TELKOM
NIKA
, Vol.13, No
.2, June 20
15
, pp. 373~3
8
0
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i2.1436
373
Re
cei
v
ed
Jan
uary 6, 2015;
Re
vised Feb
r
uar
y 9, 2015;
Acce
pted Fe
brua
ry 20, 20
15
An Introduction to Journal Phishings and Their
Detection Approach
Mehdi Dad
k
hah*
1
, Tole Sutikno
2
, Mo
ha
mma
d
D
a
va
r
p
a
n
ah
J
a
z
i
3
, Deris Stiaw
a
n
4
1
Departme
n
t of Computer a
n
d
Information T
e
chno
log
y
, F
o
u
l
ad Institute of T
e
chnolog
y
F
oula
d
sha
h
r, Isfahan 8
4
9
1
6
6
376
3, Iran
2
Departme
n
t of Electrical En
gi
neer
ing, Un
iver
sitas Ahmad D
ahl
an, Yog
y
a
k
arta, Indon
esia
3
Departme
n
t of Computer a
n
d
Information T
e
chno
log
y
, F
o
u
l
ad Institute of T
e
chnolog
y
F
oula
d
sha
h
r, Isfahan 8
4
9
1
6
6
376
3, Iran
4
Departme
n
t of Computer S
y
s
t
em Engin
eeri
n
g,
Univers
i
tas Sri
w
ij
a
y
a, Pal
e
mban
g, Indon
e
s
ia
*Corres
p
o
ndi
n
g
author; e-ma
i
l
: dadkh
a
h
80@
gmail.c
o
m
1
, tole@e
e.uad.
ac.i
d
2
A
b
st
r
a
ct
Now
adays, th
e most i
m
porta
nt risk and c
hal
len
ge i
n
on
lin
e
system ar
e on
li
ne sca
m a
nd
p
h
ishi
n
g
attacks. Phishing attacks hav
e
been alwa
ys
used to steal
important infor
m
at
ion
of user
s. In this kind
of
scam,
attacker
direct victi
m
to
fake p
ages
us
ing s
o
ci
a
l
e
ngi
neer
ing t
e
chn
i
ques, th
en, sta
r
ts stealin
g us
e
r
s`
important infor
m
ation suc
h
as
passwords. In order to confr
o
nting t
hes
e attacks, num
e
rous techniques have
bee
n inv
ente
d
w
h
ich hav
e the
abil
i
ty to confr
ont differe
nt
ki
nds of thes
e at
tacks. Our goal
in this
pap
er is
to
introd
ucin
g ne
w
kind of phish
ing attacks w
h
i
c
h are not
id
en
tifiabl
e by tech
niq
ues a
nd
me
thods w
h
ich h
a
v
e
been invent
ed
to confronting
phis
h
ing attack
s. Unlik
e other
kinds
of
phishing attacks which
target all kinds
of users, res
e
archers ar
e the
victim
s
of thes
e kinds of jour
nal phis
h
ing attacks.
Finally, we`ll
introduce an
appr
oach
bas
e
d
on cl
assificati
on al
gorit
h
m
s to ide
n
tify these
kind of jo
urn
a
l
phis
h
in
g attack
s and the
n
w
e
`l
l
check our su
gg
ested ap
pro
a
c
h
in error rate.
Ke
y
w
ords
: Phi
s
hin
g
, Hijack
ed
journ
a
l, Class
i
f
ication, Data
Minin
g
1. Introduc
tion
Phishin
g
atta
cks i
s
an effo
rt for acce
ssi
ng
peo
ple im
portant info
rmation like; usernam
e,
password, an
d credit card
s information,
using
so
cial
engin
eeri
ng t
e
ch
niqu
es [1]
.
These
attacks
were explai
n
ed in 19
87 b
y
details an
d
were u
s
ed i
n
1996 fo
r the f
i
rst time [2]. In these attacks,
t
o
incr
ea
sing
suc
c
e
s
s rat
i
o,
at
t
a
cke
r
s
t
r
y
t
o
r
eprese
n
t themselve
s
in a way that victims trust
them an
d a
c
cept th
em a
s
a le
gal
agen
ts of valid
organi
ztion
su
ch a
s
b
a
n
ks. I
n
the
s
e
kin
d
s of
attacks, phi
shers (fo
r
ge
rs
who
use phi
shing atta
cks),
begin
s
thei
r
plan by d
e
sig
n
ing a
web
s
it
e
whi
c
h i
s
simil
a
r to l
egal
website.
