Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
N
o.
2
,
A
ugus
t
20
21
,
pp.
980
~
992
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v
23
.i
2
.
pp
980
-
992
980
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
An ont
ology d
riven qu
estion ans
werin
g system f
or fat
awa
re
t
riev
al
Be
nlah
arche
Kelto
um
1
, Z
akaria
La
boudi
2
, Nab
il
a N
ouaouri
a
3
,
Djam
el
Eddine Z
ego
ur
4
1
Depa
rtment of
m
at
hemati
cs
and
computer
scie
n
c
es,
Unive
rsit
y
of
Oum
El
Bou
agh
i,
Alg
eria
2
Resea
rch
l
abor
a
tor
y
on
compute
r
scie
n
ce'
s
complex
s
y
stems
,
Uni
ver
sit
y
of
Oum
El
Bou
aghi
,
Alg
eri
a
3
La
bora
tor
y
of
C
om
ple
x
S
y
st
ems
,
Bad
ji Mokhta
r U
nive
rsit
y
,
Alge
ria
4
LCSI
La
bor
at
or
y
,
E
cole
na
ti
on
ale
Supéri
eur
e
d
'
In
form
at
ique
ESI,
Alger
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ja
n
2, 20
2
1
Re
vised
Jun
30
, 2021
Accepte
d
J
ul
7, 20
21
Thi
s
work
ai
m
s
to
propose
a
s
y
s
te
m
for
the
Alge
ria
n
Fat
awa
Ho
use
in
orde
r
to
fac
i
li
t
at
e
th
e
t
ask
of
the
Expe
r
t
Mufti
who
is
r
esponsible
of
gi
ving
fat
awa
for
Alger
ia
n
pe
ople
inqui
ri
es.
In
fac
t,
as
thi
s
house
is
rec
ent
a
nd
does
not
have
suffi
ci
en
t
h
um
an
resourc
es,
it
is
difficul
t
to
sati
sf
y
a
ll
inqui
r
ie
s
coming
dai
l
y
;
th
is l
ea
ds t
he
aske
rs t
o
wait for
a
long
ti
m
e
bef
ore
getting
an
sw
ers.
The
proposed
s
y
ste
m
al
lows
the
aske
rs
to
expr
ess
conc
ern
s
they
m
a
y
hav
e.
B
y
using
a
ca
se
-
b
ase
d
re
asoning
m
ec
han
ism
combined
with
ontol
o
g
y
dom
ain
,
the
s
y
st
em t
rie
s t
o
ret
ri
eve
sim
il
a
r
ca
ses from the
knowledge
base
.
In
the
c
ase
where
the
resp
onse
al
re
ad
y
ex
ists,
the
s
y
stem
imm
edi
at
ely
p
rovide
s
the
answer
to
the
as
ker
s.
Other
wise
,
an
inque
r
y
is
a
utomati
c
al
l
y
for
m
ula
te
d
and
sent
to
the
e
xper
t
Mufti
-
which
is
a
ce
rt
ifi
ed
schol
ar
-
in
orde
r
ei
the
r
to
val
ida
t
e
the
gene
r
ated
res
ponse
b
y
the
s
y
s
te
m
or
give
a
ne
w
answer.
Such
a
questi
on
-
answeri
ng
s
y
st
e
m
m
ay
b
e
ver
y
hel
pful
for
aske
r
s
to
get
the
ir
an
sw
ers
faste
r
since
it
a
ll
ows
both
the
storage
of
pre
vious
fat
awa
s
and
the
ir
ret
ri
eva
l
f
o
r
proc
essing
comi
ng
inqui
ri
es.
To
val
id
at
e
our
pr
oposal,
we
r
ely
on
fat
awa
s
conc
ern
ing
th
e
I
slamic
fin
anc
e
a
nd
banki
ng
tra
ns
ac
t
ions
dom
ai
n.
Overa
ll,
t
h
e
result
s a
r
e enc
ou
rag
ing and
sa
ti
sf
ac
tor
y
.
Ke
yw
or
d
s
:
Ca
se b
ase
d
-
rea
so
ni
ng
Do
m
ai
n
ontolo
gy
Ex
per
t m
uf
ti
Fata
wa
s
upport
Inform
at
ion
r
et
rieval
Qu
est
i
on an
s
w
erin
g
syst
em
s
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Be
nlahar
c
he
Kel
tou
m
Dep
a
rtm
ent o
f m
at
he
m
at
ic
s
a
nd co
m
pu
te
r
sc
ie
nces
Un
i
ver
sit
y o
f Ou
m
El Bou
a
ghi
B.P.35
8 route
de
C
on
sta
ntine
, Oum
El Bou
a
gh
i
0400
0,
Algé
rie
Em
a
il
: ben
la
ha
rch
e
.k
el
to
um@gm
ai
l.co
m
1.
INTROD
U
CTION
In
20
17,
the
A
lgerian
go
vern
m
ent
re
store
"t
he
Fata
w
a
ho
us
e"
after
a
lo
ng
ti
m
e
of
inact
ivit
y
since
the
ind
e
pe
nd
e
nc
e
in
1962.
T
hi
s
hous
e
is
an
offici
al
or
ga
ni
zat
ion
that
pro
vid
e
s
F
ata
w
a
(
answers
)
f
or
re
qu
e
sts
of
the
Alge
ria
n
people.
I
n
Islam
ic
le
gislat
ion
,
a
F
atwa
(sing
ular
of
fatawa
or
f
ata
wa
s
)
c
onsist
s
in
a
declarat
io
n
of
a
le
gal
opini
on
that
c
om
plies
with
Islam
ic
pr
e
cepts.
A
F
atwa
is
a
le
gal
sta
tem
ent
in
I
slam
,
gen
e
rated
by
a
Mufti
or
a
reli
gious
la
wyer,
on
a
sp
eci
fic
issue.
F
ata
w
a
s
are
aske
d
by
ju
dg
e
s
or
in
div
i
dual
s
in
cases w
he
re a
n i
ssu
e
of Fi
qh is unde
ci
de
d or
un
ce
rtai
n. La
w
su
it
s can
be
set
tl
ed
on the
basis of a
f
atwa
[1]
.
Accor
ding
to
"
Alge
rian
F
atwa
ho
us
e",
t
here
is
sti
ll
a
la
ck
of
a
n
a
uto
m
ated
syst
em
fo
r
storing
a
nd
retrievin
g
a
nswere
d
fat
awa
.
Indee
d,
e
ver
y
day,
the
ho
use
receives
a
la
rg
e
nu
m
ber
of
quest
io
ns
t
hro
ugh
diff
e
re
nt
m
et
ho
ds
(
e.
g.
,
phon
e,
fax
a
nd
web
sit
e
),
so
m
e
of
wh
ic
h
ha
ve
al
r
eady
b
ee
n
ask
ed
be
fore.
