T
E
L
K
O
M
N
I
K
A
T
elec
o
m
m
un
ica
t
io
n,
Co
m
pu
t
ing
,
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
19
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
1
,
p
p
.
8
9
~9
5
I
SS
N:
1
6
9
3
-
6
9
3
0
,
ac
cr
ed
ited
First Gr
ad
e
b
y
Kem
en
r
is
tek
d
i
k
ti,
Dec
r
ee
No
: 2
1
/E/KPT
/2
0
1
8
DOI
: 1
0
.
1
2
9
2
8
/TE
L
KOM
NI
K
A.
v
1
9
i1
.
1
5
7
6
6
89
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
Ext
ra
ct
iv
e multi
do
cument sum
ma
riza
tion usin
g
har
mo
ny
sea
rch alg
o
rithm
Z
uh
a
ir
H
u
s
s
ein A
li
1
,
Ahm
ed
K
a
wt
her
H
us
s
ein
2
,
H
a
it
hem
K
a
re
em
Aba
s
s
3
,
E
lha
m
F
a
d
el
4
1,
2,
4
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Co
ll
e
g
e
o
f
E
d
u
c
a
ti
o
n
,
M
u
sta
n
siriy
a
h
Un
i
v
e
rsity
,
Ba
g
h
d
a
d
,
Ira
q
3
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter E
n
g
in
e
e
rin
g
,
Al
-
M
a
n
so
u
r
Un
i
v
e
rsity
Co
l
leg
e
,
Ba
g
h
d
a
d
,
Ira
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
7
,
2
0
2
0
R
ev
is
ed
Mar
22
,
2
0
2
0
Acc
ep
ted
Sep
16
,
2
0
2
0
Th
e
e
x
p
o
n
e
n
ti
a
l
g
ro
wt
h
o
f
i
n
fo
r
m
a
ti
o
n
o
n
th
e
i
n
tern
e
t
m
a
k
e
s
it
t
ro
u
b
les
o
m
e
fo
r
u
se
rs
t
o
g
e
t
v
a
lu
a
b
le
in
f
o
rm
a
ti
o
n
.
Tex
t
su
m
m
a
riza
ti
o
n
is
th
e
p
ro
c
e
ss
to
o
v
e
rc
o
m
e
su
c
h
a
p
r
o
b
lem
.
An
a
d
e
q
u
a
te
su
m
m
a
ry
m
u
st
h
a
v
e
wid
e
c
o
v
e
ra
g
e
,
h
ig
h
d
iv
e
rsit
y
,
a
n
d
h
i
g
h
re
a
d
a
b
i
li
ty
.
I
n
th
is
a
rti
c
le,
a
n
e
w
m
e
th
o
d
fo
r
m
u
l
ti
-
d
o
c
u
m
e
n
t
s
u
m
m
a
riza
ti
o
n
h
a
s
b
e
e
n
su
p
p
o
se
d
b
a
se
d
o
n
a
h
a
rm
o
n
y
se
a
rc
h
a
lg
o
rit
h
m
th
a
t
o
p
ti
m
ize
s
th
e
c
o
v
e
ra
g
e
,
d
iv
e
rsit
y
,
a
n
d
re
a
d
a
b
il
it
y
.
C
o
n
c
e
rn
i
n
g
th
e
b
e
n
c
h
m
a
rk
d
a
ta
se
t
Tex
t
An
a
l
y
sis
Co
n
fe
re
n
c
e
(TAC
-
2
0
1
1
)
,
th
e
ROU
G
E
p
a
c
k
a
g
e
u
se
d
to
m
e
a
su
re
th
e
e
ffe
c
ti
v
e
n
e
ss
o
f
t
h
e
p
ro
p
o
se
d
m
o
d
e
l.
Th
e
c
a
lcu
late
d
re
su
lt
s su
p
p
o
rt
t
h
e
e
ffe
c
ti
v
e
n
e
ss
o
f
t
h
e
p
r
o
p
o
se
d
a
p
p
ro
a
c
h
.
K
ey
w
o
r
d
s
:
C
o
v
er
ag
e
Div
er
s
ity
Har
m
o
n
y
s
ea
r
c
h
R
ea
d
ab
ilit
y
T
AC
-
2011
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
:
Z
u
h
air
Hu
s
s
ein
Ali
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
C
o
lleg
e
o
f
E
d
u
ca
tio
n
Mu
s
tan
s
ir
iy
ah
Un
iv
er
s
ity
B
ag
h
d
ad
,
I
r
a
q
E
m
ail:
zu
h
air
7
2
h
@
u
o
m
u
s
tan
s
ir
iy
ah
.
ed
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
I
n
f
o
r
m
atio
n
o
v
er
lo
ad
is
o
n
e
o
f
th
e
m
o
s
t
co
m
m
o
n
p
r
o
b
lem
s
b
ec
au
s
e
o
f
th
e
f
ast
ev
o
lu
tio
n
o
f
in
f
o
r
m
ati
o
n
in
th
e
W
o
r
ld
W
id
e
W
eb
[
1
]
.
T
ex
t
s
u
m
m
ar
izatio
n
(
T
S)
is
th
e
d
is
m
is
s
al
o
f
s
u
ch
a
p
r
o
b
lem
.
Als
o
,
T
S,
th
e
p
r
o
ce
s
s
o
f
p
r
o
d
u
cin
g
a
d
o
cu
m
en
t
s
u
m
m
ar
y
f
r
o
m
a
s
er
ies
o
f
d
o
cu
m
en
ts
o
r
o
n
e
d
o
cu
m
en
t
with
o
u
t lo
s
in
g
its
m
ain
id
ea
s
,
aim
s
to
ex
tr
ac
t
u
s
ef
u
l
in
f
o
r
m
atio
n
f
r
o
m
th
e
s
o
u
r
ce
s
to
th
e
u
s
er
s
[
2
]
.
T
h
e
s
u
m
m
ar
y
o
f
f
er
s
a
h
elp
f
u
l
g
u
id
e
to
g
en
er
ate
atten
tio
n
o
n
in
f
o
r
m
atio
n
,
to
m
ak
e
d
ec
is
io
n
s
o
n
wh
eth
er
a
d
o
cu
m
en
t
is
u
s
ef
u
l
o
r
n
o
t
an
d
to
ass
is
t
as
a
tim
e
s
av
er
f
o
r
u
s
er
s
[
3
]
.
B
ased
o
n
th
e
q
u
an
tity
o
f
th
e
d
o
cu
m
en
t
to
b
e
s
u
m
m
ar
ized
,
T
S
ca
n
b
e
class
if
ied
as
a
s
in
g
le
d
o
cu
m
en
t
s
u
m
m
ar
izatio
n
(
SDS)
o
r
m
u
lti
-
d
o
cu
m
en
t
s
u
m
m
ar
izatio
n
(
MD
S).