Havin
g
do
ne
thi
s
step, the mu
st
find a
way to pe
rsuad
e t
heir
victims to e
n
ter thei
r o
w
n
web
s
ite a
nd
enter
his/
he
r
informatio
n. So, main targ
et on a p
h
ish
i
ng
attack is to u
s
e
a fa
ke
co
n
nectio
n
which
begi
ns
with
a e-mail in
clu
d
ing fa
ke
URL from
a
ban
ks
or gove
r
nme
n
tal agen
cy. Attacke
r or p
h
ishi
ng atta
ck de
sign
er tries to use case
s which are
attractive to victims an
d ca
n pay their at
tenti
on. Then
tries to achi
eve name, p
hone n
u
mbe
r
or
any othe
r
kin
d
s
of info
rma
t
ion which
ca
n be
u
s
ed
fo
r a
d
van
c
ing
his
goal
s, o
n
the
other ha
nd
phishing
attacks a
r
e
used
to steal victi
m
s` i
dent
ifica
t
ion
usi
ng co
mputer netwo
rks.
The
s
e ki
nds
of attacks a
r
e desi
gne
d g
enerally by mean
s
of a
c
ce
ssi
ng to p
eople IDs a
n
d
passwo
r
d
s
. Bu
t
gene
rally, includ
es a
n
y kind of info
rmation whi
c
h
illegal use of them will
follow attackers
benefits.
Many studie
s
and efforts
have bee
n d
one to
introd
ucin
g differe
nt kind
s of phishi
n
g
attacks
and
their
conf
ron
t
ing ways.
Gene
rally, di
fferent
kind
s of phi
shi
n
g
attacks in
cl
ude
deceptive phi
shin
g [3], phishin
g based
on dest
r
u
c
tive softwa
r
e [4], web trojan
s [5], pharming
[6],
phishing inje
ction [7], phishing usin
g fake application
s
[8], domain
hijacking [9], spea
r phi
shi
ng
[10] and
cha
nging
u
s
er sy
stem
setting
s attacks
[6]. To confront
t
hese
atta
cks many
te
chniq
ues
and m
e
thod
s have b
een i
n
vented
su
ch
Sign-in
S
eal
[11], develo
p
ing exp
e
rt
system ba
se
d
on
cha
r
a
c
teri
stics of web p
a
g
e
s in
ord
e
r to
detec
t p
h
ishi
ng web
s
ites [
12], geneti
c
a
l
gorithm
ba
se
d
on anti
-
phi
shi
ng techniq
u
e
s
[13], dete
c
ti
on of p
h
is
hin
g
attacks ba
sed on
catego
rizing
su
pe
r lin
ks
[14], attribute-ba
se
d p
r
e
v
ention of p
h
ishi
ng atta
cks [1
5], con
t
ent based
on anti-phi
sh
ing
approa
ch [1
6
], confro
nting
phishing
atta
cks
by tw
o
st
ep ide
n
tificati
on [17], dete
c
tion
of phi
sh
ing
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 373 – 38
0
374
page
s b
a
sed
on a
s
so
ciat
ed relation
shi
p
s [1
8], dete
c
tion
of phi
shing p
age
s
b
y
comp
arin
g
the
amount of
dif
f
eren
ce between
th
e
ad
dress
string
an
d the
white li
st [19], ra
nki
ng ba
se
d a
n
ti-
phishing a
p
p
r
oa
ch [20],
and u
s
ing
d
a
ta mining
algorith
m
s [2
1]. Mentione
d method
s
and
techni
que
s i
d
entify differen
t
kind
s
of p
h
i
s
hin
g
atta
cks by 27
recogn
ized
featu
r
e
s
but a
r
e
usele
ss
again
s
t jou
r
n
a
l phi
shin
gs
whi
c
h
we
wa
nt to introd
uce be
cau
s
e
27
key featu
r
e
s
have be
en g
o
tten
from
m
o
st rel
a
ted web
s
ite
s
with e-comm
erce.
In
[22,
23]
the
s
e attacks have bee
n
ad
dre
s
sed
as
hijacke
d
journals a
nd som
e
feature
s
of these
kind of
phishing atta
cks have b
e
e
n
mentione
d
and
gene
ral
guid
e
s
on thi
s
h
a
ve be
en
given to
re
sea
r
che
r
s. In
[24
,
25] di
scussions ab
out f
a
ke
publi
s
he
r an
d
open
a
c
cess publi
s
he
r
we
re b
een ta
ke
n ca
re fo
r, b
u
t a definite
confronting
way
has n
o
t bee
n
sugg
este
d a
nd only few
guide
s have
been give
n to re
sea
r
che
r
s abo
ut pre
d
atory
publi
s
he
rs.
Our
goal
on
this pa
pe
r is to extrac
t
re
lated featu
r
e
s
by the
s
e
ki
nds
of phi
shi
n
g
attacks an
d p
r
esent a meth
od to detectin
g
and confro
n
t
ing with the
m
.
2. Introducti
on to Jour
na
l Phishings
In this pap
er
we`ll i
n
trod
uce journal phi
shin
g
s
and t
r
y to extract th
ese
kind
s of
phishing
attacks featu
r
e and fin
a
lly pre
s
ent a
ap
proa
ch
ba
sed
on cl
assified
algorith
m
s to
confront the
m
.
