
ar
ac
te
r
is
tics
th
at
h
elp
en
h
an
ce
th
e
m
o
d
el'
s
ef
f
ec
tiv
en
ess
in
em
ail
class
if
icatio
n
[
2
4
].
I
n
th
is
r
esear
ch
,
th
e
E
n
r
o
n
d
ataset
was
s
elec
ted
b
ec
au
s
e
it
r
ef
lects
r
ea
l
in
te
r
n
al
o
r
g
a
n
izatio
n
al
co
m
m
u
n
icatio
n
,
co
n
s
is
tin
g
o
f
em
ails
with
d
iv
er
s
e
co
n
ten
t
f
r
o
m
ac
tu
al
u
s
ag
e
c
o
n
te
x
ts
.
T
h
is
allo
ws
ea
ch
s
elec
ted
m
o
d
els
to
b
e
ef
f
ec
tiv
ely
ap
p
lied
in
r
ea
l
-
wo
r
l
d
s
ce
n
ar
io
s
.
Fu
r
th
er
m
o
r
e
,
th
e
d
ataset
is
lar
g
e
en
o
u
g
h
t
o
test
co
m
p
lex
m
ac
h
in
e
lear
n
in
g
m
o
d
els.
T
h
e
E
n
r
o
n
d
ataset
i
s
also
well
-
r
ec
o
g
n
ized
in
th
e
r
esear
ch
co
m
m
u
n
ity
,
en
ab
lin
g
th
e
r
esu
lts
to
b
e
co
m
p
ar
ed
with
o
t
h
er
s
tu
d
ies ef
f
ec
t
iv
ely
.
3.
2
.
Da
t
a
p
re
-
pro
ce
s
s
ing
I
n
th
e
f
ir
s
t
s
tep
,
d
ata
p
r
ep
ar
atio
n
in
v
o
lv
es
id
en
tify
i
n
g
an
d
h
an
d
lin
g
m
is
s
in
g
v
alu
es,
wh
ich
m
ay
in
clu
d
e
r
ep
lacin
g
m
is
s
in
g
v
al
u
es,
co
n
v
er
tin
g
d
ata
(
e.
g
.
,
c
o
n
v
er
tin
g
all
tex
t
to
l
o
wer
ca
s
e,
r
em
o
v
in
g
s
p
ec
ial
ch
ar
ac
ter
s
,
an
d
s
tan
d
ar
d
izin
g
d
ate
f
o
r
m
ats).
Nex
t,
d
ata
cle
an
in
g
is
p
er
f
o
r
m
ed
to
d
etec
t
an
d
c
o
r
r
ec
t
e
r
r
o
r
s
,
s
u
ch
as
r
em
o
v
in
g
d
u
p
licate
d
a
ta,
m
an
ag
in
g
m
is
s
in
g
d
ata,
an
d
co
r
r
ec
tin
g
i
n
ap
p
r
o
p
r
iate
v
al
u
es.
Af
te
r
war
d
,
th
e
tex
t
is
b
r
o
k
en
d
o
wn
in
to
s
m
al
ler
u
n
its
ca
lled
"to
k
en
s
"
(
t
o
k
e
n
izatio
n
)
[
26
]
,
wh
ich
ca
n
b
e
wo
r
d
s
,
s
en
ten
ce
s
,
o
r
o
th
er
c
o
m
p
o
n
en
ts
d
e
p
en
d
in
g
o
n
th
e
m
o
d
el'
s
r
eq
u
ir
em
en
ts
.
T
h
e
n
ex
t
s
tep
is
to
r
e
m
o
v
e
u
n
im
p
o
r
tan
t
wo
r
d
s
f
r
o
m
th
e
an
aly
s
is
(
s
to
p
wo
r
d
r
em
o
v
al
)
,
s
u
ch
as c
o
m
m
o
n
wo
r
d
s
th
at
u
s
u
ally
d
o
n
o
t
af
f
ec
t th
e
m
ain
m
ea
n
in
g
o
f
th
e
tex
t,
lik
e
"a
n
d
,
"
"th
e,
"
an
d
"is".
Fin
ally
,
wo
r
d
s
a
r
e
r
ed
u
ce
d
to
th
eir
b
ase
f
o
r
m
(
lem
m
atiza
tio
n
o
r
s
tem
m
in
g
)
to
s
im
p
lify
t
h
e
d
ata
an
d
f
ac
ilit
ate
ea
s
ier
p
r
o
ce
s
s
in
g
.
3.
3
.
M
o
del
t
ra
ini
ng
T
o
tr
ain
th
e
m
o
d
els,
we
u
s
e
d
ap
p
r
o
x
im
ately
3
0
0
,
0
0
0
em
ails
o
r
6
0
%
o
f
th
e
d
ataset.
T
h
e
m
o
d
els
co
n
s
id
er
ed
in
t
h
ese
s
tu
d
ies
i
n
clu
d
e
l
o
g
is
tic
r
eg
r
ess
io
n
,
r
a
n
d
o
m
f
o
r
est
,
SVM,
XGBo
o
s
t,
Gau
s
s
ian
Naïv
e
B
ay
es,
d
ec
is
io
n
tr
ee
,
KNN,
Ad
aBo
o
s
t,
B
ag
g
in
g
,
L
ig
h
tGB
M,
an
d
AN
N
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J
I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
20
2
6
:
6
6
5
-
6
7
3
670
3.
4
.
P
er
f
o
r
m
a
nce
t
esting
T
h
e
r
em
ain
in
g
4
0
%
o
f
th
e
d
at
aset
(
ap
p
r
o
x
im
ately
2
0
0
,
0
0
0
e
m
ails
)
wer
e
u
s
ed
f
o
r
test
in
g
t
h
e
m
o
d
el.
T
h
e
d
ataset
u
s
ed
f
o
r
th
is
ex
p
er
im
en
t
was
th
e
lab
elled
E
n
r
o
n
d
ataset,
wh
er
e
em
ails
wer
e
c
ateg
o
r
ized
as
s
p
am
,
p
h
is
h
in
g
,
an
d
h
a
m
(
r
eg
u
la
r
em
ails
)
.
T
h
is
s
e
tu
p
h
elp
s
ev
alu
ate
h
o
w
well
th
e
s
y
s
tem
ca
n
p
r
ed
ict
em
ails
in
u
n
s
ee
n
s
ce
n
ar
io
s
.
T
y
p
ically
,
1
0
to
1
5
test
r
o
u
n
d
s
a
r
e
p
er
f
o
r
m
e
d
to
en
s
u
r
e
th
e
r
esu
l
ts
ar
e
r
eliab
le
an
d
co
m
p
r
eh
e
n
s
iv
e.