Fo
r
in
s
tan
ce
,
in
SDS
ju
s
t
a
s
in
g
le
d
o
cu
m
en
t
ca
n
b
e
s
u
m
m
ar
ized
in
to
s
h
o
r
ter
o
n
es,
wh
ile
in
MD
S a
s
et
o
f
r
elate
d
d
o
cu
m
en
ts
with
th
e
s
am
e
to
p
ic
is
s
u
m
m
ar
ized
in
to
o
n
e
s
u
m
m
ar
y
[
4
]
.
MD
S
is
m
o
r
e
co
m
p
licated
th
an
SDS
a
lth
o
u
g
h
s
o
m
e
s
im
ilar
tech
n
iq
u
es c
an
b
e
u
s
ed
f
o
r
b
o
th
MSD
an
d
SDS
d
u
e
to
in
f
o
r
m
atio
n
o
v
er
lo
ad
an
d
a
h
ig
h
d
eg
r
ee
o
f
r
ed
u
n
d
an
c
y
.
T
h
e
r
ed
u
n
d
an
cy
o
cc
u
r
s
b
ec
au
s
e
s
u
m
m
ar
ized
d
o
cu
m
en
ts
d
ea
l
with
s
im
ilar
to
p
ics
an
d
s
h
ar
e
th
e
s
am
e
id
ea
s
.
As
a
r
esu
lt,
r
ed
u
cin
g
r
ed
u
n
d
an
cy
ca
n
lead
to
a
h
ig
h
-
q
u
ality
s
u
m
m
ar
y
[
5
]
.
T
h
e
way
o
f
a
s
u
m
m
ar
y
cr
ea
tin
g
is
eith
er
ex
tr
ac
tio
n
o
r
an
ab
s
tr
ac
tio
n
ac
co
r
d
in
g
to
th
e
f
u
n
ctio
n
to
b
e
p
er
f
o
r
m
ed
[
6
]
.
E
x
tr
ac
tiv
e
s
u
m
m
ar
i
za
tio
n
is
a
m
ec
h
an
is
m
f
o
r
a
p
r
o
f
ess
io
n
al
ex
tr
ac
tio
n
o
f
th
e
liter
ar
y
co
m
p
o
n
en
ts
lik
e
s
en
ten
ce
s
,
p
ass
ag
es,
an
d
s
o
o
n
f
r
o
m
th
e
o
r
ig
in
al
m
ea
n
in
g
.
W
h
er
ea
s
,
ab
s
tr
ac
tiv
e
s
u
m
m
ar
izatio
n
will
d
ep
en
d
o
n
th
e
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
(
NL
P)
tech
n
iq
u
es
,
wh
ich
n
ee
d
a
co
m
p
licated
u
n
d
er
s
tan
d
in
g
o
f
NL
P
s
tr
ateg
ies
to
an
aly
ze
th
e
s
en
ten
ce
s
o
f
d
o
cu
m
en
ts
an
d
p
ar
ag
r
ap
h
s
wh
er
e
s
ev
er
al
ch
an
g
es h
av
e
to
b
e
m
ad
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
1
:
8
9
-
95
90
on
th
e
s
elec
ted
s
en
ten
ce
s
.
W
h
ile
in
th
e
ex
tr
ac
tiv
e
s
u
m
m
ar
izatio
n
,
n
o
n
ee
d
f
o
r
m
o
d
if
icatio
n
will
b
e
ap
p
lied
to
th
e
s
en
ten
c
es
th
at
ar
e
in
clu
d
ed
in
th
e
r
esu
lted
s
u
m
m
ar
y
.
T
h
er
ef
o
r
e,
ab
s
tr
ac
tiv
e
s
u
m
m
ar
izatio
n
is
tim
e
-
co
n
s
u
m
in
g
an
d
m
u
ch
d
if
f
icu
lt
th
an
ex
tr
ac
tiv
e
s
u
m
m
ar
izatio
n
[
7
]
.
Mo
r
eo
v
er
,
s
u
m
m
ar
izatio
n
ca
n
b
e
class
if
ied
as
eith
er
g
en
er
ic
s
u
m
m
ar
izatio
n
o
r
q
u
er
y
s
u
m
m
ar
izatio
n
.
Th
e
G
en
er
ic
s
u
m
m
ar
izatio
n
g
en
er
ate
s
a
s
u
m
m
ar
y
,
wh
ich
alwa
y
s
in
clu
d
es
th
e
ess
en
tial
co
n
ten
t
o
f
th
e
d
o
cu
m
en
ts
.
Ho
wev
er
,
th
e
r
estrictio
n
of
g
en
er
i
c
s
u
m
m
ar
izatio
n
is
n
ev
er
a
to
p
ic
o
r
q
u
er
y
is
av
ailab
le
f
o
r
th
e
s
u
m
m
ar
izatio
n
p
r
o
ce
d
u
r
e
.
W
h
ile
in
q
u
er
y
-
b
as
e
d
s
u
m
m
ar
izatio
n
,
a
s
u
m
m
ar
y
is
cr
ea
ted
d
ep
en
d
in
g
o
n
th
e
q
u
er
y
o
f
th
e
u
s
er
,
wh
er
e
th
e
d
o
cu
m
en
ts
ar
e
s
ea
r
ch
ed
to
b
e
m
atch
ed
with
s
u
ch
q
u
er
y
[
8
]
.
T
h
is
p
ap
er
ap
p
r
o
ac
h
s
u
p
p
o
s
es
a
n
ew
m
o
d
el
f
o
r
ex
tr
ac
tiv
e
g
en
er
ic
MD
S
b
ased
on
h
ar
m
o
n
y
s
ea
r
ch
alg
o
r
ith
m
(
HSA)
th
at
im
p
r
o
v
es
co
v
er
ag
e,
d
iv
er
s
ity
,
an
d
r
ea
d
ab
ilit
y
.
T
h
e
ex
p
er
im
en
ta
l
r
esu
lts
u
tili
ze
d
to
th
e
T
AC
-
2
0
1
1
d
ataset
an
d
R
OUGE
p
ac
k
ag
e
ap
p
lied
to
m
ea
s
u
r
e
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el.
2.
RE
L
AT
E
D
WO
RK
S
E
v
en
th
o
u
g
h
tex
t
s
u
m
m
ar
izatio
n
h
as
d
r
awn
co
n
s
id
er
atio
n
b
asically
af
ter
th
e
in
f
o
r
m
atio
n
ex
p
an
s
io
n
o
n
th
e
I
n
ter
n
et,
th
e
p
r
im
ar
y
wo
r
k
h
as
b
ee
n
d
o
n
e
in
1
9
5
8
[
9
]
.