I
n
f
act
,
as
the
hous
e
does
no
t
inclu
de
suffici
ent
hum
an
resour
ces
,
it
i
s
ver
y
dif
ficult
to
sat
isfy
all
t
hese
fataw
a
re
qu
e
sts
in
par
ti
c
ular
si
nce
it
s
act
ivit
y
on
ly
sta
rted
t
hr
ee
ye
ar
s
ag
o.
Th
us
,
t
he
ta
sk
of
the
E
xper
t
Mufti
beco
m
es
ver
y
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An o
nto
lo
gy dr
iv
en
quest
io
n a
ns
we
rin
g
sy
ste
m
for
fat
awa
r
et
rie
val
(
Ke
lt
ou
m
Be
nlaha
rc
he
)
981
di
ff
ic
ult
giv
e
n
that
he
is
the
on
ly
respo
nsi
ble
of
gi
ving
ans
wers
(
fat
awa
)
to
al
l
the
qu
est
io
ns
re
cei
ved
,
especial
ly
that he
m
anag
es
ot
her ta
sk
s
(
e.
g.,
te
aching at
the
un
i
ver
sit
y
,
a
nd
reli
gious
duti
es)
.
Starti
ng
from
t
hat
co
ntext,
w
e
pro
pose
a
n
e
ff
ic
ie
nt
a
uto
m
at
ic
qu
est
io
n
-
a
nsweri
ng
syst
em
to
proces
s
Fata
w
a
re
qu
es
ts
through
ans
wer
s
a
nd
ar
gu
m
ents
fr
om
Qu
ra
n,
S
unna
h
and
I
j
ti
ha
d.
T
he
idea
co
ns
is
ts
in
bu
il
di
ng
a
n
int
el
li
gen
t
case
-
ba
sed
syst
em
that
can
stor
e
p
r
evio
us
fata
wa
and
retrieve
t
hem
as
need
ed
.
T
he
need
f
or
subm
i
tt
ing
re
quest
s
c
an
be
opti
m
iz
e
d
since
sim
i
la
r
requests
a
re
al
read
y
processe
d
by
rec
ordin
g
their
answers
in
t
he
knowle
dge
ba
se
.
I
n
t
his
wa
y,
the
E
xpert
S
ch
olars
will
on
ly
fo
c
us
on
the
ne
w
issues
.
T
his
m
akes
t
he
org
anizat
ion
ha
nd
le
m
or
e
a
nd
m
or
e
us
e
rs
with
ou
t
hum
an
interve
ntio
n
wh
il
e
increas
i
ng
the
abili
ty
to
ans
we
r
in
quirie
s
.
Ce
rtai
nly,
this
hel
ps
both
as
ke
r
a
nd
the
Ex
pe
rt
Mu
fti
in
sa
ving
e
ffor
t
a
nd
tim
e
w
hile
processi
ng in
quirie
s.
Fo
r
this
pu
rpo
se,
the
case
-
ba
sed
rea
sonin
g
(CBR
)
pa
rad
i
gm
is
inv
olv
e
d.
It
is
co
ns
ide
re
d
as
on
e
of
the
m
os
t
po
pu
l
ar
ap
proac
hes
for
de
velo
ping
know
le
dge
-
ba
sed
syst
em
s
du
e
to
it
s
abili
t
y
to
retrieve
a
nd
reu
se
so
luti
ons
t
hat
hav
e
w
orke
d
f
or
sim
il
ar
sit
uations
in
the
pa
st.
This
pr
i
nci
ple
is
sim
i
la
r
to
the
way
the
hu
m
an
so
lves
pro
ble
m
s
[2]
.
CB
R
of
te
n
us
es
sim
ilarity
m
easur
em
ents
to
ide
ntify
the
cases
t
ha
t
are
m
or
e
rel
evan
t
t
o
the
co
ns
ide
red
prob
le
m
[3]
.
I
n
fact,
al
th
ough
the
re
we
re
s
om
e
atte
m
pts
t
o
in
vo
l
ve
the
CB
R
par
adi
gm
to
dea
l
with
F
ata
w
a
i
ssu
es
[
4],
[
5]
,
the
se
a
ppro
a
ch
es
are,
ho
wever,
base
d
on
s
ynta
x
-
base
d
m
easur
em
ents.
I
nd
e
e
d,
synta
x
-
base
d
m
easur
em
ents
hav
e
pr
ov
e
n
–
in
m
any
cases
–
uns
uccess
f
ul
due
t
o
sem
antic
issues
th
at
m
ay
occur.
To
over
com
e
this
lim
itati
on
,
we
pro
pose
to
buil
d
a
do
m
ai
n
ontolo
gy
of
a
giv
e
n
do
m
ai
n
to
s
upport
the
sem
antic
si
m
ilarity
m
easur
em
ent.
Me
asur
i
ng
a
dista
nce
between
c
oncepts
is
an
essenti
al
pr
oc
ess
.
Most
m
et
ho
ds use
d f
or m
easur
ing, t
hey usuall
y d
o no
t
[
6]
ta
ke
se
m
antic
f
or c
onsiderati
on
[6]
.
In
or
der
to
vali
date
our
pro
po
sal
,
we
hav
e
i
m
ple
m
ented
and
te
ste
d
a
pro
toty
pe
to
deal
with
fatawa
in
Islam
ic
finan
c
e
an
d
b
an
king
tra
ns
act
io
ns
dom
ai
n.
Global
ly
,
the
ob
t
ai
ned
res
ults
are
pr
om
isi
ng
an
d
encou
rag
i
ng and th
us sh
ow th
e eff
ect
i
ven
e
s
s
of the
prop
os
a
l.
The
re
st
of
t
hi
s
pap
e
r
is
str
uctu
red
as
f
ollow
s
,
s
ect
ion
2
pro
vi
des
som
e
basic
con
c
epts
ar
ou
nd
qu
e
sti
on
-
an
sw
erin
g
syst
e
m
s,
CB
R
par
adig
m
and
on
tol
ogie
s.
Sect
ion
3
giv
es
detai
ls
about
the
propose
d
arch
it
ect
ure
of
the
syst
em
.
Sect
ion
4
s
umm
arizes
an
d
disc
us
ses
so
m
e
il
l
us
trat
ives
exa
m
ples
and
sce
nar
i
os.
Final
ly
,
s
ect
io
n 5
pro
vid
es
som
e con
cl
usi
on
s and
final
rem
ark
s
.
2.
BASI
C CO
N
CEPTS
2.1.