As mentione
d before, in p
h
ishi
ng attacks, phi
sh
e
r
s deceive victims by desig
n
i
ng a fake we
bsite
whi
c
h i
s
sim
ilar to th
e o
r
iginal
an
d u
s
ing
soci
al e
ngine
erin
g te
chni
que
s th
e
n
ste
a
l victi
m
sen
s
itive info
rmation i
n
cl
u
d
ing p
a
ssword by dire
ctin
g them into
fake
web
s
ite
s
an
d finan
ci
al
resou
r
ces. F
a
ke jou
r
n
a
ls
works with n
a
me and cre
d
it of some valid journals
but in fact have no
relation
shi
p
with tho
s
e
jo
urnal
s
and
like phi
shi
ng
attacks, follo
w
financi
a
l moti
vations
with t
h
is
differen
c
e th
a
t
in this ki
nd
of scammi
ng,
the vict
im
do
esn
`
t give
his/her
sen
s
itive
inform
ation t
o
forgers
but delivers financial resour
ces
directly to them. In this
kind of attack which
we
will
be
calle
d "jou
rn
al phi
shin
gs"
so
on, the
forge
r
s
mo
stly deceive
their victim
who
are m
o
stly
resea
r
chers, by desig
ning
a fake we
b
page and u
s
ing valid jo
urnal
s na
me
and ISSN. The
forge
r
s g
o
after journals
whi
c
h are active in
print versio
n and
by designi
ng
a website
with
original journal features st
art
scamming from
resear
chers and
by receivin
g high sum
s
, they
will
publi
s
h the victims pa
pers. In t
hese kinds of phi
shi
ng attacks a
s
de
ceptive phishing atta
cks,
so
cial engi
ne
ering i
s
exclu
s
ively used.
The pro
c
e
s
s of a journal
phishing
s
attack is sh
own
in
Figure 1.
attack
phi
shi
ng
Figure1. The
pro
c
e
ss of a j
ourn
a
l
jacke
d
jou
r
nal
s and
i
Table 1 al
so
sho
w
s com
m
on feature
s
b
e
twee
n phi
shi
ng attacks a
n
d
h
ja
cked jo
urnal
s a
s
one
of the phishi
ng attacks.
i
justifies nami
ng h
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
An Introdu
ction to Jou
r
nal
Phishin
g
s a
n
d
Their
Dete
ction Appro
a
ch (Meh
di Da
d
k
ha
h)
375
j
a
cked jou
r
nal
s)
Hi
(
s
o
m
mon features
betwe
en phi
shing attacks and jou
r
nal p
h
ishi
ng
C
.
Table 1
Phishing attacks
high jacked journals)
(
ournal phishing
J
eature
F
verage
A
Ver
y
high
Using social engineering
Ye
s
Ye
s
Having financial
motivations
Ye
s
Ye
s
s
mail to deceive the victim
-
Sending e
Ye
s
Ye
s
Using same name or domain
phishing attacks cer
t
ain
spear
In
victim
s are attent
ional
Ye
s
Choosing victim
Phishing w
ebsite
s
are usually
period
available for a short
Fake journals
w
e
bsites are usually
available for a short period of
time
Short life time of designed fake
we
b
s
i
t
e
Ye
s
Ye
s
Using available
w
e
aknesses in
internet prot
ocols like TCP/IP
3. Identif
y
in
g Journal Ph
ishing b
y
Classifica
tion
Algorithm
To cl
assifying
data a
nd m
e
cha
n
ism
choi
ce
s
p
r
o
c
e
ss cla
ssifying al
gorithm ca
n be
u
s
ed.
Cla
ssifying
al
gorithm i
s
a
p
p
lied to
cla
ssi
fying dat
a an
d extra
c
ting t
he sample
fro
m
a set of da
ta.
By classifying algo
rithm, a sam
p
le
ca
n be extr
a
c
t
ed from a
set of data, then u
s
e extracted
sampl
e
fo
r m
a
kin
g
d
e
ci
sio
n
ab
out futu
re data.
Cl
assi
fying
algo
rith
ms have different
type
s wh
ich
we can nam
e
C5, CHAI
D
, QUEST an
d C&R tre
e
as
the most po
p
u
lar. Extracte
d sampl
e
s fro
m
these al
gorith
m
s are mo
stly as a deci
s
io
n tree [26].
We
need
dat
a to extra
c
t related jo
urnal
phishing
s
fe
ature
s
, so
we
sh
ould
coll
ect a list of
kno
w
n jo
urn
a
l
phishin
gs
which h
a
ve be
en dete
c
ted.
We stu
d
y this list from aca
demic
re
sou
r
ce
s
whi
c
h
ca
n
be
provided. A
c
cording
F
r
om
ou
r o
b
se
rvati
on
o
n
colle
ct
ed
jo
urnal phi
shin
g web
s
ite
s
,
key features
of these phishing attacks
will be ex
tracted. Table 2
present
s these features wi
th
measurable a
m
ount for ea
ch.
feature
s
to recogni
zin
g
p
h
ishi
ng jou
r
n
a
ls.
Us
ed
2.