Fro
m
th
at
y
ea
r
a
v
ar
iety
o
f
s
u
m
m
ar
izatio
n
tech
n
iq
u
es
h
av
e
b
ee
n
p
r
o
p
o
s
ed
an
d
ass
ess
ed
.
Fo
r
ex
am
p
le,
s
o
m
e
r
esear
ch
er
s
[
1
0
,
1
1
]
ap
p
lied
s
en
ten
ce
clu
s
ter
in
g
in
tex
t
s
u
m
m
ar
izatio
n
s
u
cc
ess
f
u
lly
.
T
h
e
b
asic
id
ea
b
eh
in
d
th
e
clu
s
ter
-
b
ased
ap
p
r
o
ac
h
f
o
r
MD
S
is
b
ased
o
n
s
en
ten
ce
s
with
h
ig
h
d
eg
r
ee
s
o
f
s
im
ilar
ities
th
at
ar
e
g
r
o
u
p
ed
in
to
o
n
e
clu
s
ter
,
th
en
o
n
e
s
en
ten
ce
is
s
elec
ted
f
r
o
m
ea
ch
clu
s
ter
to
b
e
in
clu
d
ed
in
th
e
g
en
er
ated
s
u
m
m
ar
y
.
Sen
ten
ce
s
elec
tio
n
d
ep
en
d
s
o
n
s
elec
tin
g
s
en
ten
ce
s
th
at
ar
e
clo
s
est
to
th
e
ce
n
tr
o
id
o
f
th
e
clu
s
ter
[
1
2
]
.
Gr
ap
h
-
b
ased
ap
p
r
o
ac
h
es,
wh
ich
ar
e
b
ased
o
n
an
ass
u
m
p
tio
n
th
at
th
e
s
en
ten
ce
im
p
o
r
tan
ce
will
in
cr
ea
s
e
if
it
h
as
m
o
r
e
s
im
ilar
ity
to
o
th
er
s
en
ten
ce
s
in
th
e
d
o
cu
m
en
t,
ar
e
also
u
s
ed
wid
ely
in
MD
S
b
y
th
e
r
esear
ch
er
.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
b
y
r
ep
r
esen
tin
g
ea
ch
s
en
ten
ce
as a
n
o
d
e
in
th
e
g
r
ap
h
an
d
th
e
co
s
in
e
s
im
ilar
ity
ca
n
b
e
u
s
ed
as a
n
ed
g
e
b
etwe
en
n
o
d
es [
1
3
]
.
T
h
e
p
ag
e
r
an
k
[
1
4
]
o
r
t
ex
t
r
an
k
[
1
5
]
is
th
en
ap
p
lied
to
s
co
r
e
th
e
s
en
ten
ce
s
,
s
en
ten
ce
s
with
h
ig
h
s
co
r
es
ar
e
in
clu
d
ed
in
th
e
f
in
al
s
u
m
m
ar
y
.
So
m
e
r
esear
ch
er
s
f
o
cu
s
o
n
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es
wh
ich
h
av
e
b
ee
n
co
m
m
o
n
ly
u
s
ed
in
th
e
f
ield
o
f
T
S.
T
h
is
ap
p
r
o
ac
h
d
ep
en
d
s
o
n
ca
teg
o
r
izin
g
th
e
s
en
ten
ce
s
in
to
two
class
es;
s
u
m
m
ar
y
s
en
ten
ce
s
o
r
n
o
n
-
s
u
m
m
ar
y
s
en
ten
ce
s
.
Su
ch
an
ap
p
r
o
ac
h
r
eq
u
ir
es
d
iv
id
in
g
t
h
e
d
ataset
in
to
tr
ain
in
g
,
test
in
g
th
at
d
ata
f
o
r
lab
elin
g
,
an
d
ca
teg
o
r
izin
g
it
ac
co
r
d
in
g
ly
.
So
m
e
o
f
th
e
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es
ar
e
Naïv
e
B
ay
es
[
1
6
]
,
n
eu
r
al
n
etwo
r
k
[
1
7
]
,
d
ec
is
io
n
tr
ee
s
[
1
8
]
an
d
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
[
1
9
]
.
Ma
n
y
r
esear
ch
er
s
h
av
e
also
in
v
esti
g
ated
o
p
tim
iz
atio
n
ap
p
r
o
ac
h
es.
Ma
n
y
o
p
tim
izatio
n
tech
n
iq
u
es
s
u
ch
as
d
if
f
er
en
tial
ev
o
lu
tio
n
(
DE
)
[
2
0
]
,
p
ar
ticle
s
war
m
o
p
tim
izatio
n
(
PS
O)
[
2
1
]
an
d
g
en
etic
alg
o
r
ith
m
(
GA)
[
2
2
]
ar
e
u
s
ed
f
o
r
T
S.
Op
tim
izatio
n
tech
n
iq
u
es
ar
e
b
ased
o
n
m
u
ltip
le
ag
en
ts
in
th
e
p
o
p
u
latio
n
th
at
s
ea
r
ch
f
o
r
ca
n
d
id
ate
s
o
lu
tio
n
s
wh
ich
ar
e
co
n
s
id
er
ed
as
p
o
in
ts
in
th
e
s
ea
r
ch
s
p
ac
e.
I
n
[
2
3
]
au
th
o
r
s
ap
p
lied
a
b
ee
co
lo
n
y
f
o
r
MD
S.
Her
e,
th
e
b
ee
s
wer
e
co
n
s
id
er
ed
as
ag
en
ts
th
at
s
ea
r
ch
f
o
r
n
ec
tar
in
f
lo
wer
s
wh
er
e
th
e
f
o
o
d
is
co
n
s
id
er
ed
as
a
ca
n
d
id
ate
s
o
lu
tio
n
,
th
er
e
is
a
s
in
g
le
b
ee
f
o
r
ev
er
y
f
o
o
d
s
o
u
r
ce
.
As a
r
esu
lt,
th
e
o
b
jectiv
e
f
u
n
ctio
n
is
f
o
r
th
e
b
ee
s
to
co
llect
a
p
o
r
tio
n
o
f
f
o
o
d
.
W
h
en
th
e
f
o
o
d
is
ab
an
d
o
n
ed
,
th
e
b
ee
th
en
tu
r
n
s
in
to
a
s
co
u
t
an
d
lo
o
k
s
f
o
r
an
o
th
er
f
o
o
d
s
o
u
r
ce
.
T
h
ey
s
ea
r
ch
f
o
r
n
eig
h
b
o
r
in
g
ar
ea
s
an
d
s
elec
t
th
e
b
est
ca
n
d
id
ate.