Que
stion
n
a
nswe
ri
ng
s
ys
te
m
Qu
est
i
on
-
a
nsw
erin
g
(
QA)
sy
stem
s
in
inf
orm
at
ion
retriev
al
are
ta
sk
s
t
ha
t
autom
at
ic
a
lly
answ
e
r
th
e
qu
e
sti
on
s
as
ke
d
by
hu
m
ans
in
na
tural
la
ngua
ge
us
i
ng
e
it
her
a
pre
-
st
r
uctu
red
data
ba
se
or
a
colle
ct
ion
of
natu
ral
la
ngua
ge
do
c
um
ents
[7]
,
[8]
.
I
n
orde
r
to
bette
r
unde
rstan
d
what
QA
syst
em
s
are,
t
he
ass
oc
ia
te
d
te
rm
ino
lo
gy
is
firstly
prov
i
de
d,
nam
ely
Qu
e
sti
on
P
hrase
,
Qu
est
i
on
Type
,
Answer
Ty
pe
and
Q
uestio
n
Foc
us
.
The
te
rm
Qu
es
ti
on
P
hrase
is
the
par
t
of
th
e
quest
io
n
that
spe
ci
fies
w
hat
e
xa
ct
ly
is
so
ug
ht.
The
te
rm
Qu
e
sti
on
Type
re
fer
s
t
o
a
cat
ego
rizat
io
n
of
the
quest
ion
reg
a
rd
i
ng
it
s
purpose
.
The
te
r
m
Answer
Type
re
fer
s
t
o
a
cl
ass
of
obj
ect
s
that
the
qu
est
i
on
i
s
lookin
g
f
or
.
Finaly
,
the
te
r
m
Qu
est
ion
F
ocus
is
the
pro
per
ty
or
entit
y
bein
g
searche
d by th
e questi
on
[
9]
.
2.2.
Isla
mi
c fi
na
nce
and
banking
In
la
st
fe
w
de
cades,
Islam
ic
finan
ce
a
nd
bankin
g
(
IF
B
)
has
gro
w
n
expo
nen
ti
al
ly
and
it
is
now
reg
a
rd
e
d
as
one
of
the
m
os
t
i
m
po
rtant
tren
ds
in
the
global
finan
ci
al
in
dustry.
In
dee
d
,
t
hat
kind
of
fi
na
ncial
transacti
ons
at
tract
s
not
only
I
slam
ic
co
m
m
u
niti
es,
but
al
so
m
any
people
over
al
l
the worl
d
f
or
m
any
reaso
ns
,
a
m
on
g
oth
e
rs:
i
)
m
aking
sound
eco
nom
ic
in
vestm
ent,
ii
)
exch
a
ngin
g
pr
oducts
,
iii
)
sat
isfyi
ng
pe
op
le
li
vi
ng
in
non
-
Musli
m
C
ountries.
I
n
IFB
,
fina
ncial
and
banki
ng
in
sti
tuti
on
s
do
not
m
ake
us
ur
y
and
inte
r
est
-
ba
sed
transacti
ons w
i
th
their
c
us
tom
ers
bu
t
rath
er
t
o
tra
ns
act
io
ns
that
le
ad
to
be
ne
fits
by
ta
king into
acc
ount
t
he
ris
k
to
resu
lt
in
som
e
finan
ci
al
losses
[10]
.
I
n
fa
ct
,
th
e
com
plian
ce
of
IF
B
re
gula
ti
o
ns
m
ay
v
ary
from
on
e
ban
k
t
o
ano
t
her.
T
his
us
ua
ll
y
le
ads
custom
ers
to
as
k
s
om
e
sp
eci
fic
co
un
ci
ls
(
e
.g.,
Alger
ia
n
Fat
a
wa
H
ouse
)
ab
out
the
conf
or
m
ity of
t
heir pr
oducts
a
nd servic
es
with
IF
B
pr
i
nciple
s.
2.3.
C
as
e
bas
ed rea
soni
ng
Ca
se
base
d
re
aso
ning
(CBR
)
is
an
arti
fici
al
intel
li
gen
ce
par
a
dig
m
that
so
lves
new
pro
blem
s
by
retrievin
g
st
ored
rec
ords
of
pr
i
or
pro
blem
-
so
lvi
ng
e
pis
odes
(cases
)
an
d
ada
pting
t
heir
so
luti
ons
t
o
f
it
new
ci
rcu
m
sta
nces.
P
ro
ce
ssin
g
e
a
ch
epis
od
e
pro
vid
es
a
new
ca
se
that
is
store
d
f
or
fu
t
ur
e
re
us
e,
m
aking
le
arn
i
ng
a
natu
ral
side
-
e
f
fect
of
the
rea
so
ni
ng
proc
ess
[11]
.
T
he
CB
R
can
be
ve
ry
adv
a
nta
geous
for
a
decisi
on
m
aker
wh
ic
h
knows
a
la
rg
e
num
ber
of
cases
a
nd
is
able
to
ind
e
x
them
so
that
the
m
os
t
relevant
on
es
com
e
to
m
ind
wh
e
n nee
ded
[
4]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
2
,
A
ugust
20
21
:
980
-
992
982
2.4.
On
t
ology
Ont
ologies
pla
y
an
im
po
rtant
ro
le
for
m
any
knowle
dge
-
i
ntensiv
e
a
ppli
cat
ion
s
[12]
.
T
he
y
prov
i
de
a
nu
m
ber
of
us
e
fu
l
featu
res
f
or
intel
li
gen
t
sys
tem
s,
as
well
a
s
f
or
knowle
dg
e
represe
ntati
on
in
ge
ner
al
a
nd
f
or
knowle
dge
e
nginee
rin
g
pr
oc
esses
[
13]
.
Cu
rr
e
ntly
,
ontolo
gies
are
us
e
d
to
m
ake
infor
m
at
ion
ex
plici
t
and
sh
ara
b
le
.
I
n
ad
diti
on
,
t
o
furth
er
i
m
pr
ove
dat
a
and
in
form
ation
m
anag
em
e
nt,
one
w
ou
l
d
create
an
ontol
og
y
-
base
d
ap
plica
ti
on
that
proce
sses
al
l
kn
owle
dg
e
r
el
at
ed
to
a
giv
e
n
topi
c.
Su
ch
a
pr
ocess
w
ould
involv
e
knowle
dge
m
od
el
ing t
hro
ugh
an on
t
ology
[
14]
.
3.
RELATE
D
W
O
RKS
The
li
te
ratur
e
rev
ie
w
on
studi
es
inv
ol
ving
assist
ance
to
Fa
tawa
ge
ner
at
i
on
syst
e
m
s
sh
owed
us
that
there
is
only
a
ver
y
sm
all
nu
m
ber
of
resear
ch
-
w
orks.
T
he
authors
[
5]
ha
ve
pro
posed
a
n
intel
li
gen
t
F
ata
w
a
Qu
est
i
on
An
s
wer
i
ng
syst
em
us
in
g
SOPH
I
A
-
te
xtu
al
case
-
base
d
reas
oni
ng
(TCBR
)
w
hi
ch
was
fir
stl
y
us
e
d
in
Patt
erson
’s
w
ork
[
15
]
.