Table
3.1. Domain ranking
This feature will be checked in a relati
on with
website dom
ain. Because journal
phishing
s
a
r
e
the copy of t
he legal
we
b
s
ite, so
they
don`t h
a
ve hi
gh ra
nks in
search e
ngin
e
s
.
But, this feature i
s
not co
rrect all the ti
me, bec
au
se,
it might be a journal
witho
u
t web
s
ite an
d
sea
r
ch engi
n
e
s dete
c
t fake web
s
ite in
stead of the
legal one
(like hijacked
Jokull jou
r
n
a
l or
Measures
Kind
A
d
j
ect
i
v
e Na
me
Rank
Having page ran
k
=
1
Not having page
rank
=
0
Logical
Domain r
anking
1
Numbers of e
x
te
r
nal link less than 2
=
L
Numbers of e
x
te
r
nal links betw
een
2 and 7
=
M
Numbers of e
x
te
r
nal links more than 7
=
H
Discr
ete
Using extern
al links
2
Short lifetime
=
0
Long lifetime
=
1
Logical
Domain lifetime
3
Indexed
=
1
Not indexe
d
=
0
Logical
Indexing in pop
ular databases
4
Contained first 2
results
=
L
Contained 2 to 4
results
=
M
Other results
=
H
Discr
ete
Sequence in searching results
5
Among 1 to 4 co
untries
=
H
Among 4 to 8 co
untries
=
M
More than 8
cou
n
tries
=
L
No infor
m
ation
=
NA
Discr
ete
Entered count
ries to journal
w
ebsite
6
available
=
1
Not available
=
0
Logical
Availablity
of pre
v
ious issues
7
Long URL
=
1
Suitable URL
=
0
Logical
Long URL
8
Gene
ral aim and
scope
=
1
Specific aim and
scop
=
0
Logical
Journal aim and
scope
9
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 373 – 38
0
376
www.jo
kulljou
r
nal.com
whi
c
h h
a
s
ran
k
i
n
Goo
g
le sea
r
ch
engi
ne). I
n
mention
ed
method, we u
s
e
Googl
e search e
ngin
e
b
e
ca
use of th
e abilit
y to ranki
ng
web
s
i
t
e based o
n
page
ran
k
.
This
feature i
s
co
nce
r
ne
d a
s
a
Boolean va
ri
able in a
way
that if checki
ng we
bsite
h
a
s ran
k
ing, t
h
e
variable
will be 1 otherwise it will be 0.
3.2. Using ex
ternal links
This featu
r
e
con
c
ent
rate
s on che
c
ke
d web
s
it
e
s
cod
e
s
st
ru
ct
ure.
I
n
t
he ca
se t
hat
external lin
ks provide i
m
a
ges
with
che
c
ked
we
b
s
ite
conte
n
t, the web
s
ite i
s
susp
ecte
d to
be
journ
a
l phi
shi
ng be
cau
s
e
most of journ
a
l phish
ing
s
u
s
e othe
r we
b
s
ites
copi
ed content.
3.3. Domain lifetime
Acco
rdi
ng to
our
survey o
n
journ
a
l phi
sh
i
ngs, mo
st of these
web
s
ite
s
dom
ain hav
e been
regi
stered fe
w month
s
b
e
fore d
e
si
gnin
g
the fake
web
s
ite while so
me pap
ers a
c
cording to
ma
ny
years ag
o a
r
e availabl
e in
journal a
r
chi
v
e. So
, by using Whio
s da
tabases,
we
can
extra
c
t the
amount of th
is featu
r
e a
n
d
get on
det
ecting
phi
shi
ng jou
r
nal
s.
Suitable lifetime is m
e
a
s
u
r
ed
according to the first issue i
n
journ
a
ls.
3.4. Indexing
in popular datab
ase
s
In
gene
ral, in
dexed jo
urn
a
l
s in a
pop
ul
ar d
a
taba
se
are int
e
re
ste
d
and
have
value to
victims a
nd
can attra
c
t the
m
. One
of the
s
e i
ndexin
g
d
a
taba
se i
s
T
h
omso
n-Re
ute
r
s. Almo
st all
of
the jou
r
nal
p
h
ishi
ng
s a
r
e
detecte
d a
r
e
indexing
in
this d
a
ta b
a
se. But, we
should
con
s
id
er
cor
r
e
c
t data
b
a
se
be
cau
s
e
Cite Fa
cto
r
(
h
ttp://www.
cit
e
factor.
o
rg
) h
a
s in
dex
ed
al
most all
of the
hijacke
d
journals
with fake
addre
s
se
s a
nd it`s not
sui
t
able for su
rv
eying.
3.5. Sequenc
e in searchin
g results
This featu
r
e
has b
een
ad
ded to in
cre
a
se a
c
cu
ra
cy on detecti
ng phi
shin
g
page
s in
mentione
d m
e
thod. In this method the
title of
the concern
ed jo
u
r
nal h
a
s
bee
n se
archin
g
on
sea
r
ch engi
n
e
and the website ad
dre
ss will b
e
re
verse
d
. This
feature will b
e
used to de
tec
t
journ
a
l phi
shi
ngs
whi
c
h the
i
r legal on
e h
a
s ele
c
tro
n
ic
versio
n.