W
h
en
m
o
v
in
g
to
a
n
eig
h
b
o
r
,
a
s
en
ten
ce
is
d
elete
d
r
an
d
o
m
ly
f
r
o
m
th
e
p
r
esen
t
s
u
m
m
ar
y
an
d
an
o
th
er
s
en
ten
ce
is
in
clu
d
ed
s
o
th
e
len
g
th
lim
itatio
n
is
n
o
t v
io
lated
.
3.
P
RO
P
O
SE
D
F
RA
M
E
WO
R
K
I
n
th
is
p
ap
er
,
a
n
ew
ap
p
r
o
ac
h
f
o
r
MD
S
is
p
r
o
p
o
s
ed
.
I
t
is
d
ec
o
m
p
o
s
ed
o
f
f
o
u
r
m
ain
s
tep
s
.
First
th
e
p
r
ep
r
o
ce
s
s
in
g
is
d
o
n
e.
Seco
n
d
ly
,
wo
r
d
s
im
ilar
ity
m
ea
s
u
r
e
an
d
s
u
m
m
ar
y
q
u
ality
f
ac
to
r
s
ar
e
ap
p
lied
,
an
d
f
i
n
ally
,
h
ar
m
o
n
y
s
ea
r
ch
is
p
er
f
o
r
m
ed
.
T
h
ese
f
o
u
r
s
tep
s
ar
e
d
escr
ib
ed
as f
o
llo
ws
;
3
.
1
.
P
re
pro
ce
s
s
ing
T
h
er
e
ar
e
f
o
u
r
s
tep
s
f
o
r
p
r
ep
ar
in
g
th
e
d
ata,
th
ese
s
tep
s
in
clu
d
e:
-
Sen
ten
ce
s
eg
m
en
tatio
n
:
ea
ch
d
o
cu
m
e
n
t
is
d
iv
id
ed
in
d
iv
id
u
ally
in
to
s
ev
er
al
s
en
ten
ce
s
b
a
s
ed
o
n
th
e
d
o
t
b
etwe
en
th
em
.
-
T
o
k
en
izatio
n
:
th
e
p
r
o
ce
s
s
o
f
s
ep
ar
atin
g
s
en
ten
ce
s
in
to
te
r
m
s
.
-
Sto
p
wo
r
d
r
em
o
v
al:
i
n
v
o
lv
es
r
em
o
v
in
g
r
e
d
u
n
d
an
t a
n
d
r
ep
ea
t
ed
ter
m
s
i
n
th
e
d
o
cu
m
en
t
th
at
d
o
n
o
t
o
f
f
er
th
e
r
eq
u
ir
ed
in
f
o
r
m
atio
n
f
o
r
r
ec
o
g
n
izin
g
an
im
p
o
r
tan
t
s
en
s
e
o
f
t
h
e
d
o
c
u
m
en
t.
-
Stem
m
in
g
:
th
e
p
r
o
ce
s
s
o
f
g
en
er
atin
g
th
e
r
o
o
t o
f
th
e
wo
r
d
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
E
xtra
ctive
mu
lti d
o
cu
men
t su
mma
r
iz
a
tio
n
u
s
in
g
h
a
r
mo
n
y
s
ea
r
ch
a
lg
o
r
ith
m
(
Zu
h
a
ir
Hu
s
s
ein
A
li
)
91
3
.
2
.
Si
m
ila
rit
y
m
ea
s
ure
Similar
ity
m
ea
s
u
r
e
p
lay
s
a
s
ig
n
if
ican
t
r
o
le
in
th
e
f
ield
o
f
tex
t
m
in
in
g
[
2
0
]
.
T
o
co
m
p
u
te
th
e
s
im
ilar
ity
b
etwe
en
ea
ch
ter
m
,
th
ey
m
u
s
t
b
e
r
ep
r
esen
ted
as
a
v
ec
to
r
.
T
h
e
well
-
k
n
o
wn
r
ep
r
esen
tatio
n
s
ch
em
e
f
o
r
ter
m
s
u
n
its
is
th
e
v
ec
to
r
s
p
ac
e
m
o
d
el
(
VSM)
.
L
et
T
=
{t
1
,
t
2
,..,t
p
}
r
ep
r
esen
t
th
e
d
is
tin
ct
ter
m
s
th
at
ex
is
t
in
th
e
d
o
cu
m
en
t
co
llectio
n
D,
wh
er
e
p
is
th
e
n
u
m
b
er
o
f
ter
m
s
in
D.
T
h
r
o
u
g
h
VSM
ev
er
y
s
en
ten
ce
(s
i
)
is
r
ep
r
esen
ted
u
s
in
g
th
ese
ter
m
s
as
a
v
ec
to
r
in
n
-
d
im
en
s
io
n
al
s
p
ac
e,
s
i
=
{w
i,
1
,
w
i,
2
,
…,
w
i,
p
},
f
o
r
i=1
to
p
.
E
ac
h
elem
en
t
in
th
e
v
ec
to
r
r
ep
r
esen
t a
ter
m
with
in
a
g
iv
en
s
en
ten
ce
.
T
h
e
v
alu
e
o
f
ea
ch
elem
en
t
in
th
e
v
ec
to
r
ass
ig
n
s
a
weig
h
t
u
s
in
g
ter
m
f
r
eq
u
en
cy
-
in
v
er
s
e
-
s
en
ten
ce
-
f
r
eq
u
en
cy
as e
x
p
lain
ed
s
h
o
wn
in
(
1
)
[
2
4
]
.
,
=
,
∗
l
og
(
)
(
1
)
w
h
er
e:
TF
i,
k
is
th
e
ter
m
f
r
eq
u
en
cy
,
r
ep
r
esen
ts
h
o
w
m
an
y
ter
m
k
ap
p
ea
r
s
in
th
e
s
en
ten
ce
(
S
i
).
n
.
th
e
n
u
m
b
er
o
f
s
en
ten
ce
s
in
D.
n
k.
th
e
n
u
m
b
er
o
f
s
en
ten
ce
s
in
wh
ich
ter
m
ap
p
ea
r
s
.
T
h
e
weig
h
t
,
o
f
th
e
ter
m
s
h
o
u
ld
b
e
ze
r
o
if
it d
o
es n
o
t e
x
is
t in
th
e
s
en
ten
ce
S
i
.
T
h
e
VSM
r
eq
u
ir
es
h
ig
h
d
im
en
s
io
n
ality
o
f
f
ea
tu
r
e
s
p
ac
e
th
at
af
f
ec
ts
th
e
p
er
f
o
r
m
an
ce
o
f
T
S.
Dep
en
d
in
g
o
n
th
e
n
u
m
b
er
o
f
ter
m
s
in
ea
ch
s
en
ten
ce
th
e
s
p
ec
if
ied
v
ec
to
r
d
im
en
s
io
n
p
is
v
er
y
lar
g
e
an
d
h
as n
u
m
er
o
u
s
n
u
ll
elem
en
ts
,
wh
ich
ca
n
b
e
a
m
ajo
r
d
is
ad
v
an
tag
e
o
f
VSM.