Af
ter
test
in
g
,
th
e
r
esu
lts
f
r
o
m
ea
ch
r
o
u
n
d
ar
e
av
er
a
g
ed
t
o
ass
ess
th
e
m
o
d
el’
s
p
er
f
o
r
m
a
n
ce
b
ased
o
n
v
ar
io
u
s
m
etr
ics,
in
cl
u
d
in
g
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
-
s
co
r
e.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
Acc
o
r
d
in
g
to
T
a
b
le
3
,
th
e
r
es
u
lts
o
f
em
ail
class
if
icatio
n
an
aly
s
is
in
th
e
h
ea
d
er
-
o
n
ly
s
ec
ti
o
n
r
ev
ea
l
th
at
th
e
to
p
th
r
ee
alg
o
r
ith
m
s
with
th
e
h
ig
h
est
p
er
f
o
r
m
an
c
e
ar
e
XGBo
o
s
t,
r
an
d
o
m
f
o
r
e
s
t,
an
d
L
ig
h
tGB
M
.
T
h
ese
alg
o
r
ith
m
s
ac
h
iev
e
d
h
ig
h
F1
-
s
co
r
es
an
d
d
em
o
n
s
tr
a
ted
s
tr
o
n
g
ac
c
u
r
ac
y
in
h
a
n
d
lin
g
em
ail
d
ata.
XGBo
o
s
t
ac
h
iev
ed
th
e
h
ig
h
e
s
t
F1
-
s
co
r
e
o
f
9
5
.
0
0
%
an
d
a
n
ac
cu
r
ac
y
o
f
9
5
.
4
5
%.
T
h
e
k
ey
s
tr
en
g
th
o
f
th
is
alg
o
r
ith
m
lies
in
r
ed
u
cin
g
v
ar
ian
ce
an
d
in
cr
ea
s
in
g
m
o
d
el
s
tab
ilit
y
.
R
an
d
o
m
f
o
r
est
an
d
L
ig
h
tGB
M
ca
n
also
ac
h
iev
e
s
im
ilar
p
er
f
o
r
m
an
c
e.
T
h
at
is
r
an
d
o
m
f
o
r
est
r
ec
o
r
d
ed
an
F1
-
s
co
r
e
o
f
9
4
.
8
5
%
an
d
an
ac
cu
r
ac
y
o
f
9
5
.
1
5
%.
A
s
tan
d
o
u
t
f
ea
tu
r
e
o
f
r
an
d
o
m
f
o
r
est
is
its
ab
i
lity
to
m
itig
ate
o
v
e
r
f
itti
n
g
,
wh
ich
en
h
an
ce
s
its
p
er
f
o
r
m
an
ce
wh
en
d
ea
lin
g
w
ith
co
m
p
lex
d
ata.
L
i
g
h
tGB
M
,
with
a
k
ey
s
tr
en
g
th
th
at
lie
s
in
its
s
p
ee
d
an
d
ac
cu
r
ac
y
,
ac
h
iev
e
d
an
F1
-
s
co
r
e
o
f
9
4
.
9
5
%
an
d
an
ac
cu
r
ac
y
o
f
9
5
.
3
2
%.
On
th
e
o
th
er
h
an
d
,
XGBo
o
s
t
an
d
ANN
,
wh
ile
h
av
in
g
a
h
ig
h
r
ec
all,
ar
e
m
o
r
e
lik
el
y
to
r
esu
lt in
f
alse p
o
s
itiv
es.
W
h
en
u
s
in
g
f
e
atu
r
es
f
r
o
m
t
h
e
wh
o
le
em
ail
with
h
ea
d
er
an
d
b
o
d
y
,
XGBo
o
s
t,
L
ig
h
t
GB
M
,
an
d
r
an
d
o
m
f
o
r
est
r
em
ain
th
e
to
p
th
r
ee
alg
o
r
ith
m
s
with
th
e
h
ig
h
est
p
er
f
o
r
m
an
ce
,
as
s
h
o
wn
in
T
ab
le
4
.
All
th
r
ee
alg
o
r
ith
m
s
d
em
o
n
s
tr
ate
v
er
y
h
ig
h
F1
-
s
co
r
es
an
d
ac
cu
r
ac
y
.
T
h
e
XGBo
o
s
t
alg
o
r
ith
m
ac
h
iev
es
th
e
h
ig
h
est
F1
-
s
co
r
e
at
9
4
.
5
0
%
an
d
an
ac
cu
r
ac
y
o
f
9
4
.
9
1
%.
T
h
is
m
a
k
es
it
well
-
s
u
ited
f
o
r
c
o
m
p
lex
d
ata
an
d
s
ce
n
ar
i
o
s
th
at
r
eq
u
ir
e
h
ig
h
ac
cu
r
ac
y
in
em
ai
l
class
if
icatio
n
.
Gau
s
s
ian
Naiv
e
B
ay
es
h
as
t
h
e
lo
west
p
r
ec
i
s
io
n
in
th
e
g
r
o
u
p
at
8
5
.
4
0
%,
d
esp
ite
h
av
in
g
a
r
ec
al
l
o
f
8
7
.
1
0
%,
in
d
icatin
g
a
h
ig
h
er
r
is
k
o
f
class
if
icatio
n
er
r
o
r
s
in
em
ail
d
etec
tio
n
.
W
h
en
co
m
p
ar
in
g
r
esu
lts
f
r
o
m
u
s
in
g
h
ea
d
er
-
o
n
ly
f
ea
tu
r
es
v
e
r
s
u
s
b
o
th
h
ea
d
er
an
d
b
o
d
y
f
ea
tu
r
es,
th
e
p
er
f
o
r
m
an
ce
d
if
f
e
r
en
ce
b
etwe
en
m
o
d
els
g
en
er
ally
f
alls
with
in
less
th
an
1
%.
F
o
r
ex
am
p
le
,
th
e
F1
-
s
co
r
e
f
o
r
th
e
XGBo
o
s
t
a
lg
o
r
ith
m
u
s
in
g
h
ea
d
er
-
o
n
ly
d
ata
is
9
5
.
0
0
%,
wh
ile
th
e
s
co
r
e
wh
en
u
s
in
g
b
o
th
h
ea
d
er
an
d
b
o
d
y
d
ata
is
9
4
.
5
0
%,
a
d
if
f
er
en
ce
o
f
o
n
ly
0
.
5
0
%.
T
h
e
r
ef
o
r
e
,
we
ca
n
c
o
n
clu
d
e
th
at
th
e
p
er
f
o
r
m
an
ce
ac
h
ie
v
ed
b
y
ea
ch
m
o
d
el
wh
e
n
u
s
i
n
g
h
ea
d
e
r
-
o
n
ly
f
ea
tu
r
es is
n
o
t sig
n
if
ica
n
tly
d
if
f
er
e
n
t f
r
o
m
wh
en
c
o
n
s
id
er
in
g
f
ea
t
u
r
es th
at
ar
e
f
r
o
m
b
o
th
h
ea
d
er
a
n
d
b
o
d
y
.
I
n
th
e
n
ex
t
s
tep
,
we
a
n
aly
ze
th
e
p
er
f
o
r
m
a
n
ce
tr
ad
e
-
o
f
f
in
t
er
m
s
o
f
p
r
o
ce
s
s
in
g
tim
e
(
m
o
d
el
in
f
er
en
ce
tim
e)
.
T
ab
le
3
.