3.6. Entered
countries to
journal
w
e
bs
ite
Acco
rdi
ng to
our stu
d
ie
s
on availa
ble
journ
a
l p
h
ishi
ng
web
s
ite
s
i
t
has be
en d
e
tected
that ea
ch
jou
r
nal
phi
shin
g
victim b
e
lon
g
s to
a
certa
i
n count
ry o
r
incl
ude
s li
mi
ted po
pulatio
n.
Therefore
jo
urnal
p
h
ishi
n
g
s ca
n be d
e
tected
by A
l
exa data
b
a
s
e (http://www.alexa.com
)
and
cla
ssifying th
e web
s
ite gu
ests b
a
sed o
n
the cou
n
try.
3.7. Av
ailablit
y
of pre
v
iou
s
issues
Previou
s
issu
es i
n
jou
r
nal
phishing
s
a
r
e
not u
s
ually
a
v
ailable o
r
ju
st some
of th
em a
r
e.
Phishe
rs pre
v
ent user a
c
ce
ssi
ng previous is
su
e b
y
designin
g
a login page
for acce
ssi
ng
previou
s
i
s
su
e or me
ntioni
ng writers` n
a
mes
or
the
pape
rs
su
bje
c
ts. The
rea
s
on of this is t
hat
desi
gning a
web
s
ite with
previou
s
issu
e gets a
lot of time and sometime
s be
cau
s
e the forger
doe
sn't acce
ss all previo
us
issue
s
.
3.8. Long URL
Some journal
phishi
n
g
s
use a long URL
.
Long URL
s
are u
s
ually u
s
ed to hid
e
d
oubtful
parts o
n
a
ddress b
a
r. A
c
cordin
g to
scie
ntific p
r
in
ciple
s
, the
r
e i
s
no
stan
da
rd l
e
n
g
th to d
e
tecti
n
g
legal URL
s
from illegal o
n
e
s but n
o
rm
a
lly, if a
URL seems l
ong thi
s
might bel
on
g to a phishin
g
web
s
ite.
3.9. Journal aim and sco
pe
Most
of the j
ourn
a
l p
h
ishi
ngs a
r
e i
n
a
way t
hat
accept pa
pers wi
th differe
nt subje
c
ts
or
have ge
neral
aim and
scop
e. Most of th
ese j
our
nal
s
have spe
c
ific
name
s
which
don`t
rep
r
e
s
ent
s
ubjec
t domain
(lik
e Walia or
http://www.
waliaj.
c
o
m)
or their
s
ubjec
t
s
are in a way that c
o
nc
lude
different re
se
arch fields
(like Jou
r
nal of tech
nolo
g
y).
In orde
r to
choo
se the
su
itable alg
o
rith
m to dete
c
t journ
a
l phi
shi
ngs, first we
need to
provide
a training d
a
taset
of journal p
h
i
s
hin
g
s
and
th
e amo
unt of 9
mention
ed fe
ature
s
in ta
bl
e 2
for each phi
shin
g web
s
it
e need to be measure
d
, then classifi
ed algo
rithm
s
be appli
e
d
on
provide
d
trai
ning data
s
et
and a
c
cordi
ng to erro
r
ratio, the mo
st suita
b
le al
gorithm
will
be
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
An Introdu
ction to Jou
r
nal
Phishin
g
s a
n
d
Their
Dete
ction Appro
a
ch (Meh
di Da
d
k
ha
h)
377
cho
s
e
n
. Used
training
data
s
et mu
st in
clu
de phi
shi
ng
website
s
in
ad
dition to leg
a
l
and m
a
in on
es
to be able to detec
t original
webs
ites
t
oo. We
use IBM SPSS
Modeler data
mining tools
for
classifying data. Accordi
n
g to fi
nished feature Selectio
n analysi
s
, t
h
e possibility
of each feature
in dete
c
ting
phishing
atta
cks
ha
s b
e
e
n
re
presente
d
in T
able
3.
All of the fe
ature
s
with fi
rst
prio
rities
can
be sele
cted a
s
the tre
e
ro
ot. This
is i
m
portant, be
ca
use it i
s
po
ssible that one
of
the feature in
the site ca
n`t
be measure
d
and d
e
ci
sio
n
can
be ma
de by usi
ng
different feat
ure
as the tree root. It is important to say that if
the root feature
chang
es, the
deci
sio
n
tree
will
cha
nge.
Table 3. The
effect of the each featu
r
e o
n
detectin
g
jo
urnal p
h
ishin
g
s
Impor
tan
t
Feature
1
Domain lifetime
1
Availablity
of pre
v
ious issues
1
Domain ranking
1
Journal aim and
scope
1
Entered count
ries to journal
w
ebsite
0.997
Indexing in pop
ular databases
0.964
Sequence in searching results
0.85
Using extern
a
l links
0.838
Long URL
Figure 2 rep
r
ese
n
ts differe
nt algorithm
s error ratio.