T
h
e
ce
n
ter
o
f
th
e
d
o
cu
m
en
t
co
llectio
n
(
o
)
ca
n
b
e
ca
lcu
lated
as th
e
av
er
ag
e
o
f
weig
h
ts
,
o
f
ter
m
t
k
f
o
r
all
S
i
in
th
e
d
o
cu
m
en
t c
o
llectio
n
as
s
h
o
wn
in
(
2
)
[
2
5
]
.
=
1
∑
,
=
1
=
1
(
2
)
3
.
3
.
Su
mm
a
r
y
o
f
qu
a
lity
f
a
ct
o
rs
I
n
th
is
s
ec
tio
n
,
th
e
im
p
o
r
tan
t
f
ac
to
r
s
f
o
r
s
u
m
m
ar
y
q
u
ality
ar
e
d
em
o
n
s
tr
ated
.
T
h
at
co
n
s
is
ts
o
f
co
v
er
ag
e
,
d
iv
er
s
ity
an
d
r
ea
d
ab
ilit
y
.
E
ac
h
f
ac
to
r
p
lay
s
im
p
o
r
tan
t
r
o
le
in
th
e
s
u
m
m
ar
izatio
n
p
r
o
ce
s
s
.
T
h
ese
f
ac
to
r
s
ar
e
d
escr
ib
ed
as b
elo
w:
3
.
3
.
1
.
Co
v
er
a
g
e
T
h
e
g
o
al
o
f
T
S is
to
co
v
er
th
e
m
ain
co
n
ten
t
o
f
th
e
s
u
m
m
ar
ized
d
o
cu
m
en
ts
b
y
ch
o
o
s
in
g
s
u
b
s
et
S
D
th
at
co
v
er
s
as
m
an
y
co
n
ce
p
tu
al
s
en
ten
ce
s
as
p
o
s
s
ib
le.
Su
m
m
ar
y
co
v
er
ag
e
ca
n
b
e
ca
lcu
lated
b
y
m
ea
s
u
r
in
g
th
e
co
s
in
e
s
im
ilar
ity
b
etwe
en
th
e
ce
n
ter
o
f
d
o
cu
m
en
t c
o
llectio
n
(
O)
an
d
ea
ch
s
en
ten
ce
(
S
i
)
as
s
h
o
wn
in
(
3
)
.
(
,
)
=
∑
=
1
√
∑
(
)
2
=
1
∗
√
∑
(
)
2
=
1
=
1
(
3
)
T
h
e
s
im
ilar
ity
b
etwe
en
th
e
ce
n
ter
o
f
d
o
cu
m
en
t
co
llectio
n
an
d
ea
ch
s
en
ten
ce
d
ec
id
es
th
e
im
p
o
r
tan
ce
o
f
th
e
s
en
ten
ce
an
d
wh
eth
er
it is
in
clu
d
ed
in
th
e
g
en
er
ated
s
u
m
m
ar
y
[
2
6
]
.
3
.
3
.
2
.
Div
er
s
it
y
A
s
u
m
m
ar
y
th
at
h
as
a
h
ig
h
d
iv
er
s
ity
b
etwe
en
its
s
en
ten
ce
s
ca
n
b
e
co
n
s
id
er
ed
as
a
g
o
o
d
s
u
m
m
ar
y
b
ec
au
s
e
its
s
en
ten
ce
s
s
o
lv
e
th
e
p
r
o
b
lem
o
f
in
f
o
r
m
atio
n
r
ed
u
n
d
an
cy
th
at
o
cc
u
r
s
in
m
o
s
t
s
u
m
m
ar
izatio
n
m
o
d
els
,
esp
ec
ially
in
MD
S.
T
h
u
s
,
to
ac
h
iev
e
an
ad
eq
u
ate
s
u
m
m
ar
y
,
th
e
s
en
ten
ce
s
s
h
o
u
ld
h
av
e
a
h
ig
h
d
iv
er
s
ity
am
o
n
g
th
em
.
Su
m
m
ar
y
d
iv
er
s
ity
is
co
m
p
u
ted
b
y
co
n
s
id
er
in
g
th
e
to
tal
v
alu
e
o
f
s
en
ten
ce
s
im
ilar
ity
.
A
g
o
o
d
s
u
m
m
ar
y
is
ass
o
ciate
d
with
lo
wer
d
iv
er
s
ity
v
alu
es
th
at
en
s
u
r
e
m
in
im
u
m
in
f
o
r
m
atio
n
r
ed
u
n
d
an
cy
.
As
s
h
o
ws
in
(
4
)
th
e
f
o
r
m
u
latio
n
to
co
m
p
u
te
s
en
ten
ce
s
d
iv
er
s
ity
[
2
7
]
.
(
)
=
∑
∑
(
=
+
1
−
1
=
1
,
)
(
4
)
3
.
3
.
3
.
Rea
da
bil
it
y
R
ea
d
ab
ilit
y
is
an
im
p
o
r
tan
t
f
ac
to
r
f
o
r
d
o
cu
m
en
t
s
u
m
m
ar
y
th
at
in
d
icate
s
th
e
s
en
ten
ce
s
in
th
e
s
u
m
m
ar
y
ar
e
h
ig
h
ly
r
elat
ed
to
th
e
n
ex
t
s
en
ten
ce
in
th
e
d
o
cu
m
en
t
s
u
m
m
ar
y
.
T
h
e
r
ea
d
ab
ilit
y
(
R
s
)
o
f
s
u
m
m
ar
y
(
s
)
with
len
g
th
(
S)
ca
n
b
e
f
o
r
m
u
lated
as
s
h
o
wn
in
(
5
)
an
d
(
6
)
r
esp
ec
tiv
ely
[
2
8
]
.
=
∑
(
0
≤
≤
,
+
1
(
5
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
1
:
8
9
-
95
92
=
∀
×
(
6
)
T
h
e
o
b
jectiv
e
f
u
n
ctio
n
is
to
m
ax
im
ize
th
e
th
r
ee
f
ac
to
r
s
co
v
er
ag
e,
d
iv
er
s
ity
an
d
r
ea
d
ab
ilit
y
as
s
h
o
wn
in
(
7
)
.
(
)
=
(
)
+
(
)
+
(
)
(
7
)
3
.
4
.
H
a
r
m
o
ny
s
ea
rc
h ba
s
ed
M
DS
Har
m
o
n
y
s
ea
r
ch
alg
o
r
ith
m
(
HS
A
)
is
a
m
eta
-
h
eu
r
is
tic
alg
o
r
ith
m
th
at
was
d
ev
elo
p
ed
b
y
Z.