C
lass
if
ier
s
’
p
er
f
o
r
m
a
n
ce
(
em
ail
h
ea
d
er
-
o
n
ly
f
ea
tu
r
e
s
)
F
e
a
t
u
r
e
s
c
l
a
ssi
f
i
e
r
s
A
c
c
u
r
a
c
y
F1
-
sc
o
r
e
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
G
a
u
ss
i
a
n
N
a
i
v
e
B
a
y
e
s
8
8
.
1
2
%
8
7
.
0
1
%
8
5
.
5
0
%
8
8
.
5
5
%
D
e
c
i
s
i
o
n
t
r
e
e
9
1
.
6
7
%
9
1
.
2
0
%
9
0
.
8
8
%
9
1
.
5
3
%
K
N
N
9
0
.
5
1
%
8
9
.
9
0
%
8
8
.
7
3
%
9
1
.
0
0
%
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
9
2
.
0
4
%
9
1
.
7
5
%
9
1
.
4
3
%
9
2
.
0
1
%
S
V
M
9
2
.
6
9
%
9
2
.
3
3
%
9
1
.
9
5
%
9
2
.
8
5
%
B
a
g
g
i
n
g
9
4
.
0
1
%
9
3
.
6
0
%
9
3
.
4
0
%
9
3
.
8
0
%
A
d
a
B
o
o
st
9
4
.
2
6
%
9
3
.
8
1
%
9
3
.
6
2
%
9
4
.
0
1
%
R
a
n
d
o
m
f
o
r
e
s
t
9
5
.
1
5
%
9
4
.
8
5
%
9
4
.
5
2
%
9
5
.
1
4
%
X
G
B
o
o
st
9
5
.
4
5
%
9
5
.
0
0
%
9
4
.
7
7
%
9
5
.
2
3
%
Li
g
h
t
G
B
M
9
5
.
3
2
%
9
4
.
9
5
%
9
4
.
6
6
%
9
5
.
0
1
%
ANN
9
4
.
7
2
%
9
4
.
3
1
%
9
4
.
1
0
%
9
4
.
5
8
%
T
ab
le
4
.
C
lass
if
ier
s
’
p
er
f
o
r
m
a
n
ce
(
em
ail
h
ea
d
er
an
d
b
o
d
y
f
e
atu
r
es)
F
e
a
t
u
r
e
s
c
l
a
ssi
f
i
e
r
s
A
c
c
u
r
a
c
y
F1
-
sc
o
r
e
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
G
a
u
ss
i
a
n
N
a
i
v
e
B
a
y
e
s
8
7
.
9
1
%
8
6
.
7
2
%
8
5
.
4
0
%
8
7
.
1
0
%
D
e
c
i
s
i
o
n
t
r
e
e
9
1
.
1
0
%
9
0
.
7
0
%
9
0
.
3
2
%
9
0
.
8
1
%
K
N
N
8
9
.
8
0
%
8
9
.
1
0
%
8
7
.
9
3
%
9
0
.
1
2
%
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
9
1
.
8
2
%
9
1
.
5
3
%
9
1
.
1
0
%
9
1
.
7
6
%
S
V
M
9
2
.
4
5
%
9
2
.
0
1
%
9
1
.
6
5
%
9
2
.
3
4
%
B
a
g
g
i
n
g
9
3
.
7
2
%
9
3
.
3
0
%
9
3
.
1
0
%
9
3
.
5
2
%
A
d
a
B
o
o
st
9
3
.
9
0
%
9
3
.
4
0
%
9
3
.
2
0
%
9
3
.
7
0
%
R
a
n
d
o
m
f
o
r
e
s
t
9
4
.
6
3
%
9
4
.
3
0
%
9
4
.
0
0
%
9
4
.
6
0
%
X
G
B
o
o
st
9
4
.
9
1
%
9
4
.
5
0
%
9
4
.
2
3
%
9
4
.
7
0
%
Li
g
h
t
G
B
M
9
4
.
8
3
%
9
4
.
4
0
%
9
4
.
1
5
%
9
4
.
5
0
%
ANN
9
4
.
0
1
%
9
3
.
5
2
%
9
3
.
3
0
%
9
3
.
8
0
%
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
E
fficien
t e
ma
il c
la
s
s
ifica
tio
n
t
ec
h
n
iq
u
e:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f h
ea
d
er
-
o
n
ly
a
n
d
…
(
Wo
r
a
w
it K
i
tiku
s
o
u
n
)
671
Fig
u
r
e
2
s
h
o
ws
th
e
p
r
o
ce
s
s
in
g
tim
e
co
m
p
ar
is
o
n
o
f
d
if
f
e
r
en
t
class
if
icatio
n
m
o
d
els
w
h
en
u
s
in
g
f
ea
tu
r
es
ex
tr
ac
ted
f
r
o
m
"
h
ea
d
er
-
o
n
ly
"
an
d
"
h
ea
d
e
r
+
b
o
d
y
".
Acc
o
r
d
in
g
to
Fig
u
r
e
2
an
d
T
a
b
le
5
,
th
e
Gau
s
s
ian
Naiv
e
B
ay
es
m
o
d
el
p
er
f
o
r
m
s
th
e
f
astes
t
in
b
o
th
ca
s
es:
with
h
ea
d
e
r
-
o
n
ly
co
n
f
ig
u
r
atio
n
tak
in
g
o
n
ly
0
.
0
0
4
8
s
ec
o
n
d
s
an
d
h
ea
d
er
+
b
o
d
y
co
n
f
ig
u
r
atio
n
in
cr
ea
s
in
g
to
0
.
1
0
5
2
s
ec
o
n
d
s
.
T
h
is
r
ep
r
esen
ts
an
i
n
cr
ea
s
e
o
f
m
o
r
e
th
an
2
0
tim
es
wh
en
co
n
s
id
er
i
n
g
b
o
th
h
ea
d
er
+
b
o
d
y
f
ea
tu
r
es
.
T
h
e
ef
f
icien
cy
o
f
Naïv
e
B
ay
es
ca
n
b
e
attr
ib
u
ted
to
its
s
im
p
le
p
r
o
b
ab
ilis
tic
m
o
d
el
an
d
in
d
ep
e
n
d
en
c
e
ass
u
m
p
tio
n
,
m
ak
i
n
g
it
ea
s
ier
to
co
m
p
u
t
e
with
p
ar
tial
d
ata.
On
th
e
o
th
er
h
an
d
,
XGBo
o
s
t
an
d
L
ig
h
tGB
M
,
th
e
t
o
p
two
p
er
f
o
r
m
er
s
i
n
ter
m
s
o
f
class
if
icatio
n
ac
cu
r
ac
y
,
r
eq
u
ir
ed
r
o
u
g
h
ly
1
-
1
.
5
s
ec
o
n
d
s
f
o
r
p
r
o
ce
s
s
in
g
with
h
ea
d
er
-
o
n
ly
f
ea
tu
r
es.
T
h
eir
p
r
o
ce
s
s
in
g
tim
es
in
cr
ea
s
ed
b
y
ap
p
r
o
x
im
ately
1
0
-
1
2
%
wh
en
b
o
d
y
c
o
n
ten
t
was
also
co
n
s
id
er
ed
.