So, ac
c
o
rding to this
er
ror rati
o, C5 algo
rith
m
will be the most suitabl
e algorithm for
det
ecting journal
phishi
ng attacks.
Figure 2. Algorithm
s with their e
rro
r rati
o in cla
ssifyin
g data
Mean
while, F
i
gure
3 sh
ows de
cisi
on tre
e
s in a
ca
se
t
hat different f
eature
s
a
r
e
selecte
d
as
ro
ot.
But we
shoul
d co
nsi
der thi
s
if ro
ot featu
r
e
chan
ge
s, e
rro
r ratio will
cha
nge
a little but the a
m
ount
of these
changes are lo
w.
We
used IBM
SPSS Model
er appli
c
ation to analy
z
e
data and extract
deci
s
io
n tree
based o
n
ou
r gathe
red
dat
a and
built dif
f
erent d
e
ci
sio
n
tree
s by
ch
oosi
n
g differe
n
t
feature
s
a
s
th
e ro
ot. This is impo
rtant th
at t
he value
o
f
root featu
r
e
may not b
e
measurable
a
nd
cho
o
si
ng an
o
t
her feature can do ma
king
deci
s
ion a
s
the root.
0
0.5
1
1.5
2
2.5
3
3.5
4
C5
CHAID
QUEST
C&R Tree
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 373 – 38
0
378
a) Domain lif
etime as root feature
b) Availabilit
y of previous
issue as a root feature.
c)
Jou
r
nal
scope a
s
root fe
ature
d)
Page ran
k
in
g as ro
ot feature
Figure 3. Usi
ng of different
feature
s
as root in deci
s
io
n tree
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
An Introdu
ction to Jou
r
nal
Phishin
g
s a
n
d
Their
Dete
ction Appro
a
ch (Meh
di Da
d
k
ha
h)
379
4. Measuring
error ratio
To mea
s
u
r
e
mention
ed
method
error ratio, di
ffere
nt dataset from previou
s
trainin
g
dataset shoul
d be
u
s
ed
th
en a
c
co
rding
to a
c
hieve
d
results, cal
c
u
l
ate
e
rro
r rati
o.
According
to
experim
ents,
our ap
pro
a
c
h resi
sts a
gain
s
t errors
beca
u
se if just one the
root feature
s
be
inacce
ssible,
makin
g
de
cisi
on is po
ssib
le
by choo
sing
anothe
r featu
r
e as
root.
5. Conclusio
n
and Futu
r
e
Work
In this pap
er
we di
scusse
d
about a n
e
w
kind
of phishi
ng attacks wh
ich
were dete
c
ted a
s
journ
a
l p
h
ishi
ng a
nd
menti
oned
ou
r
rea
s
on
for th
is n
a
ming. T
hen,
we
d
e
tecte
d
key
features of
this kin
d
of phishing atta
cks an
d presented
a ne
w approa
ch fo
r dete
c
ting them. Mentio
ned
approa
ch u
n
l
i
ke all
pa
st method
s to
confront
ing
p
h
ishi
ng, ha
s
the ability to detect p
h
ish
i
ng
journ
a
ls.
Usi
ng this meth
od with
the
combinatio
n o
f
past u
s
e
d
method
s to
confront
phi
sh
ing
attacks
can
b
e
a pa
rt of future
efforts. New
p
r
e
s
ente
d
feature
s
in t
h
is p
ape
r
can
be ad
ded to
27
kno
w
n p
h
ishi
ng attacks feature
s
an
d prese
n
t a
more
perfe
ct meth
od rath
er than pas
t methods
to confronting
to different ki
nds of phi
shi
ng attacks.
Ackn
o
w
l
e
dg
ements
We would li
ke to ackno
w
l
edge to Ma
rwan M. Ob
ei
dat from De
partme
n
t of English
Lang
uage
an
d Literatu
re, Ha
shemite
University, Zarqa, Jordan.
Referen
ces
[1]
Dadk
hah M, Ja
zi M.D, L
y
as
he
nko V. Predicti
on of
ph
ishi
ng
w
e
bsit
es usi
n
g
classificati
on
alg
o
rithm
s
base
d
on
w
e
i
g
ht of
w
e
b
p
age
s char
acteristic
s.
Jo
u
r
na
l o
f
Ma
th
em
a
t
i
c
s an
d Te
ch
no
lo
g
y
. 2
014;
5(2)
:
24-35. DOI: 10.7813/jmt.2014/5-2/4.
[2]
San Marti
no A,
Perramo
n
X. Phish
i
ng S
e
cre
t
s: Histor
y
, Effe
cts, and Co
unt
ermeas
ures.
In
ternatio
nal
Journ
a
l of Net
w
ork Security
. 201
0; 11(3): 16
3-17
1.