T
his
is
a
su
bf
ie
ld
of
case
-
base
d
reas
on
i
ng
(
CB
R)
in
wh
ic
h
knowle
dge
so
urces
(expertise)
a
re
avail
able
in
te
xtu
al
f
or
m
at
.
It
is
based
on
the
conditi
onal
pr
oba
bili
ty
di
stribu
ti
ons
of
te
rm
s
within
docum
ents.
T
he
li
m
i
tation
of
t
his
sys
tem
li
es
in
the
fact
that
S
OPHI
A
-
TCB
R
is
no
t
e
ndowe
d
with
a
m
echan
ism
to
identify
w
ords'
order
,
neg
at
i
on
an
d
syn
on
y
m
s
and
thu
s
it
la
cks
of
ob
j
e
ct
iveness
since
these
el
e
m
ents p
la
y a key
ro
le
i
n
in
te
rp
retat
i
on
of
qu
e
sti
on
s
.
The
a
uthors
[
4]
ha
ve
de
sig
ned
a
n
i
ntell
i
gen
t
to
ol
-
cal
le
d
El
Ba
ya
ne
-
t
o
gi
ve
F
ata
w
a
f
or
ne
w
sit
uations
,
by
us
in
g
t
he
F
at
awa
of
past
si
tuati
on
s
.
T
he
syst
e
m
delas
with
fat
awa
r
el
at
ed
to
t
he
f
ie
ld
of
dr
i
nk
i
ng
a
nd
s
m
ok
ing
in
Isla
m
ic
legisl
ation. I
t i
s also
bu
il
t around the CB
R par
a
dig
m
w
her
e the
knowl
edg
e is
orga
nized
int
o
cases
store
d
i
n
a
case
-
base.
T
hen,
a
n
infe
re
nc
e
pr
ocess
is
pe
rfor
m
ed
in
order
t
o
fi
nd
a
nd
reu
se
the
ap
pro
pr
ia
te
Fatwa
a
nd
i
ts
argum
entat
i
on.
Ev
en
s
o,
t
his
syst
em
retrieves
on
ly
si
m
il
ar
cases
based
on
synta
ct
ic
an
al
ysi
s and th
us i
t do
e
s
no
t c
onsider
the sem
antic
asp
ect
.
The
ai
m
targ
et
ed by t
his
pa
pe
r
is t
w
ofol
d:
processi
ng
Fat
awa
i
n
IF
B
fiel
d
to
assist
the
Ex
per
t
M
uf
ti
i
n
his task
an
d
us
in
g
a
n
e
w
se
m
antic
si
m
i
la
rity
a
ssessm
ent
m
echan
ism
b
ased
on a
do
m
ai
n
ontolo
gy
for I
FB.
Th
us
,
we
de
ve
lop
a
n
a
ut
om
atic
QA
syst
em
t
o
deal
wit
h
F
at
awa
requests
by
m
eans
of
a
n
intel
li
ge
nt
case
-
ba
sed
sys
tem
that
can
stor
e
previ
ous
fataw
a
and
retr
ie
ve
them
as
need
ed
.
For
this
purpose,
the
c
ase
-
base
d
reas
onin
g
(CBR
)
par
a
di
gm
is
inv
ol
ve
d
by
us
in
g
sim
il
arit
y
m
easur
e
m
ents
to
identi
fy
the
cases
th
at
are
m
or
e
relevan
t.
And
as
sy
nta
x
-
base
d
m
easur
em
ents
hav
e
pro
ven
–
in
m
any
cases
–
un
su
ccess
fu
l
bec
ause
of
sem
a
ti
c
issues
,
we
buil
d
a
n
ontolo
gy
to
s
upport
the
sem
antic
si
m
il
arit
y
m
easur
em
ent.
It
shou
l
d
be
no
te
d
that
our
proposal
is
diff
e
re
nt
from
fatawa
we
b
si
te
s
for
w
hich
use
rs
ca
n
ei
ther
ask
qu
est
io
ns
and
wait
f
or
ge
tt
ing
answers
or sea
rch m
anu
al
ly
f
or
fatw
a
from
b
oo
klet t
hro
ugh keyw
ords.
4.
SY
STE
M CO
MPO
NENT
S
4.1.
F
atawa q
uestion
an
sw
e
ri
ng
s
ys
t
em
ar
chitecture
Islam
ic
Fatwa
is
wh
at
t
he
Isla
m
ic
ru
le
s
jud
ge
ab
ou
t
a
n
iss
ue
.
Wh
e
n
ge
ne
r
at
ing
a
fatw
a
,
Muftis
m
us
t
hav
e
the
do
c
um
ent
of
pr
i
or
f
ata
w
a
an
d
al
l
updates
giv
e
n
by
ex
per
t
sc
ho
la
rs.
T
his
doc
um
ent
keep
s
hu
ndre
ds
of
m
il
l
ion
s
of
su
c
h
Islam
ic
Fatwa
,
searc
hing
the
rig
ht
fat
wa
m
ay
ta
kes
days.
On
the
oth
er
ha
nd,
wh
e
n,
Im
a
m
(Muf
ti
N
o
vice)
hav
e
n’
t
the
do
cum
ent
of
fatw
a
.
T
hen,
they
c
ou
l
d
w
ron
gly
try
to
giv
e
fatw
a
.
I
n
pr
inci
ple
,
the
fatw
a
m
us
t
be
giv
e
n
by
J
ur
is
pr
ece
de
nt
(e
xpert
sch
olars
).
These
reasons
m
ot
ivate
us
to
create
a
know
le
dg
e
-
base
d
syst
em
f
or assist
ing M
uf
t
i.
The
arc
hitec
tur
e
of
the
syst
e
m
sh
own
in
Fi
gure
1
i
nclu
de
s
three
com
ponen
ts
(
case
m
e
m
or
y,
Isla
m
ic
fina
nce
dom
ain
ontol
og
y,
s
yst
e
m
interfac
es
)
an
d
fi
ve
m
od
ules
(
in
de
xing
m
od
ule,
extracti
ng
m
odule,
adap
ta
ti
on
m
od
ule,
validat
io
n
m
od
ule
and
stora
ge
m
od
ule
)
.
The
syst
em
ta
kes
the
desc
riptio
n
of
the
s
it
uation
as in
pu
t a
nd
prov
i
des
a
n
a
nsw
er
with arg
um
e
nts as o
utput.
The
pro
po
se
d
syst
e
m
is
inten
de
d
for
bo
t
h
exp
e
rt
an
d
no
vice
(lear
ner)
Mufti
s
.