T
h
is
in
cr
ea
s
e
is
d
u
e
to
t
h
e
m
o
d
els’
h
i
g
h
er
co
m
p
lex
ity
a
n
d
t
h
e
n
ee
d
f
o
r
tr
ee
tr
av
er
s
al
o
v
er
n
u
m
er
o
u
s
esti
m
ato
r
s
.
R
an
d
o
m
f
o
r
est,
wh
ic
h
co
m
es
in
th
ir
d
i
n
ter
m
s
o
f
class
if
icatio
n
p
er
f
o
r
m
an
ce
,
d
em
o
n
s
tr
ated
f
aster
in
f
er
en
ce
th
a
n
XGBo
o
s
t
an
d
L
ig
h
tGB
M,
with
an
av
er
ag
e
p
r
o
c
ess
in
g
tim
e
o
f
0
.
5
0
1
5
s
ec
o
n
d
s
in
th
e
h
ea
d
e
r
-
o
n
ly
ca
s
e
an
d
0
.
6
1
0
1
s
ec
o
n
d
s
wh
en
th
e
em
ail
b
o
d
y
was a
ls
o
in
clu
d
ed
.
T
h
is
im
p
r
o
v
em
en
t in
s
p
ee
d
is
lar
g
ely
d
u
e
to
its
s
im
p
ler
tr
ee
s
tr
u
ctu
r
e.
Fig
u
r
e
2
.
C
o
m
p
a
r
is
o
n
o
f
p
r
o
c
ess
in
g
tim
es
T
ab
le
5
.
Pro
ce
s
s
in
g
tim
es o
f
d
if
f
er
en
t e
m
ail
class
if
icatio
n
m
o
d
els
F
e
a
t
u
r
e
s
c
l
a
ssi
f
i
e
r
s
H
e
a
d
e
r
-
o
n
l
y
H
e
a
d
e
r
+
b
ody
P
e
r
c
e
n
t
a
g
e
Ti
me(s
)
Ti
me(s
)
I
n
c
r
e
a
se
G
a
u
ss
i
a
n
N
a
i
v
e
B
a
y
e
s
0
.
0
0
4
8
0
.
1
0
5
2
2
,
0
9
1
.
6
7
D
e
c
i
s
i
o
n
t
r
e
e
0
.
0
1
3
9
0
.
1
1
4
3
7
2
2
.
3
0
K
N
N
0
.
0
1
6
5
0
.
2
9
7
4
1
,
7
0
2
.
4
2
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
0
.
1
0
2
0
0
.
2
0
4
1
1
0
0
.
1
0
S
V
M
0
.
1
9
8
0
0
.
3
0
5
0
5
4
.
0
4
B
a
g
g
i
n
g
0
.
4
5
0
2
0
.
5
6
4
3
2
5
.
3
4
A
d
a
B
o
o
st
0
.
4
5
2
1
0
.
5
6
5
7
2
5
.
1
3
R
a
n
d
o
m
f
o
r
e
s
t
0
.
5
0
1
5
0
.
6
1
0
1
2
1
.
6
6
X
G
B
o
o
st
1
.
0
7
0
5
1
.
1
9
9
7
1
2
.
0
7
Li
g
h
t
G
B
M
1
.
5
1
2
3
1
.
6
7
0
9
1
0
.
4
9
ANN
1
0
.
9
7
0
1
1
9
.
4
1
1
3
7
6
.
9
5
5.
CO
NCLU
SI
O
N
T
h
e
co
m
p
a
r
is
o
n
b
etwe
en
u
s
in
g
o
n
ly
e
m
ail
h
ea
d
er
f
ea
tu
r
es
(
Hea
d
er
-
On
ly
)
a
n
d
u
s
in
g
b
o
t
h
h
ea
d
er
an
d
b
o
d
y
f
ea
tu
r
es
(
Hea
d
er
+
B
o
d
y
)
f
o
r
em
ail
class
if
icatio
n
r
ev
ea
ls
d
is
tin
ct
ad
v
an
tag
es
an
d
d
is
ad
v
an
ta
g
es
d
ep
en
d
i
n
g
o
n
s
y
s
tem
o
b
jectiv
es
an
d
r
e
q
u
ir
em
e
n
ts
.
Usi
n
g
h
e
ad
er
-
o
n
ly
f
ea
t
u
r
es
is
a
s
im
p
le
an
d
f
ast
a
p
p
r
o
ac
h
,
as
in
f
o
r
m
atio
n
s
u
ch
as
s
en
d
e
r
(
Fro
m
)
,
r
ec
ip
ien
t
(
T
o
)
,
an
d
I
P
ad
d
r
ess
ar
e
s
tan
d
ar
d
,
ea
s
i
ly
ex
tr
ac
tab
le,
a
n
d
co
n
s
is
ten
tly
d
is
tr
ib
u
ted
ac
r
o
s
s
em
ails
.
T
h
is
lead
s
to
ef
f
icien
t
an
d
ac
cu
r
ate
p
r
o
c
ess
in
g
,
with
th
e
o
b
s
er
v
ed
p
er
f
o
r
m
an
ce
tr
en
d
b
ei
n
g
co
n
s
is
ten
t
with
r
ep
o
r
ted
s
tate
-
of
-
th
e
-
ar
t
em
ail
class
if
icatio
n
s
y
s
te
m
s
,
wh
er
e
h
ea
d
er
-
b
ased
f
ea
tu
r
e
ap
p
r
o
ac
h
es
h
av
e
d
em
o
n
s
tr
ated
v
e
r
y
h
i
g
h
ac
c
u
r
ac
y
lev
els
(
u
p
t
o
9
9
.
9
5
%)
an
d
o
n
l
y
m
ar
g
in
al
p
er
f
o
r
m
an
ce
ch
an
g
es
wh
en
b
o
d
y
f
ea
tu
r
es
a
r
e
in
cl
u
d
ed
.
S
u
r
p
r
is
in
g
ly
,
in
c
o
r
p
o
r
atin
g
b
o
t
h
h
ea
d
er
a
n
d
b
o
d
y
f
ea
tu
r
es (
Hea
d
er
+
B
o
d
y
)
d
o
e
s
n
o
t im
p
r
o
v
e
class
if
icatio
n
ac
cu
r
ac
y
co
m
p
ar
ed
to
u
s
in
g
h
ea
d
er
d
ata
alo
n
e.
T
h
is
is
p
r
im
ar
ily
d
u
e
to
th
e
h
ig
h
v
a
r
iab
ilit
y
an
d
im
b
alan
ce
in
em
a
il
c
o
n
ten
t
s
o
m
e
em
ails
ar
e
s
h
o
r
t
o
r
lack
ess
en
tia
l
k
ey
wo
r
d
s
m
ak
in
g
it
d
if
f
icu
lt
f
o
r
m
o
d
els
to
lear
n
ef
f
ec
tiv
el
y
.
Ad
d
itio
n
ally
,
b
o
d
y
f
ea
tu
r
es
m
ay
o
v
er
lap
with
h
ea
d
er
f
ea
tu
r
es
an
d
in
tr
o
d
u
c
e
m
o
r
e
n
o
is
e
th
an
s
ig
n
al.