[3]
Mahmo
od M, Rajam
a
n
i
L. APD: ARM Deceptive
Ph
ishi
n
g
Detector S
y
s
t
em Phishi
ng
Detectio
n in
Instant Messe
n
gers Us
in
g D
a
ta Min
i
n
g
Ap
pro
a
ch.
Glo
bal
T
r
ends
in
Co
mpu
t
ing
and
Co
mmunic
a
tio
n
System
s
. 20
12
; 269; 490-
502.
[4]
Li S, Schmitz R.
A Novel Anti-Phis
h
in
g F
r
amew
ork Base
d on Ho
neyp
o
ts. eCrime Re
search
ers
Summit (IEEE).
T
a
coma, WA. 2009; 1-
13. DOI:10.1109/ECRI
ME.2009.5342609.
[5]
Johns
on M. Eric.
Man
agi
ng
Informatio
n
Risk and
the
Econom
ics
of
Secur
i
t
y
. 20
0
9
th
Editi
on.
German
y
: S
p
ri
nger. 20
09: 1-4
0
.
[6]
Dadk
hah M,
Davar
pan
ah J
a
zi M. Secure
Pa
y
m
ent in
E-commerce:
Deal
w
i
th Ke
ylog
gers a
n
d
Phish
i
ngs.
Inte
rnatio
nal
Jour
n
a
l of E
l
ectro
n
i
cs Co
mmu
n
ica
t
ion a
n
d
Co
mp
uter En
gin
eeri
n
g
. 2
014
;
5(3): 656-
66
0.
[7]
Alkhate
eb F
,
Manasr
ah A, Bsoul A. Bank
W
eb Sites Ph
ishin
g
Det
e
ctio
n and
Notificat
i
on S
y
ste
m
Based
o
n
Sem
antic W
e
b tec
hno
log
i
es.
Inte
rnatio
nal
Jo
urn
a
l
of Secur
i
ty
& Its Appl
icati
ons
. 20
12
;
6(4): 53-6
6
.
[8]
Schle
gel R, Z
h
ang K, Z
hou
X. Y,
Int
w
al
a M,
Kapad
ia A, Wang
X. Sound
comber: A Stealth
y
an
d
Conte
x
t-A
w
a
r
e
Sound T
r
ojan
for Smartphon
es.
In NDSS
. 2011; 11: 1
7
-33.
[9]
Cha
ndav
al
e A
.
A, Sapkal A.
M.
Algorith
m
f
o
r Secur
ed O
n
lin
e Aut
henti
c
ation
Usin
g
CAPT
CHA
.
Procee
din
g
s o
f
the third Int
e
rnati
ona
l C
o
nf
erenc
e o
n
Emergi
ng T
r
ends i
n
En
gin
eeri
ng a
n
d
T
e
chnology
.
Goa. 201
0: 292
– 297. DOI: 10.110
9/ICET
E
T
.2010.126.
[10]
Hong J.
T
he Stat
e of Phishing Attacks.
Commu
n
ic
ations of
the ACM
. 201
2
;
55(1): 74-81.
[11]
Agar
w
a
l
N, R
enfro S, B
e
j
a
r A.
Yah
o
o
Sign-In
Se
al and
Curr
ent Anti-Phis
h
in
g Soluti
ons
.
Procee
din
g
s of
W
eb 2.0 Secu
rit
y
& Privac
y
W
o
rkshop. 20
0
7
: 1-4.
[12]
Aburro
us M,
Hossai
n
M. A,
Dah
a
l
K, T
habat
ah
F
.
Intelli
gent P
h
ish
i
n
g
Detectio
n S
y
st
em for E-
Banki
ng Usi
ng
F
u
zz
y
Dat
a
Minin
g
.
Expert Systems w
i
th App
licatio
ns
. 20
10;
37; 791
3–7
92
1.
[13]
Shreer
am
V,
Suba
n
M, S
h
anthi P,
Manjula K.
A
n
ti-phis
h
ing detection of
phis
hing
attacks using
gen
etic al
gorit
hm
. Proc
eedings of IEEE In
ternational Conference
on
Communication Control and
Comp
uting
T
e
chnolog
ies
(ICCCCT
). Rama
nath
a
puram. 2
0
1
0
:
447
–
450.
DOI:
10.11
09/ICCC
CT
.2010.567
0
593.
[14]
Chen J, Guo C.
Onlin
e
D
e
tection an
d Preventi
on of
Phish
i
n
g
Attacks
. Procee
d
i
ngs
of first
International Conferen
ce on Communications
and
Net
w
o
r
king
(IEEE). China.
2006: 1-7.
DOI:
10.11
09/C
H
INACOM.2006.3
447
18.
[15]
Atighetch
i M, Pal P. A
ttrib
ute-bas
ed Pre
v
entio
n of Phishi
n
g
Attacks. Proceedi
ngs
of eight
h
Internatio
na
l Sy
mp
osi
u
m on
Net
w
ork C
o
mp
uting a
nd A
ppl
i
c
ations (IEEE). Cambri
dge. 2
009: 2
66
–
269. DOI: 10.1109/
NCA.2009.13.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 373 – 38
0
380
[16]
Dunl
op M, Gr
oat S, She
l
l
y
D.