W
.
Gr
ee
n
,
et
a
l
.
in
2
0
0
1
[
2
9
]
.
HS
A
r
eq
u
ir
es
less
m
ath
em
atica
l
o
p
er
atio
n
s
an
d
ca
n
b
e
ea
s
ily
u
s
ed
in
m
an
y
o
p
tim
izatio
n
p
r
o
b
lem
s
co
m
p
ar
ed
to
o
th
er
m
eta
-
h
eu
r
is
tic
alg
o
r
ith
m
s
.
HS
A
alg
o
r
ith
m
tr
ies
to
s
ea
r
ch
f
o
r
a
g
lo
b
a
l
s
o
lu
tio
n
s
p
ec
if
ied
b
y
th
e
o
b
jectiv
e
f
u
n
ctio
n
.
T
h
e
d
ec
is
io
n
v
ar
iab
les
ass
ig
n
v
alu
es
to
d
eter
m
in
e
th
e
o
b
jectiv
e
f
u
n
ctio
n
,
is
s
im
ilar
to
to
n
es o
f
m
u
s
ical
in
s
tr
u
m
en
ts
th
at
d
ec
id
e
th
e
ae
s
th
etic
q
u
ality
.
T
h
u
s
,
th
e
HS
A
alg
o
r
ith
m
wo
r
k
s
s
im
ilar
ly
to
a
m
u
s
ician
wh
o
is
lo
o
k
in
g
f
o
r
th
e
b
est h
ar
m
o
n
y
[
3
0
]
.
T
h
e
h
ar
m
o
n
y
v
ec
to
r
v
alu
es a
r
e
s
to
r
ed
i
n
th
e
h
ar
m
o
n
y
m
em
o
r
y
(
HM
)
m
atr
ix
as f
o
llo
ws
;
=
[
1
1
2
1
1
⋮
⋮
⋮
1
2
]
wh
er
e
[
x
1
i
,
x
2
i
,..,
x
n
i
]
is
a
ca
n
d
id
ate
s
o
lu
tio
n
.
T
h
e
HM
is
in
iti
alize
d
b
y
r
an
d
o
m
v
ar
iab
les.
A
ls
o
,
two
im
p
o
r
tan
t
p
ar
am
eter
s
th
at
s
h
o
u
ld
b
e
in
itialized
ar
e
Har
m
o
n
y
m
em
o
r
y
c
o
n
s
id
er
in
g
r
ate
(
HM
C
R
)
an
d
p
itch
ad
ju
s
tin
g
r
ate
(
PAR
)
.
T
h
ese
two
p
ar
am
eter
s
ar
e
u
p
d
ated
b
y
h
a
r
m
o
n
y
m
em
o
r
y
c
o
n
s
id
er
atio
n
(
HM
C
)
an
d
p
itch
ad
ju
s
tin
g
(
PA)
r
esp
ec
tiv
ely
.
T
h
e
HM
C
R
p
lay
s
an
im
p
o
r
tan
t
r
o
le
in
s
elec
tin
g
a
v
alu
e
f
r
o
m
m
e
m
o
r
y
wh
ile
th
e
PA
is
im
p
o
r
tan
t
f
o
r
b
o
th
ex
p
lo
itatio
n
an
d
ex
p
lo
r
atio
n
.
T
h
e
ex
p
lo
itatio
n
is
u
s
ed
to
f
in
d
o
p
tim
al
s
o
lu
tio
n
s
,
wh
er
ea
s
ex
p
lo
r
atio
n
is
u
s
ed
to
av
o
id
l
o
ca
l m
in
im
a
[
3
1
]
.
T
h
e
f
o
llo
win
g
alg
o
r
ith
m
s
h
o
ws h
o
w
HSA
is
u
s
ed
f
o
r
te
x
t su
m
m
ar
izatio
n
.
-
Step
1
:
co
llect
a
s
et
o
f
m
u
ltip
le
d
o
cu
m
e
n
ts
D=
{D1
,
D2
,
.
.
,
DN}
wh
er
e
ea
ch
D
i
r
ep
r
esen
t
in
d
i
v
id
u
al
d
o
c
u
m
en
t
-
Step
2
:
ap
p
ly
p
r
ep
r
o
ce
s
s
in
g
s
tep
s
to
ea
ch
D
i
-
Step
3
: f
o
r
ea
c
h
D
i
ca
lcu
late
th
e
co
v
er
ag
e
as
s
h
o
wn
in
(
3
)
-
Step
4
: f
o
r
ea
c
h
D
i
ca
lcu
late
th
e
d
iv
er
s
ity
as
s
h
o
wn
in
(
4
)
-
Step
5
: f
o
r
ea
c
h
D
i
ca
lcu
late
th
e
r
ea
d
ab
ilit
y
as
s
h
o
wn
in
(
6
)
-
Step
6
: in
itialize
th
e
HM
with
r
an
d
o
m
s
o
lu
tio
n
s
an
d
also
in
iti
alize
HM
C
R
,
PAR
-
Step
7
: so
r
t th
e
en
tire
s
o
lu
tio
n
o
f
HM
an
d
r
an
k
th
em
ac
co
r
d
i
n
g
to
s
h
o
w
n
in
(
7
)
-
Step
8
:
im
p
r
o
v
es
a
n
ew
s
o
lu
tio
n
f
r
o
m
HM
as f
o
llo
ws
;
a
.
if
r
an
d
(0
.
1
)
<
HM
C
R
th
en
ch
o
o
s
e
n
ew
s
o
lu
tio
n
f
r
o
m
HM
else c
h
o
o
s
e
a
s
o
lu
tio
n
r
a
n
d
o
m
ly
.
b
.
if
r
an
d
(
0
.
1
)
<
PAR
th
en
ch
o
o
s
in
g
an
ad
jace
n
t
v
alu
e
o
f
th
e
s
elec
ted
v
alu
e
to
d
e
p
en
d
o
n
b
a
n
d
wid
th
.
-
Step
9
:
if
th
e
n
ew
s
o
lu
tio
n
is
b
etter
th
an
wo
r
s
t
s
to
r
ed
{
b
ased
s
h
o
wn
in
(
7
)
}
o
n
e
th
en
u
p
d
ate
th
e
H
M
b
y
th
e
n
ew
s
o
lu
tio
n
E
ls
e
elim
in
ate
th
e
n
ew
s
o
lu
tio
n
-
Step
1
0
: c
h
ec
k
s
to
p
co
n
d
itio
n
i
f
th
e
r
esu
lt b
e
in
a
s
tab
le
s
tate
th
en
en
d
E
ls
e
g
o
to
s
tep
7
4.