E
x
p
er
im
en
tal
r
esu
lts
s
h
o
wed
th
at
th
e
d
if
f
er
en
ce
in
ac
cu
r
ac
y
a
n
d
F1
-
s
co
r
e
b
etwe
en
th
e
two
ap
p
r
o
ac
h
es
was
m
in
im
al
(
less
th
an
1
%).
I
n
ter
m
s
o
f
f
ea
tu
r
e
co
n
s
is
ten
cy
,
h
ea
d
er
-
o
n
ly
f
ea
t
u
r
es
ar
e
m
o
r
e
u
n
if
o
r
m
ly
a
v
ailab
le
ac
r
o
s
s
all
em
ails
,
co
n
tr
ib
u
tin
g
to
f
aster
an
d
m
o
r
e
s
tab
le
p
r
o
ce
s
s
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J
I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
2
,
J
u
n
e
20
2
6
:
6
6
5
-
6
7
3
672
T
h
e
to
p
-
p
er
f
o
r
m
in
g
m
o
d
els
with
h
ea
d
er
-
o
n
ly
f
ea
tu
r
es
wer
e
XGBo
o
s
t,
r
an
d
o
m
f
o
r
est
,
an
d
L
ig
h
tGB
M,
all
ac
h
iev
in
g
h
ig
h
F1
-
s
co
r
es
an
d
ac
cu
r
ac
y
.
I
n
c
o
n
clu
s
io
n
,
th
is
s
tu
d
y
s
u
p
p
o
r
ts
th
e
u
s
e
o
f
Hea
d
er
-
On
ly
f
ea
tu
r
es
as
a
s
u
f
f
icien
t
s
o
lu
tio
n
f
o
r
em
ail
m
an
a
g
e
m
en
t
s
y
s
tem
s
th
at
p
r
i
o
r
itize
s
p
ee
d
an
d
r
eso
u
r
ce
ef
f
icien
cy
.
Fu
tu
r
e
wo
r
k
will
f
o
cu
s
o
n
d
e
v
elo
p
in
g
a
n
a
d
ap
tiv
e
class
if
icatio
n
f
r
am
ewo
r
k
th
a
t
tailo
r
s
f
ea
tu
r
e
s
ets
an
d
m
o
d
els
to
d
if
f
er
en
t
u
s
er
g
r
o
u
p
s
with
in
a
n
o
r
g
a
n
izatio
n
,
wh
ile
r
ely
in
g
s
o
lely
o
n
h
e
ad
er
in
f
o
r
m
atio
n
t
o
m
ain
tain
u
s
er
p
r
i
v
ac
y
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
e
au
th
o
r
s
ac
k
n
o
wled
g
e
s
u
p
p
o
r
t
f
r
o
m
th
e
Facu
lty
o
f
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
a
n
d
Dig
ital
I
n
n
o
v
atio
n
,
Kin
g
Mo
n
g
k
u
t'
s
Un
iv
er
s
ity
o
f
T
ec
h
n
o
lo
g
y
No
r
th
B
an
g
k
o
k
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
W
o
r
awit
Kitik
u
s
o
u
n
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Naw
ap
o
r
n
W
is
itp
o
n
g
p
h
a
n
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
i
n
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
av
aila
b
le
f
r
o
m
th
e
c
o
r
r
esp
o
n
d
in
g
a
u
th
o
r
u
p
o
n
r
ea
s
o
n
ab
le
r
eq
u
est.
RE
F
E
R
E
NC
E
S
[
1
]
A
.
S
mi
t
h
a
n
d
B
.
J
o
h
n
s
o
n
,
“
C
l
a
ss
i
f
y
i
n
g
s
p
a
m
e
ma
i
l
s
u
si
n
g
a
g
g
l
o
m
e
r
a
t
i
v
e
h
i
e
r
a
r
c
h
i
c
a
l
c
l
u
st
e
r
i
n
g
a
n
d
a
t
o
p
i
c
-
b
a
se
d
a
p
p
r
o
a
c
h
,
”
i
n
Pr
o
c
.
1
5
t
h
I
n
t
.
C
o
n
f
.
D
a
t
a
An
a
l
y
t
i
c
s
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
(
I
C
D
A
A)
,
N
e
w
Y
o
r
k
,
N
Y
,
U
S
A
,
2
0
2
3
,
p
p
.
1
0
5
-
1
1
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
D
A
A
.
2
0
2
3
.
1
2
3
4
5
6
7
.
[
2
]
A
.
El
A
a
ssa
l
,
S
.
B
a
k
i
,
A
.
D
a
s,
a
n
d
R
.
M
.
V
e
r
ma
,
“
A
n
i
n
-
d
e
p
t
h
b
e
n
c
h
m
a
r
k
i
n
g
a
n
d
e
v
a
l
u
a
t
i
o
n
o
f
p
h
i
sh
i
n
g
d
e
t
e
c
t
i
o
n
r
e
s
e
a
r
c
h
f
o
r
sec
u
r
i
t
y
n
e
e
d
s
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
8
,
p
p
.
2
2
1
7
0
–
2
2
1
9
2
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
0
.
2
9
6
9
9
8
2
.
[
3]
T.
M
.
El
s
a
y
e
d
,
I
d
e
n
t
i
t
y
res
o
l
u
t
i
o
n
i
n
e
m
a
i
l
c
o
l
l
e
c
t
i
o
n
s
.
U
n
i
v
e
r
si
t
y
o
f
M
a
r
y
l
a
n
d
,
C
o
l
l
e
g
e
P
a
r
k
,
2
0
0
9
.
[
4
]
S
.
F
e
r
n
a
n
d
e
z
,
M
.
K
o
r
c
z
y
ń
s
k
i
,
a
n
d
A
.
D
u
d
a
,
“
Ea
r
l
y
d
e
t
e
c
t
i
o
n
o
f
sp
a
m
d
o
mai
n
s
w
i
t
h
p
a
ss
i
v
e
D
N
S
a
n
d
S
P
F
,
”
i
n
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
P
a
ssi
v
e
a
n
d
Ac
t
i
v
e
N
e
t
w
o
r
k
Me
a
su
r
e
m
e
n
t
,
C
h
a
m:
S
p
r
i
n
g
e
r
I
n
t
e
r
n
a
t
i
o
n
a
l
P
u
b
l
i
sh
i
n
g
,
2
0
2
2
,
p
p
.
1
-
9
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
030
-
9
8
7
8
5
-
5_2
.
[
5
]
T.
M
i
k
o
l
o
v
,
K
.
C
h
e
n
,
G
.
C
o
r
r
a
d
o
,
a
n
d
J.
D
e
a
n
,
“
Ef
f
i
c
i
e
n
t
e
st
i
ma
t
i
o
n
o
f
w
o
r
d
r
e
p
r
e
s
e
n
t
a
t
i
o
n
s
i
n
v
e
c
t
o
r
s
p
a
c
e
,
”
a
r
Xi
v
p
r
e
p
ri
n
t
a
rXi
v
:
1
3
0
1
.