Gold Ph
ish
:
Using I
m
a
g
e
s
for Conte
n
t-
Based P
h
is
hin
g
Ana
l
ysis
.
Procee
din
g
s o
f
the fifth
Internatio
nal C
onfe
r
enc
e o
n
Internet Monitor
i
ng
and Protectio
n
(IEEE).
Barcel
ona. 2
0
1
0
: 123–
12
8.
[17]
Mishra M, Ga
urav, Jai
n
A.
A Prevent
iv
e Anti-Phis
h
in
g T
e
chnique
usi
ng
C
o
d
e
w
o
r
d
.
Internatio
na
l
Journ
a
l of Co
mputer Scie
nc
e
and Infor
m
ati
o
n T
e
chno
lo
gies
. 2012; 3(3): 42
48-4
250.
[18]
Liu G, Qiu B
,
W
e
n
y
i
n
L.
Auto
matic D
e
tection
of Phis
hin
g
T
a
rg
et from
Phis
hin
g
W
ebpa
ge
.
Procee
din
g
s of
Internatio
nal C
onfere
n
ce o
n
P
a
tte
rn Reco
gn
ition (IEEE). Is
tanbu
l. 201
0: 41
53-4
156.
DOI: 10.1109/I
C
PR.2010.1010.
[19]
Red
d
y
V.P, Ra
dha V, Ji
nd
al
M. Client Sid
e
protectio
n
from Phishi
ng attac
k
.
Internation
a
l
Journa
l of
Advanc
ed En
gi
neer
ing Sci
enc
es and T
e
ch
no
logi
es
. 20
11; 3
(
1): 39-45.
[20]
Khon
ji M, Jo
ne
s A, Iraqi Y.A.
Novel
Phis
hi
ng
Class
ificatio
n
Based
On UR
L F
eatur
es
. Procee
din
g
s
of GCC Conference and
E
x
hibition (IEEE). Duba
i. 2011:
221 –
224. DOI:
10.11
09/IEEE
GCC.
201
1.57
5
250
5.
[21]
Ruth R
a
m
y
a K
,
Pri
y
a
n
ka K, A
nush
a
K, J
y
ost
hna
Devi C
H
, Siva Pras
ad Y.
A. An Effective Strateg
y
for Identif
yin
g
Phish
i
ng W
e
bs
ites usi
ng C
l
as
s-Based A
ppr
o
a
ch.
Intern
atio
nal J
ourn
a
l
of Scientific &
Engi
neer
in
g R
e
searc
h
. 201
1; 2(12): 1-7.
[22]
M Jala
lia
n, H
Mahb
oob
i. Hi
ja
cked J
ourn
a
ls
and
Pred
ator
y
Publ
isher
s: Is
T
here a N
eed
to Re-T
hi
n
k
Ho
w
to Assess
the Qualit
y
of Academ
ic Res
earch
?.
Walail
ak J Sci & Tech
. 2014; 1
1
(5): 389-
394.
[23]
Dadk
hah M,
Obeid
a
t MM,
Jazi MD, Suti
k
no T
,
Riyad
i
MA. Ho
w
C
an W
e
Identif
y Hi
jack
ed
Journ
a
ls?.
B
u
lletin
of E
l
ec
trical E
ngi
ne
e
r
ing
an
d Inf
o
rmatics
. 2015; 4(2):
83-87. DOI:
10.12
92
8/eei.v
4i2.4
49.
[24]
Luki
æ
T
i
n, Blešiæ Ivana, Ba
sarin Bi
lja
na, Iv
anov
iæ Bi
bi
æ Lju
b
ic
a, Mil
o
ševi
æ Drag
a
n
, Sakulsk
i
Duša
n. Pre
dat
or
y a
n
d
F
a
ke
Scientific
Jour
n
a
ls/P
ub
lish
e
rs–
A Glo
bal
Outb
reak
w
i
th
Risi
n
g T
r
end: A
Revie
w
.
G
eogr
aph
ica Pa
nno
n
i
ca
. 201
4; 18(3
)
: 69-81.
[25]
W
illiams E. N
w
ag
w
u
. Ope
n
A
ccess Initiative
s in
Africa- Structure, Inc
entiv
es and D
i
sinc
e
n
tives.
Th
e
Journ
a
l of Aca
d
e
m
ic L
i
brar
ia
nshi
p
. 2013; 39(1): 3-10. DOI: 10.
1016/j.acalib.2012.11.024
[26]
R Kumar, R Verma. Classification
Al
gorithm
s for Data Mining:
A
Surv
e
y
.
Internati
o
n
a
l
Journ
a
l of
Innovati
ons i
n
Engi
neer
in
g an
d T
e
chno
lo
gy
. 201
2; 1(2): 7-1
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.