It
sa
ve
s
fatawa
i
n
a
case
-
ba
se
m
e
mo
ry
(
i
.e.,
al
l pr
i
or
F
at
awa
)
f
or
furthe
r
re
us
e
.
The
e
xp
e
rt
Mu
fti
can
a
dd,
up
date
or d
el
et
e
fataw
a
via
a
s
pecific
i
nterf
ace
.
In
ad
diti
on
,
both
e
xpert
a
nd
le
ar
ne
r
Mu
fti
can
r
e
trie
ve
fat
awa
by
inter
r
og
at
i
ng
the
syst
e
m
via
anot
her
inter
face
t
hat
al
lows
run
ning
queries
on
the
s
pecific
f
ie
lds.
The
s
yst
e
m
analy
zes
qu
er
ie
s
synta
ct
ic
al
ly
a
nd
sem
antic
al
l
y.
The
f
irst
ve
r
sion o
f
this
app
ro
ac
he was
pro
po
s
ed
on [1
6].
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An o
nto
lo
gy dr
iv
en
quest
io
n a
ns
we
rin
g
sy
ste
m
for
fat
awa
r
et
rie
val
(
Ke
lt
ou
m
Be
nlaha
rc
he
)
983
Figure
1
.
Fata
wa
syst
em
a
rch
it
ect
ur
e
4.2.
D
om
ain
o
nt
ol
ogy
The
ontolo
gy
reu
se
is
hi
ghly
rec
omm
end
ed
as
it
hel
ps
r
e
searche
r
s
a
nd
dev
el
oppe
rs
to
save
ti
m
e,
effor
ts
a
nd
c
ost
s
[17]
.
In
t
he
co
nte
xt
of
IF
B
fiel
d,
ve
ry
f
e
w
ontol
og
ie
s
a
lready
exist
;
th
e
m
os
t
relevent
are:
Sem
antic
-
based
Mult
il
ing
ual
Isla
m
ic
Finan
ce
Thesaur
us
[
18
]
a
nd
I
-
F
I
K
R
O
nt
ol
o
gy
.
T
he
f
or
m
e
r
w
a
s
us
e
d
e
a
r
l
i
e
r
(
i
.
e
.
,
i
n
pa
s
t
t
im
e
)
i
n
t
he
M
I
M
O
S
w
e
b
s
i
t
e:
ht
t
p:
/
/
w
w
w
.
m
im
os
.m
y/
I
F
T
#
a
nd
i
t
i
s
c
ur
r
e
nt
l
y
no
t
a
va
i
l
a
bl
e
f
or
r
e
us
e
.
T
he
l
a
tt
e
r
i
s
no
t
a
va
i
l
a
bl
e
a
s
a
n
op
e
n
da
t
a
s
i
nc
e
i
t
i
s
s
ti
ll
un
de
r
c
on
s
t
r
uc
t
i
on
a
n
d
t
he
r
e
f
o
r
e
no
t
r
e
a
dy
f
or
r
e
us
e
.
A
c
c
or
di
n
gl
y
,
w
e
b
ui
l
t
t
he
I
s
l
am
i
c
F
i
na
nc
e
a
nd
B
a
nk
i
ng
O
nt
ol
og
y
(
I
F
B
O
)
(
s
e
e
F
i
gu
r
e
2
)
,
a
do
m
a
i
n
r
e
f
e
r
e
nc
e
o
nt
o
l
og
y
f
or
I
F
B
,
by
f
ollow
i
ng
the
desig
n
pro
cess
sp
eci
fied
by
Ne
on
m
eth
od
ology
[
19
]
.
This
on
t
ology
pro
vi
des
a
c
onsens
ual
sens
e
f
or
each
word
w
hi
ch
hel
ps
in
re
trie
ving
an
d
le
arn
i
ng
from
existi
ng
cases
in
orde
r
to
en
dow
the
base
with
ne
w
cases.
I
t
is
c
r
e
a
t
e
d
f
or
f
r
e
e
r
e
us
e
by
m
a
ki
ng
i
t
a
va
i
l
a
bl
e
to
us
e
r
s
un
de
r
t
he
f
ol
l
ow
i
ng
l
i
nk
:
https:/
/ww
w.re
search
gate.
net/publi
cat
ion
/
32345489
5_IF
B
O_Islam
ic
_F
inance_
a
nd_Ba
nkin
g_O
nto
lo
gy
F
o
r
m
or
e
de
t
a
il
s
a
bo
ut
I
F
B
O
o
nt
ol
og
y
a
s
w
e
l
l
a
s
t
he
de
v
e
l
op
m
e
nt
pr
oc
e
s
s
,
t
he
r
e
a
de
r
s
ho
ul
d
r
e
f
e
r
t
o
[
20
]
.
Figure
2
.
I
FB
O
o
nto
lo
gy
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
2
,
A
ugust
20
21
:
980
-
992
984
4.3.
C
as
e
d
esc
ri
pt
ion
A
case
is
a
co
nt
extuali
zed
pie
ce
of
kn
ow
le
dge
re
prese
ntin
g
a
n
e
xp
e
rienc
e.
It
c
on
ta
i
ns
t
he
past
le
sson
ou
tl
inin
g
t
he
c
on
te
xt
of
t
he
c
ase
an
d
t
he
c
onte
xt
in
w
hic
h
the
le
sson
c
an
be
us
e
d.
A
cas
e
can
be
an
acc
ount
of
an
e
ven
t,
a
st
ory
,
or
a
rec
ord
ty
pical
ly
co
m
pr
isi
ng
[
21
]
.
R
egardin
g
our
QA
syst
em
,
the
case
c
on
ta
in
s
two
par
ts a
s s
how
n i
n
F
ig
ure
3
.
Figure
3
.
Case
descr
i
ption
The pr
oble
m
d
escripti
on
par
t
(the q
uer
y)
tha
t con
ta
in
s:
Acti
on
: <B
uy, Sal
e, Off
e
r, Re
nt,
L
oa
n,
....>.
Be
nef
it
N
am
e:
<Car,
Ap
a
rtm
e
nt, Hom
e, …>.
Con
tract
: <M
uraba
ha
h,
M
ud
a
rab
a
h, Leasi
ng,
I
sti
sn
a
’…
>,
Ba
nk Type: <
Conve
ntion
al
,
Islam
ic
, W
in
dow
in
c
onve
ntion
al
,
>.
The
s
olu
ti
on
de
scriptio
n part
(the
a
nswer
).
4.4.
C
as
e
b
as
e
Ca
se
-
b
ase
is
use
d
to
sto
re
previo
us
an
d
ne
w
le
arn
e
d
cas
es.