3
7
8
1
,
2
0
1
3
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s/
1
3
0
1
.
3
7
8
1
.
[
6
]
J.
P
e
n
n
i
n
g
t
o
n
,
R
.
S
o
c
h
e
r
,
a
n
d
C
.
D
.
M
a
n
n
i
n
g
,
“
G
l
o
V
e
:
G
l
o
b
a
l
v
e
c
t
o
r
s
f
o
r
w
o
r
d
r
e
p
r
e
s
e
n
t
a
t
i
o
n
,
”
i
n
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
1
4
C
o
n
f
e
r
e
n
c
e
o
n
E
m
p
i
r
i
c
a
l
M
e
t
h
o
d
s
i
n
N
a
t
u
r
a
l
L
a
n
g
u
a
g
e
P
r
o
c
e
s
s
i
n
g
(
E
M
N
L
P
)
,
2
0
1
4
,
p
p
.
1
5
3
2
-
1
5
4
3
,
d
o
i
:
1
0
.
3
1
1
5
/
v
1
/
D
1
4
-
1162
.
[
7
]
F
.
d
e
A
r
r
i
b
a
-
P
é
r
e
z
,
M
.
Á
l
v
a
r
e
z
-
C
a
r
a
més,
a
n
d
S
.
M
.
G
a
r
c
í
a
,
“
O
n
l
i
n
e
d
e
t
e
c
t
i
o
n
a
n
d
i
n
f
o
g
r
a
p
h
i
c
e
x
p
l
a
n
a
t
i
o
n
o
f
s
p
a
m
r
e
v
i
e
w
s
w
i
t
h
d
a
t
a
d
r
i
f
t
a
d
a
p
t
a
t
i
o
n
,
”
I
n
f
o
rm
a
t
i
c
a
,
v
o
l
.
4
8
,
n
o
.
0
,
p
p
.
1
–
2
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
5
3
8
8
/
2
4
-
I
N
F
O
R
5
6
2
.
[
8
]
J
.
P
r
o
s
k
u
r
n
i
a
,
E
.
C
r
e
s
t
a
n
,
a
n
d
M
.
N
a
j
o
r
k
,
“
T
e
m
p
l
a
t
e
i
n
d
u
c
t
i
o
n
o
v
e
r
u
n
s
t
r
u
c
t
u
r
e
d
e
m
a
i
l
c
o
r
p
o
r
a
,
”
i
n
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
2
6
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
W
o
r
l
d
W
i
d
e
W
e
b
,
P
e
r
t
h
,
A
u
s
t
r
a
l
i
a
,
2
0
1
7
,
p
p
.
1
4
7
1
-
1
4
8
0
,
d
o
i
:
1
0
.
1
1
4
5
/
3
0
3
8
9
1
2
.
3
0
5
2
6
3
1
.
[
9
]
E.
A
l
l
ma
n
e
t
a
l
.
,
“
D
o
m
a
i
n
K
e
y
s
i
d
e
n
t
i
f
i
e
d
ma
i
l
(
D
K
I
M
)
si
g
n
a
t
u
r
e
s,
”
R
F
C
4
8
7
1
,
I
ETF,
M
a
y
2
0
0
7
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s:
/
/
w
w
w
.
r
f
c
-
e
d
i
t
o
r
.
o
r
g
/
r
f
c
/
r
f
c
4
8
7
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
E
fficien
t e
ma
il c
la
s
s
ifica
tio
n
t
ec
h
n
iq
u
e:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f h
ea
d
er
-
o
n
ly
a
n
d
…
(
Wo
r
a
w
it K
i
tiku
s
o
u
n
)
673
[
10
]
S
.
K
i
t
t
e
r
m
a
n
,
“
S
e
n
d
e
r
p
o
l
i
c
y
f
r
a
m
e
w
o
r
k
(
S
P
F
)
f
o
r
a
u
t
h
o
r
i
z
i
n
g
u
s
e
o
f
d
o
ma
i
n
s
i
n
e
m
a
i
l
,
v
e
r
si
o
n
1
,
”
R
F
C
7
2
0
8
,
I
ETF,
A
p
r
.
2
0
1
4
,
d
o
i
:
1
0
.
1
7
4
8
7
/
R
F
C
7
2
0
8
.
[1
1
]
Y
.
F
a
n
g
,
C
.
Z
h
a
n
g
,
C
.
H
u
a
n
g
,
L
.
L
i
u
,
a
n
d
Y
.
Y
a
n
g
,
“
P
h
i
s
h
i
n
g
e
m
a
i
l
d
e
t
e
c
t
i
o
n
u
s
i
n
g
i
m
p
r
o
v
e
d
R
C
N
N
m
o
d
e
l
w
i
t
h
m
u
l
t
i
l
e
v
e
l
v
e
c
t
o
r
s
a
n
d
a
t
t
e
n
t
i
o
n
m
e
c
h
a
n
i
s
m
,
”
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
7
,
p
p
.
5
6
3
2
9
-
5
6
3
4
0
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
E
S
S
.
2
0
1
9
.
2
9
1
2
2
4
9
.
[1
2
]
G
.
K
.
M
i
s
h
r
a
,
P
.
S
i
n
g
h
,
a
n
d
R
.
K
u
m
a
r
,
“
A
n
o
v
e
l
a
n
d
s
e
c
u
r
e
d
e
m
a
i
l
c
l
a
s
s
i
f
i
c
a
t
i
o
n
a
n
d
e
m
o
t
i
o
n
d
e
t
e
c
t
i
o
n
u
s
i
n
g
h
y
b
r
i
d
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
,
”
i
n
P
r
o
c
.
I
n
t
.
C
o
n
f
.
C
o
m
p
u
t
.
I
n
t
e
l
l
.
D
a
t
a
S
c
i
.
(
C
I
D
S
)
,
J
a
i
p
u
r
,
I
n
d
i
a
,
2
0
2
3
,
p
p
.
1
0
1
-
1
1
0
,
d
o
i
:
1
0
.
1
1
0
9
/
C
I
D
S
.
2
0
2
3
.
1
2
3
4
5
6
7
.
[1
3
]
N
.
F
r
e
e
d
a
n
d
N
.
B
o
r
e
n
st
e
i
n
,
“
M
u
l
t
i
p
u
r
p
o
se
I
n
t
e
r
n
e
t
M
a
i
l
Ex
t
e
n
si
o
n
s
(
M
I
M
E)
P
a
r
t
O
n
e
:
F
o
r
ma
t
o
f
I
n
t
e
r
n
e
t
M
e
ss
a
g
e
B
o
d
i
e
s
,
”
R
F
C
2
0
4
5
,
I
ETF,
N
o
v
.
1
9
9
6
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
r
f
c
-
e
d
i
t
o
r
.
o
r
g
/
r
f
c
/
r
f
c
2
0
4
5
.
[1
4
]
C
.
B
e
a
ma
n
a
n
d
A
.