DATAS
E
T
AND
E
VAL
UAT
I
O
N
M
E
T
R
I
CS
T
AC
-
2
0
1
1
d
ataset
was
u
s
ed
to
test
th
e
s
y
s
tem
p
er
f
o
r
m
an
ce
.
T
h
e
d
ataset
co
n
s
is
ts
o
f
s
ev
en
lan
g
u
ag
es
(
E
n
g
lis
h
,
Ar
ab
ic,
Gr
ee
k
,
C
ze
ch
,
Fre
n
ch
,
Hin
d
i,
Heb
r
ew)
.
T
h
er
e
ar
e1
0
to
p
ics,
ea
ch
o
f
1
0
d
o
cu
m
en
ts
f
o
r
ea
ch
lan
g
u
ag
e
[
3
2
]
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
d
ea
ls
with
th
e
E
n
g
lis
h
lan
g
u
ag
e
o
n
ly
.
R
ec
all
-
o
r
ien
ted
u
n
d
er
s
tu
d
y
f
o
r
g
is
tin
g
ev
alu
atio
n
(
R
OUGE
)
[
3
3
]
was
u
s
ed
to
ev
alu
ate
th
e
p
r
o
p
o
s
ed
s
y
s
tem
th
e
o
u
tp
u
ts
o
f
a
r
o
u
g
e
p
ac
k
ag
e
ar
e
th
r
ee
n
u
m
b
er
s
wh
ich
r
ep
r
esen
t Pr
ec
is
io
n
,
R
ec
all,
an
d
F−sco
r
e.
T
h
ey
ar
e
f
o
r
m
u
lated
as f
o
llo
ws.
=
∩
ℎ
(
8
)
=
∩
ℎ
(
9
)
−
=
2
∗
∗
+
(
1
0
)
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
E
xtra
ctive
mu
lti d
o
cu
men
t su
mma
r
iz
a
tio
n
u
s
in
g
h
a
r
mo
n
y
s
ea
r
ch
a
lg
o
r
ith
m
(
Zu
h
a
ir
Hu
s
s
ein
A
li
)
93
5.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
R
OUGE
-
1
an
d
R
OUGE
-
2
m
atr
ices
h
av
e
b
ee
n
u
s
ed
to
m
ea
s
u
r
e
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
s
u
m
m
ar
y
.
T
h
ese
m
atr
ices
ar
e
v
er
y
s
im
ilar
to
h
u
m
an
ju
d
g
m
en
t.
T
h
e
s
u
m
m
ar
y
p
er
f
o
r
m
an
ce
is
m
ea
s
u
r
ed
b
y
co
m
p
u
tin
g
th
e
o
v
er
lap
b
etwe
en
s
y
s
tem
s
u
m
m
ar
ies
with
h
u
m
an
s
u
m
m
ar
ies.
R
OUGE
-
1
is
co
n
ce
r
n
ed
with
co
m
p
u
tin
g
u
n
ig
r
am
s
o
v
er
lap
s
wh
ile
R
OUGE
-
2
is
co
n
ce
r
n
ed
with
co
m
p
u
tin
g
b
ig
r
am
s
o
v
er
lap
s
.
T
h
e
r
esu
lts
ar
e
co
m
p
a
r
ed
with
th
e
r
esu
lts
o
f
[
1
2
]
th
at
in
clu
d
ed
p
ee
r
s
u
m
m
ar
ies
in
th
e
T
AC
-
2
0
1
1
d
ata
s
et
.
T
ab
les
f
r
o
m
1
to
4
s
h
o
w
th
e
r
esu
lts
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
[
1
2
]
u
s
in
g
R
OUGE
-
1
an
d
R
OUGE
-
2
r
esp
ec
tiv
ely
.
As
s
ee
n
f
r
o
m
T
ab
le
s
1
an
d
2
,
co
m
p
ar
ed
to
th
e
r
esu
lt
o
f
[
1
2
]
,
u
s
in
g
R
OUGE
-
1
.
T
h
e
tab
les
s
h
o
w
th
e
r
ec
all
an
d
F
-
s
co
r
e
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
ar
e
h
ig
h
er
.
Ho
wev
er
,
th
e
p
r
ec
is
io
n
is
lo
wer
.
T
h
e
J
u
d
g
m
en
t b
etwe
en
th
e
r
ec
all
an
d
th
e
p
r
ec
is
io
n
is
th
e
F
-
s
co
r
e
th
at
co
n
s
id
er
th
em
b
o
th
.
As
k
n
o
wn
,
th
e
p
r
ec
is
io
n
is
co
m
p
u
ted
b
y
d
iv
id
in
g
th
e
n
u
m
b
er
o
f
s
en
ten
ce
s
o
v
er
lap
b
etwe
en
s
y
s
tem
s
u
m
m
ar
y
an
d
id
ea
l
s
u
m
m
ar
y
b
y
th
e
n
u
m
b
er
o
f
s
en
ten
ce
s
in
th
e
s
y
s
tem
s
u
m
m
ar
y
.
W
h
er
ea
s
th
e
r
ec
all
is
co
m
p
u
ted
b
y
d
iv
id
in
g
th
e
n
u
m
b
er
o
f
s
en
ten
ce
s
o
v
er
la
p
b
etwe
en
s
y
s
tem
s
u
m
m
ar
y
an
d
id
ea
l
s
u
m
m
ar
y
b
y
th
e
n
u
m
b
er
o
f
s
en
ten
ce
s
in
th
e
id
ea
l
s
u
m
m
ar
y
.
T
h
u
s
,
b
y
in
cr
ea
s
in
g
th
e
n
u
m
b
er
o
f
wo
r
d
s
in
th
e
s
y
s
tem
s
u
m
m
ar
y
lead
s
to
d
ec
r
ea
s
in
g
p
r
ec
is
io
n
.
W
h
ile
d
ec
r
ea
s
in
g
th
e
n
u
m
b
er
o
f
wo
r
d
s
in
th
e
s
y
s
tem
s
u
m
m
ar
y
lead
s
to
d
ec
r
ea
s
in
g
th
e
r
ec
all.
T
h
e
len
g
th
o
f
ea
ch
id
ea
l
s
u
m
m
ar
y
b
et
wee
n
2
4
0
-
2
5
0
wo
r
d
s
,
wh
ile
th
e
len
g
th
o
f
ea
ch
in
d
iv
id
u
al
g
en
er
ated
s
u
m
m
ar
y
is
m
o
r
e
th
an
2
5
0
,
th
e
r
ea
s
o
n
b
eh
in
d
th
e
len
g
th
o
f
th
e
g
en
er
ated
s
u
m
m
ar
y
is
th
e
m
ec
h
an
is
m
o
f
cr
ea
tio
n
th
at
is
b
ased
o
n
ad
d
in
g
s
en
ten
ce
s
to
th
e
s
u
m
m
ar
y
with
o
u
t
an
y
ch
an
g
e
to
th
e
len
g
th
o
f
th
e
s
en
ten
ce
.