As
re
porte
d
in
[
22
]
,
t
he
r
e
are
sever
al
m
et
ho
ds f
or
m
e
m
or
y
orga
niz
at
ion
:
The fla
t o
rg
a
ni
zat
ion
(a f
la
t m
e
m
or
y):
cases
are
se
qu
e
ntial
ly
stor
e
d
on
e
af
te
r
the
oth
e
r
i
n a si
m
ple li
st, a
n
arr
ay
, o
r
a
file
.
Sh
a
red featu
re
Netw
ork: this
r
egro
up
s
cases
pr
ese
ntin
g
sim
i
la
riti
es in one
c
luster.
Hiera
rc
hies are
for
m
ed
wh
e
n
cl
us
te
rs
are s
ub
div
i
ded
into
unde
r
-
cl
ust
ers.
T
he
m
et
ho
ds
use
d i
n
re
gro
upin
g
cl
us
te
r
ing
a
re
t
he
on
e
s
us
e
d
in
m
achine learn
i
ng.
Discrim
inati
on
N
et
w
orks:
t
he reg
rou
ping
cl
ust
ering
m
ade in
s
har
e
d feat
ur
e n
et
w
orks
leads t
o
a
discrim
inati
on
.
In this
orga
niz
at
ion
netw
orks
,
eve
ry
inter
nal
nod
e
is a
ques
ti
on
that
disc
rim
inate
s
accor
ding t
o
th
e ans
wer cases
con
ta
ine
d
i
n
t
he
b
ase
. T
he
m
os
t im
po
rtant
qu
est
ion
s a
re
put
in f
ir
st.
In
our
QA
sys
tem
,
cases
are
store
d
in
a
flat
case
m
e
m
or
y
giv
en
t
hat
th
e
current
nu
m
ber
of
ca
ses
le
ads
t
o
bet
te
r
perform
ances.
4.5.
In
ference
c
ycle
Re
ason
i
ng
m
od
ules
are
the
op
erati
onal
pa
rt
of
the
syst
e
m
:
retrieval
m
od
ule,
ada
ptati
on
m
od
ul
e,
evaluati
on
a
nd
le
arn
in
g
m
odule
.
4.5.1.
Re
triev
al
m
od
ule
Wh
e
n,
a
ne
w
pro
blem
(q
uer
y)
is
po
sed
a
nd
i
s
descr
i
bed
as
the
pro
blem
par
t
of
a
new
case
.
The
sem
antic
si
m
il
arity
m
ea
su
rem
ent
is
co
m
pu
te
d
betwee
n
new
case
pro
blem
descr
ipto
rs
a
nd
ol
d
case
s
pro
blem
descr
ipto
rs
t
o
identify
the
m
os
t
sim
il
ar
cas
es
to
the
ne
w
pro
blem
f
ro
m
the
case
m
e
m
o
ry
(case
base
).
The
sem
antic
si
m
il
arity
m
easur
em
ent
assum
ed
by
t
he
Applic
at
ion
On
t
ology
buil
t
befor
e
.
It
was
i
nteg
rated
to
the
CB
R
cy
cl
e
for
i
m
pr
ovin
g
thei
r
re
su
lt
s
of cas
es r
et
rie
val b
ec
ause t
he
synta
c
ti
c
m
easur
em
e
nt
oft
en
g
i
ve fa
il
. F
or
e
xam
ple
:
Pr
oble
m
d
escri
ption =
<
Act
i
on'
, b
ene
fit
',
tra
nsa
cti
on'
, B
ank
Ty
pe'
>.
In
this
case
,
synta
ct
ic
m
easur
e
m
ent
giv
e
fai
l.
Thus,
the
sy
stem
sh
ou
l
d
be
able
to
sel
ect
the
case
or
s
et
of
cases
with
great
er
si
m
il
arity
to
reuse
the
so
luti
on
after
possible
adap
ta
ti
on.
T
hat
is
wh
y,
t
he
app
li
cat
ion
ontolo
gy
is
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An o
nto
lo
gy dr
iv
en
quest
io
n a
ns
we
rin
g
sy
ste
m
for
fat
awa
r
et
rie
val
(
Ke
lt
ou
m
Be
nlaha
rc
he
)
985
need
e
d
to
asse
s
sim
i
la
rity
between
the
case
pro
blem
descr
ipto
rs.
T
he
pro
po
s
ed
a
ppro
ac
h
is
d
et
ai
le
d
in
[16]
an
d
su
m
m
arized as foll
ow
:
Syntac
ti
c m
eas
ur
em
ent:
if
∃
problem description=<action, benefit, transaction, Bank Type>
then < Solution>
else <Fail>;
Sem
antic
m
eas
ur
em
ent:
if
∃
problem description=<Action', benefit', transaction', Bank Type
> And
then< Solution>
else <fail>;
Si
m
il
arity att
ri
bu
te
s:
Sim(D,D') =
4.5.2.
A
verage
sema
nt
ic
me
asure
men
t
:
The
A
ver
a
ge
Si
m
il
arity
m
ea
su
rem
ent
(A
-
S
i
m
)
between
ne
w
case
pro
ble
m
descr
ipto
rs
(N
D
P)
a
nd
ol
d
cases p
r
ob
le
m
descr
i
ptors
(ODP), in
the
propo
s
ed
syst
em
, i
s g
i
ven b
y:
A
-
Sim (NPD,
OPD)
=
∑
(
,
′
)
=
1
;
wh
e
re
j
ref
e
rs
to
a
pro
blem
d
escripto
rs
at
trib
ute.
n refe
rs
to
a
num
ber
o
f desc
riptors att
rib
ute.
4.5.3.
Adap
tat
ion
m
od
ule
The
resu
lt
of
r
et
rievin
g
m
odule
is
a
set
of
cases
with
gre
at
er
sim
il
arit
y.
The
n,
t
he
ada
ptati
on
m
odule
reu
se
s
the
so
lu
ti
on
of
old
cas
es
retrieve
d
after
possible
ada
ptati
on
.
T
his
adap
ta
ti
on
us
es
a
set
of
base
r
ules,
th
e
app
li
cat
io
n on
t
ology, a
nd is
pe
rfor
m
e
d
by t
he
ad
a
ptati
on A
l
gorithm
.
Rule B
ase
The rule
base
c
on
ta
in
s
4
pr
inc
ipals’
ru
le
s a
s s
how
n
in
table
a
s sho
wn in
(Ta
ble
1
).