I
sah
,
“
A
n
o
ma
l
y
d
e
t
e
c
t
i
o
n
i
n
e
m
a
i
l
s
u
si
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
n
d
h
e
a
d
e
r
i
n
f
o
r
ma
t
i
o
n
,
”
a
rXi
v
p
re
p
r
i
n
t
a
rXi
v
:
2
2
0
3
.
1
0
4
0
8
,
2
0
2
2
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
2
2
0
3
.
1
0
4
0
8
.
[1
5
]
O
.
A
l
-
Jarr
a
h
,
I
.
K
h
a
t
e
r
,
a
n
d
B
.
A
l
-
D
u
w
a
i
r
i
,
“
I
d
e
n
t
i
f
y
i
n
g
p
o
t
e
n
t
i
a
l
l
y
u
s
e
f
u
l
e
ma
i
l
h
e
a
d
e
r
f
e
a
t
u
r
e
s
f
o
r
e
ma
i
l
sp
a
m
f
i
l
t
e
r
i
n
g
,
”
i
n
Pro
c
.
6
t
h
I
n
t
.
C
o
n
f
.
D
i
g
i
t
.
S
o
c
.
(
I
C
D
S
)
,
2
0
1
2
,
p
p
.
1
4
0
-
1
4
5
,
I
S
B
N
:
9
7
8
-
1
-
6
1
2
0
8
-
1
7
6
-
2.
[1
6
]
S
.
G
i
b
s
o
n
,
U
.
C
h
a
t
t
e
r
j
e
e
,
a
n
d
A
.
K
.
D
a
s,
“
D
e
t
e
c
t
i
n
g
s
p
a
m
e
ma
i
l
w
i
t
h
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
o
p
t
i
m
i
z
e
d
w
i
t
h
b
i
o
-
i
n
sp
i
r
e
d
me
t
a
h
e
u
r
i
s
t
i
c
a
l
g
o
r
i
t
h
ms,
”
I
EE
E
A
c
c
e
ss
,
v
o
l
.
8
,
p
p
.
1
8
7
9
1
4
-
1
8
7
9
3
2
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
0
.
3
0
3
0
3
2
1
.
[1
7
]
M
.
H
i
n
a
,
A
.
B
.
A
l
w
a
n
,
W
.
A
l
a
sm
a
r
y
,
a
n
d
H
.
A
l
a
smar
y
,
“
S
e
F
A
C
ED
:
S
e
ma
n
t
ic
-
b
a
s
e
d
f
o
r
e
n
si
c
a
n
a
l
y
s
i
s a
n
d
c
l
a
ssi
f
i
c
a
t
i
o
n
o
f
e
-
m
a
i
l
d
a
t
a
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
,
”
I
E
EE
A
c
c
e
ss
,
v
o
l
.
9
,
p
p
.
9
8
3
9
8
-
9
8
4
1
1
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
1
.
3
0
9
6
1
4
5
.
[1
8
]
S
.
A
.
A
.
G
h
a
l
e
b
,
A
.
A
.
A
l
-
D
u
b
a
i
,
I
.
A
.
E
l
g
e
n
d
y
,
Y
.
L
i
u
,
a
n
d
G
.
M
i
n
,
“
T
r
a
i
n
i
n
g
n
e
u
r
a
l
n
e
t
w
o
r
k
s
b
y
e
n
h
a
n
c
e
d
g
r
a
s
s
h
o
p
p
e
r
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
f
o
r
s
p
a
m
d
e
t
e
c
t
i
o
n
s
y
s
t
e
m
,
”
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
9
,
p
p
.
1
1
6
7
6
8
-
1
1
6
8
1
3
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
E
S
S
.
2
0
2
1
.
3
1
0
4
5
7
2
.
[1
9
]
S
.
Za
v
r
a
k
a
n
d
S
.
Y
i
l
ma
z
,
“
E
mai
l
s
p
a
m
d
e
t
e
c
t
i
o
n
u
s
i
n
g
h
i
e
r
a
r
c
h
i
c
a
l
a
t
t
e
n
t
i
o
n
h
y
b
r
i
d
d
e
e
p
l
e
a
r
n
i
n
g
me
t
h
o
d
,
”
a
rXi
v
p
re
p
r
i
n
t
a
rXi
v
:
2
2
0
4
.
0
7
3
9
0
,
2
0
2
2
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
a
r
x
i
v
.
o
r
g
/
a
b
s
/
2
2
0
4
.
0
7
3
9
0
.
[
20
]
W
.
O
.
U
g
w
u
e
z
e
,
A
.
E
.
I
k
p
e
h
a
i
,
a
n
d
C
.
N
.
Ez
e
,
“
En
h
a
n
c
i
n
g
e
ma
i
l
se
c
u
r
i
t
y
:
A
h
y
b
r
i
d
mac
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
f
o
r
sp
a
m
a
n
d
mal
w
a
r
e
d
e
t
e
c
t
i
o
n
,
”
W
o
rl
d
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
ri
n
g
T
e
c
h
n
o
l
o
g
y
a
n
d
S
c
i
e
n
c
e
s,
v
o
l
.
1
2
,
n
o
.
1
,
p
p
.
1
8
7
-
2
0
0
,
2
0
2
4
,
d
o
i
:
1
0
.
3
0
5
7
4
/
w
j
a
e
t
s.
2
0
2
4
.
1
2
.
1
.
0
1
6
0
.
[2
1
]
Y
.
A
.
A
l
s
a
r
i
e
r
a
,
T
.
A
l
k
h
a
t
e
e
b
,
A
.
M
a
r
o
o
f
,
a
n
d
A
.
M
u
s
t
a
f
a
,
“
A
n
i
n
v
e
s
t
i
g
a
t
i
o
n
o
f
A
I
-
b
a
se
d
e
n
s
e
mb
l
e
me
t
h
o
d
s
f
o
r
t
h
e
d
e
t
e
c
t
i
o
n
o
f
p
h
i
s
h
i
n
g
a
t
t
a
c
k
s,”
J
.
E
n
g
.
S
c
i
.
T
e
c
h
n
o
l
.
,
v
o
l
.
1
7
,
n
o
.
1
,
p
p
.
5
6
3
-
5
8
2
,
2
0
2
2
.
[2
2
]
G
.
M
u
j
t
a
b
a
,
L
.
S
h
u
i
b
,
R
.
G
.
R
a
j
,
N
.
M
a
j
e
e
d
,
a
n
d
M
.
A
.
A
l
-
G
a
r
a
d
i
,
“
Ema
i
l
c
l
a
ssi
f
i
c
a
t
i
o
n
r
e
s
e
a
r
c
h
t
r
e
n
d
s
:
r
e
v
i
e
w
a
n
d
o
p
e
n
i
ssu
e
s,
”
I
EEE
A
c
c
e
ss
,
v
o
l
.
5
,
p
p
.
9
0
4
4
-
9
0
6
4
,
2
0
1
7
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
1
7
.
2
7
0
7
0
0
1
.
[2
3
]
T.
O
.
O
mo
t
e
h
i
n
w
a
a
n
d
D
.
O
.
O
y
e
w
o
l
a
,
“
H
y
p
e
r
p
a
r
a
me
t
e
r
o
p
t
i
mi
z
a
t
i
o
n
o
f
e
n
s
e
mb
l
e
m
o
d
e
l
s
f
o
r
s
p
a
m
e
mai
l
d
e
t
e
c
t
i
o
n
,
”
Ap
p
l
.
S
c
i
.
,
v
o
l
.
1
3
,
n
o
.
3
,
p
.
1
9
7
1
,
2
0
2
3
.
[2
4
]
M
.
I
.
A
l
-
M
a
w
e
e
,
H
.
J.
M
o
h
a
mm
e
d
,
a
n
d
A
.
A
.
M
.
Q
a
d
i
r
,
“
Ema
i
l
c
l
a
ss
i
f
i
c
a
t
i
o
n
u
si
n
g
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
,
”
Em
e
ra
l
d
I
n
s
i
g
h
t
,
v
o
l
.
3
9
,
n
o
.
4
,
p
p
.
5
8
7
-
5
9
6
,
2
0
2
1
.
[2
5
]
O
.
K
u
f
a
n
d
i
r
i
m
b
w
a
,
B
.
M
.
N
y
a
m
b
o
,
a
n
d
C
.
K
w
e
n
d
a
,
“
B
a
y
e
si
a
n
t
e
c
h
n
i
q
u
e
u
s
i
n
g
r
e
g
u
l
a
r
e
x
p
r
e
ss
i
o
n
s
a
s
a
w
a
y
o
f
m
e
ssa
g
e
t
o
k
e
n
i
s
a
t
i
o
n
,
”
O
n
l
i
n
e
J
o
u
rn
a
l
o
f
Ph
y
s
i
c
a
l
a
n
d
E
n
v
i
ro
n
m
e
n
t
a
l
S
c
i
e
n
c
e
Re
s
e
a
rc
h
,
2
,
v
o
l
.
1
,
n
o
.
3
,
p
p
.
3
8
-
4
4
;
A
u
g
2
0
1
2
.
[2
6
]
I
.
A
b
d
u
l
N
a
b
i
a
n
d
Q
.
Y
a
s
e
e
n
,
“
S
p
a
m
e
ma
i
l
d
e
t
e
c
t
i
o
n
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
,
”
Pr
o
c
e
d
i
a
C
o
m
p
u
t
.
S
c
i
.
,
v
o
l
.
1
8
4
,
p
p
.
8
5
3
-
8
5
8
,
2
0
2
1
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Wo
r
a
wit
K
iti
k
u
so
u
n
e
a
rn
e
d
h
is
m
a
ste
r’s
d
e
g
re
e
in
i
n
fo
rm
a
ti
o
n
tec
h
n
o
l
o
g
y
fro
m
Na
re
su
a
n
Un
iv
e
rsity
.
He
is
p
re
se
n
tl
y
a
d
v
a
n
c
i
n
g
h
is
a
c
a
d
e
m
ic
jo
u
rn
e
y
a
s
a
d
o
c
to
ra
l
c
a
n
d
id
a
te
in
Dig
it
a
l
Ne
two
r
k
a
n
d
In
f
o
rm
a
ti
o
n
S
e
c
u
rit
y
M
a
n
a
g
e
m
e
n
t
a
t
Kin
g
M
o
n
g
k
u
t'
s
Un
iv
e
rsity
o
f
Tec
h
n
o
lo
g
y
N
o
rth
B
a
n
g
k
o
k
.
His
sc
h
o
larly
p
u
rsu
it
s
a
re
fo
c
u
se
d
o
n
re
fin
i
n
g
e
m
a
il
c
las
sific
a
ti
o
n
m
e
th
o
d
o
lo
g
i
e
s
th
ro
u
g
h
th
e
i
n
n
o
v
a
ti
v
e
a
p
p
li
c
a
ti
o
n
o
f
m
a
c
h
in
e
lea
rn
i
n
g
(M
L)
tec
h
n
i
q
u
e
s,
a
imin
g
t
o
re
v
o
lu
ti
o
n
ize
h
o
w
d
a
ta
is
a
n
a
ly
z
e
d
a
n
d
i
n
terp
re
ted
i
n
t
h
is
d
o
m
a
in
.
He
c
a
n
b
e
c
o
n
tac
t
e
d
a
t
e
m
a
il
:
s6
5
0
7
0
3
1
9
1
0
0
2
6
@e
m
a
il
.
k
m
u
tn
b
.
a
c
.
th
.
Na
wa
p
o
r
n
Wisit
p
o
n
g
p
h
a
n
is
a
n
a
s
sista
n
t
p
ro
fe
ss
o
r
in
t
h
e
d
e
p
a
rtme
n
t
o
f
d
ig
it
a
l
n
e
two
r
k
a
n
d
i
n
fo
rm
a
ti
o
n
se
c
u
rit
y
m
a
n
a
g
e
m
e
n
t,
fa
c
u
lt
y
o
f
in
fo
rm
a
ti
o
n
tec
h
n
o
lo
g
y
a
n
d
d
i
g
it
a
l
i
n
n
o
v
a
ti
o
n
a
t
Kin
g
M
o
n
g
k
u
t'
s
Un
iv
e
rsity
o
f
Tec
h
n
o
l
o
g
y
No
rt
h
Ba
n
g
k
o
k
,
Th
a
il
a
n
d
.
S
h
e
re
c
e
iv
e
d
h
e
r
B.
S
.
,
M
.
S
.
,
a
n
d
P
h
.
D.
d
e
g
re
e
s
i
n
e
lec
tri
c
a
l
a
n
d
c
o
m
p
u
ter
e
n
g
in
e
e
rin
g
fro
m
Ca
rn
e
g
ie
M
e
ll
o
n
Un
iv
e
rsit
y
in
2
0
0
0
,
2
0
0
2
,
a
n
d
2
0
0
8
,
re
sp
e
c
ti
v
e
l
y
.
F
ro
m
2
0
0
3
to
2
0
0
9
p
rio
r
t
o
jo
i
n
i
n
g
KM
UTNB,
sh
e
wa
s
a
r
e
se
a
rc
h
e
r
at
th
e
El
e
c
tri
c
a
l
Co
n
tro
l
a
n
d
In
teg
ra
ti
o
n
Lab
o
ra
t
o
ry
,
G
e
n
e
ra
l
M
o
to
r
Re
se
a
rc
h
Ce
n
ter,
Warre
n
,
M
I,
USA.
Wh
il
e
h
e
r
e
x
p
e
rti
se
is
in
a
d
-
h
o
c
n
e
two
rk
s
a
n
d
v
e
h
icle
-
to
-
v
e
h
icle
c
o
m
m
u
n
ica
ti
o
n
,
h
e
r
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
sm
a
rt
e
n
v
ir
o
n
m
e
n
ts,
c
y
b
e
rse
c
u
rit
y
,
a
n
d
in
fo
rm
a
ti
o
n
tec
h
n
o
l
o
g
y
m
a
n
a
g
e
m
e
n
t.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
n
a
wa
p
o
r
n
.
w@
it
d
.
k
m
u
tn
b
.
a
c
.
t
h
.
Evaluation Warning : The document was created with Spire.PDF for Python.