Table
1
. R
ule
b
ase
(
e
xce
r
pt)
Ru
les
Descripti
o
n
s
Co
n
d
itio
n
s
Exa
m
p
les
Ru
le 1
D inv
erse of
D'
D,D'
in
stan
ces o
f
Actio
n
(Sale,
Bay
)
Ru
le 2
D su
b
class d
irect
D' =
H
alal Ben
ef
it
/
D'= Ha
ra
m
Ben
ef
it/
D'= Co
n
tracts
(Reality
,
H
alal Be
n
ef
it)/
(Alcoh
o
l,
Har
a
m
Ben
ef
it)/
(M
u
sh
arakah
,Co
n
tracts)
Ru
le 3
D su
b
class in
d
irec
t of
D'
D' =
H
alal Ben
ef
it
/
D'= Ha
ra
m
Ben
ef
it/
D'= Co
n
tracts
(Co
m
m
e
rcial
Reali
ty
,
Halal B
en
ef
it)/
(Gelatine,
Ha
ra
m
Ben
ef
it)/
(Bay
'
Bi T
h
a
m
an
'
Ajil,
Co
n
tracts)
Ru
le 4
D ins
tan
ce of
D'
-
(T
er
rain,
Halal B
e
n
ef
it)
Ru
le 5
Si
m
(D,
D')=1
-
(Car,
Auto
m
o
b
ile)
Ad
a
ptati
on
a
lg
or
it
hm
The
ad
aptat
io
n
Algorithm
is
t
he
co
re
of
ada
ptati
on
m
od
ule
,
it
con
sist
s
of
set
of
cases
to
be
exec
uted
i
n
dep
e
ndent
with
the
pro
blem
d
escripto
rs
a
nd
base
d on the a
pp
li
cat
io
n on
t
ol
og
y.
Sim(Action, Action')=1;
Sim(benefit, benefit')=1;
Sim(transaction, transaction')=1;
Sim (Bank Type, Bank Type')=1
1
0
if (D same as D') or (D equivalent to D');
// D value of the new problem case
// D' value of retrieved problem case.
otherwise;
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
2
,
A
ugust
20
21
:
980
-
992
986
begin
if [case retrieved] then
<
Extracted Answer>
else
case 1: D inverse of D' do D=D';
case 2: D sub class direct of D' do D=D';
case 3: D sub class indirect of D' do D=D';
case 4: D instance of D' do D=D';
case 5: Sim(D,D')=1 do D=D';
.....
.....
otherwise: <send the query to expert Mufti>;
end;
Fo
r
e
xam
ple:
User ask
ed
the
qu
e
sti
on
:
w
hat
is t
he
r
ulin
g on buyi
ng a tr
uc
k by Mu
ra
bah
a
thro
ugh
a
n Isla
m
ic
b
ank
?
New p
roblem
d
escri
ption = <
Buy, T
ru
c
k, Mura
ba
ha,
Islam
ic
Bank
>;
Thr
ee
Hy
po
t
he
ses m
ay
su
gg
e
st
, as fo
ll
ow:
Hyp
oth
esi
s
o
ne
:
suppose t
hat
exists,
Old p
roblem
d
escripti
on
= <
Bu
y,
Tru
ck,
Mu
rab
aha
h,
Isl
ami
c B
ank
>;
The
n by sem
antic
m
easur
em
e
nt: M
ur
a
ba
ha
i
s equivale
nt to
Mura
bah
a
h,
Sim(Buy,Buy)=1;
Sim(Truck,Truck)=1;
Sim(Murabaha, Murabahah)=1;
Sim(Islamic Bank,Islamic Bank)=1;
A
-
Sim=
(1+1+1+1)/4=1
New p
roblem
d
escri
ption
=
<
B
u
y, Truck, Murabahah, Islamic Bank>;
then <Extract the exact Answer>;
Hyp
oth
esi
s
t
w
o:
suppose
th
at
,
after
com
pu
ti
ng
sem
antic
si
m
i
la
rity
m
e
asur
em
ent
between
t
he
bo
t
h
descr
i
ptors,
th
e
aver
a
ge
sim
i
la
rity
is
between
0
a
nd
1,
(
0
<
A
-
Sim
<1
).
In
t
his
sit
uation,
t
he
adap
ta
ti
on
m
od
ule
searc
h i
n
the
r
ule b
a
se
for
t
he
a
ppr
opriat
e ru
le
.
Fr
om
the
on
t
olo
gy:
Buy is inverse of Sale, then Rule 1, executed ;
Truck is instance of Halal benefit, then Rule 4, executed ;
Murabahah
is sub class direct of
Contracts, then Rule 2, executed ;
Sim(Islamic Bank, Islamic Bank)=1, then Rule 5, executed;
New problem description= <Sale, Truck, Murabahah, Islamic Bank>;
then <A
-
Sim Recalculated>;
If A
-
Sim =1 then
<Extract Answer>;
<
New case learned>;
else
<Extract set of similar cases>; // will show as similar cases
<Se
nd
the
que
ry
wi
th
a
set
of
sim
ila
r
cas
es
to
be
eva
lu
ate
d>;
//
by
the expert Mufti
Hyp
oth
esi
s
t
hr
ee:
su
pp
os
e
th
at
,
after
com
pu
ti
ng
s
em
antic
si
m
il
ari
ty
m
e
asur
em
ent
no
l
ikeness
betwe
e
n
the
both
desc
r
iptor
s;
al
l
dim
ensio
ns
giv
e
z
ero
(
A
-
Sim
=
0
);
In
t
his
sit
ua
ti
on
,
t
he
ada
ptati
on
m
odule
sen
d
this
qu
e
ry
to
ex
per
t
Mufti.
For
givi
ng
ne
w
a
ns
w
e
rs
(m
anu
al
le
ar
ning)
.
And,
show
to
t
he
us
er
a
m
essage:
"Sorry
the
re
is n
o
a
ns
we
r for yo
ur que
ry, y
our req
uest
wa
s sen
t t
o
the
expert M
uf
ti
. Pl
ease, try a
gain a
fter
few days".
4.5.4.
Le
arnin
g and ev
alu
at
i
on
m
odul
e
The
E
xpert
-
M
uf
ti
e
nr
ic
he
s
th
e
case
ba
se
a
nd
the
ap
plica
ti
on
ontol
og
y
w
hen
re
quired
.
The
E
xpert
-
M
uf
ti
has
a
spe
ci
fic
interface
t
o
ad
d,
m
od
ify
and
de
le
ti
ng
cases.
His
r
ole
is
to
evaluate
the
pro
po
s
ed
a
nswers
for
cases
hav
i
ng
b
et
wee
n
(
0an
d
1
)
.
He
m
a
y
validate
the
m
or
gen
e
rate
new
ones
t
o
m
ake
the
sys
tem
le
arn
.
The
autom
at
ic
le
arn
is
reali
zed
on
ly
wh
e
n
t
he
Avera
ge Si
m
il
arit
y
(A
-
Si
m=
1
).
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An o
nto
lo
gy dr
iv
en
quest
io
n a
ns
we
rin
g
sy
ste
m
for
fat
awa
r
et
rie
val
(
Ke
lt
ou
m
Be
nlaha
rc
he
)
987
5.
RESU
LT
S
:
I
LL
US
TR
ATI
VE E
X
A
M
PL
E
To
validat
e
thi
s
work,
we
ha
ve
im
ple
m
ented
a
prototyp
e
of
the
pro
po
se
d
Q
A
syst
em
.
It
con
sist
s
in
a
m
ul
ti
la
ng
ua
ge
app
li
cat
io
n
de
velo
ped
us
i
ng
the
Java
la
ngua
ge
.
In
w
hat
f
ollows,
sc
reensho
ts
a
re
il
lust
rated
in
bot
h
En
glish
a
nd
A
rab
ic
la
ngua
ge
s.
T
hree
us
e
rs'
cl
asses
are
co
ns
ide
red
:
sim
ple
us
e
r
(
ل
ئاس
)
(
is
sim
ple
us
er
a
no
vice
or
le
arn
er
I
m
a
m
)
, exper
t M
uf
ti
(
ر
يب
خ
)
an
d
a
dm
i
nistrato
r
(
ر
يدم
);
Fig
ur
e
4
sho
w
s the m
ai
n
pa
ne
l.
Figure
4. The
m
ai
n
m
enu
It
shou
l
d
be
no
te
d
that
the
cu
r
ren
t
ver
si
on
of
the
prot
otype
do
e
s
not
al
low
us
ers
t
o
intera
ct
with
the
syst
e
m
thro
ug
h
qu
est
io
ns
f
orm
ulate
d
in
na
tural
la
ng
uages.
T
his
li
es
to
the
fact
tha
t
the
us
e
of
na
tural
la
nguag
e
s
m
ay
po
se
se
ve
ral
pro
blem
s
relat
e
d
to
na
tural
la
ngua
ge
proc
es
sing
(N
L
P).
Con
s
eq
ue
ntly
,
we
hav
e
lim
it
ed
oursel
ves
by
im
po
sing
so
m
e
restri
ct
ion
s
on
the
way
the
quest
ion
s
are
f
or
m
ulate
d
since
the
m
a
in
obj
ect
ive
h
e
re
is o
nly t
o
c
hec
k
the
ef
fecti
ve
ness of t
he
app
ro
ac
h
.
User spa
ce
,
t
he
us
e
r
s
pace all
ows sim
ple u
ser
s to form
ulate
t
heir q
uestions
by spec
ify
ing (
see Fig
ur
e
5):
a)
The kin
d of q
ue
sti
on
w
hich
c
an be: (
What
is
the
j
ud
gm
ent
?
Wh
at
is the
pu
nish
m
ent
?
):
(
؟د
ح
ام ؟
مكح
ام
(
b)
T
he
act
io
n
(Buy
, Rent, Sel
l,
Mortage
)
: (
ن
هر,
عيب ,ءارك
,ء
ا
رش
)
.
c)
T
he be
nef
it
)
Com
m
ercial
locati
on
, H
ouse,
Ca
r
(
::
(
ة
را
يس
, لزن
م
,ل
ح
م
)
.
d)
T
he
ty
pe
of
co
ntract
(Mur
a
ba
ha,
M
ushara
ka
h,
M
udha
rab
a
h
)
: (
ة
كرا
شم
,ة
ب
راضم ,ةح
بارم
).
e)
T
he
ty
pe
of
ba
nk (
c
onve
ntio
na
l, Islam
ic
): (
,ي
ديل
قت
ي
م
سإ
)
.
Figure
5
.
Ques
ti
on
form
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
2
,
A
ugust
20
21
:
980
-
992
988
On
ce
the
quest
ion
is
for
m
ulate
d,
t
he user
as
ks
for
t
he
a
ns
w
er
by
cl
ic
king
on the
butt
on (
An
al
yz
e:
ل
ل
ح
)
.
Th
e
r
es
ul
t i
s sh
ow
n
in
Figure
6
.
Figure
6
.
Resp
on
s
e f
orm
The
syst
e
m
al
lows
the
sim
ple
us
e
r
t
o
ta
ke
a
lo
ok
at
the
sim
il
ar
cases
(S
im
il
ar
cases:
ة
هب
اشت
مل
ا
ت
اح
لا
),
as
dep
ic
te
d i
n Fi
gure
7
.
Figure
7
.
Sim
i
l
ar cases
for
m
Ad
m
inist
rator
sp
ace
The
a
dm
inist
rator
is
res
pons
i
bl
e for
set
ti
ng
-
up a
nd m
anag
in
g
the
case
base
b
e
fore
sta
rtin
g t
o use
t
he
QA syst
em
. A
t any ti
m
e, h
e c
an br
owse the
case base
(see
Figure
8
).
Figure
8
.
Adm
i
nistrato
r
m
ai
n
m
enu
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An o
nto
lo
gy dr
iv
en
quest
io
n a
ns
we
rin
g
sy
ste
m
for
fat
awa
r
et
rie
val
(
Ke
lt
ou
m
Be
nlaha
rc
he
)
989
Ex
per
t M
uf
ti
s
pace
T
h
e
e
x
p
e
r
t
M
u
f
t
i
e
n
s
u
r
e
t
h
r
e
e
r
o
l
e
s
:
a
d
d
n
e
w
q
u
e
s
t
i
o
n
s
t
o
e
n
r
i
c
h
t
h
e
l
i
s
t
o
f
a
l
l
q
u
e
s
t
i
o
n
s
,
i
n
t
r
o
d
u
c
e
a
n
s
w
e
r
s
t
o
p
r
e
v
i
o
u
s
l
y
a
s
k
e
d
q
u
e
s
t
i
o
n
s
a
n
d
f
i
n
a
l
l
y
m
o
di
f
y
a
r
e
s
p
o
n
s
e
i
n
t
h
e
c
a
s
e
b
a
s
e
,
a
s
g
i
v
e
n
o
n
F
i
g
u
r
e
9
.
D
e
p
e
n
d
i
n
g
o
n
t
h
e
s
e
l
e
c
t
e
d
o
p
t
i
o
n
,
t
h
e
s
y
s
t
e
m
d
i
s
p
l
a
y
s
t
h
e
c
o
r
r
e
s
p
o
n
d
i
n
g
s
c
r
e
e
n
a
s
s
h
o
w
n
i
n
F
i
g
u
r
e
s
1
0
a
n
d
1
1
.
Figure
9
.
Expe
rt
m
uf
ti
m
ai
n
m
enu
Figure
10
. Int
r
oducin
g res
ponse
s
Figure
11
. A
dju
sti
ng r
e
spo
nse
6.
AS
SES
SME
N
T OF THE
W
ORK
Finall
y
, w
e t
he
or
et
ic
al
ly
asses
s
an
d
c
om
par
e
our
c
ontrib
utio
n wit
h othe
r
rel
at
ed
w
orks
, as
giv
e
n on
Table
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.