W
h
ich
ca
u
s
es
th
e
s
u
m
m
ar
y
len
g
th
o
f
m
o
r
e
th
an
2
5
0
wo
r
d
s
,
esp
ec
ially
wh
en
th
e
last
s
en
ten
ce
is
to
o
lo
n
g
.
T
ab
le
1
.
C
o
m
p
a
r
is
o
n
b
etwe
en
th
e
p
r
o
p
o
s
ed
m
o
d
el
an
d
[
1
2
]
r
esu
lts
u
s
in
g
R
OUGE
-
1
M
o
d
e
l
To
p
i
c
P
r
o
p
o
se
d
m
o
d
e
l
R
e
s
u
l
t
s
[
1
2
]
R
e
s
u
l
t
s
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F
-
S
c
o
r
e
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F
-
S
c
o
r
e
I
D
1
0
.
3
7
2
0
0
0
.
4
8
8
1
8
0
.
4
2
2
2
4
0
.
4
1
2
5
3
0
.
4
0
5
2
4
0
.
4
0
7
7
6
I
D
2
0
.
3
9
2
2
0
0
.
5
2
2
2
3
0
.
4
4
7
9
7
0
.
4
5
6
5
5
0
.
4
6
4
8
1
0
.
4
6
0
6
2
I
D
3
0
.
4
1
2
6
6
0
.
5
6
8
8
7
0
.
4
7
8
3
3
0
.
4
7
9
0
9
0
.
4
3
1
6
9
0
.
4
5
4
0
4
I
D
4
0
.
4
0
2
1
6
0
.
5
6
6
3
4
0
.
4
7
0
3
3
0
.
4
4
9
6
6
0
.
4
4
4
2
3
0
.
4
4
6
9
1
I
D
5
0
.
4
1
2
1
3
0
.
6
1
6
8
6
0
.
4
9
4
1
2
0
.
4
3
5
1
3
0
.
4
1
0
9
2
0
.
4
2
2
4
3
I
D
6
0
.
3
7
6
2
5
0
.
4
6
8
2
5
0
.
4
1
7
2
3
0
.
4
5
1
2
2
0
.
3
5
4
7
0
.
3
9
6
1
7
I
D
7
0
.
3
5
3
2
1
0
.
5
5
3
6
7
0
.
4
3
1
2
8
0
.
3
9
5
3
0
.
3
9
5
8
6
0
.
3
9
5
4
7
I
D
8
0
.
4
2
1
4
2
0
.
6
9
5
8
3
0
.
5
2
4
9
2
0
.
3
9
2
6
5
0
.
3
8
7
1
4
0
.
3
8
9
8
5
I
D
9
0
.
3
7
2
5
1
0
.
5
5
1
2
3
0
.
4
4
4
5
8
0
.
3
7
7
2
6
0
.
3
8
1
0
5
0
.
3
7
9
1
2
I
D
1
0
0
.
4
3
7
1
1
0
.
6
0
2
5
1
0
.
5
0
6
6
5
0
.
5
1
8
0
6
0
.
5
2
4
8
8
0
.
5
2
1
4
1
T
ab
le
2
.
Av
e
r
ag
e
p
r
ec
is
io
n
,
r
e
ca
ll
an
d
F
-
s
co
r
e
u
s
in
g
R
OUG
E
-
1
M
o
d
e
l
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F
-
S
c
o
r
e
P
r
o
p
o
se
d
m
o
d
e
l
0
.
3
9
5
1
6
5
0
.
5
6
3
3
9
6
0
.
4
6
3
7
6
[
1
2
]
0
.
4
3
6
7
4
5
0
.
4
2
0
0
5
2
0
.
4
2
7
3
7
6
T
ab
le
s
3
an
d
4
s
h
o
w
th
e
r
esu
lts
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
u
s
in
g
R
OUGE
-
2
.
T
h
e
ef
f
icien
cy
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
was
ev
id
en
t
wh
en
it
was
u
s
ed
R
OUGE
-
2
b
ec
au
s
e
R
OUGE
-
2
is
clo
s
er
to
h
u
m
an
s
u
m
m
ar
y
th
an
R
OUGE
-
1
.
T
h
e
Av
er
ag
e
o
f
th
e
th
r
ee
m
etr
ics r
ec
all,
p
r
ec
is
io
n
an
d
F
-
s
co
r
e
ar
e
b
etter
th
an
[
1
2
]
.
T
h
is
is
b
ec
au
s
e
o
f
th
e
g
o
o
d
d
ef
in
itio
n
s
o
f
co
v
er
ag
e,
d
iv
er
s
ity
,
an
d
r
ea
d
ab
ilit
y
an
d
d
u
e
to
th
e
g
o
o
d
p
er
f
o
r
m
an
ce
o
f
HS
A
in
r
eg
ar
d
s
to
ch
o
o
s
in
g
th
e
m
o
s
t su
itab
le
s
en
ten
ce
s
to
b
e
in
clu
d
ed
in
th
e
f
in
al
s
u
m
m
ar
y
.
T
ab
le
3
.
C
o
m
p
a
r
is
o
n
b
et
wee
n
p
r
o
p
o
s
ed
m
o
d
el
an
d
[
1
2
]
r
esu
l
ts
u
s
in
g
R
OUGE
-
2
M
o
d
e
l
To
p
i
c
P
r
o
p
o
se
d
m
o
d
e
l
R
e
s
u
l
t
s
[
1
2
]
R
e
s
u
l
t
s
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F
-
S
c
o
r
e
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F
-
S
c
o
r
e
I
D
1
0
.
2
4
5
5
7
0
.
1
7
5
9
1
0
.
2
0
4
9
8
0
.
1
2
4
4
8
0
.
1
2
1
2
5
0
.
1
2
2
4
7
I
D
2
0
.
1
4
2
1
1
0
.
1
7
8
2
9
0
.
1
5
8
1
5
0
.
1
6
7
7
9
0
.
1
7
0
5
2
0
.
1
6
9
1
4
I
D
3
0
.
1
9
2
3
0
0
.
2
6
7
1
0
0
.
2
2
3
6
1
0
.
1
9
2
5
6
0
.
1
7
3
3
0
.
1
8
2
3
7
I
D
4
0
.
1
4
3
9
3
0
.
2
6
1
8
2
0
.
1
8
5
7
4
0
.
1
5
3
6
9
0
.
1
5
1
7
0
.
1
5
2
6
9
I
D
5
0
.
1
7
2
5
6
0
.
3
2
5
5
1
0
.
2
2
5
5
5
0
.
1
4
4
0
4
0
.
1
3
6
0
5
0
.
1
3
9
8
5
I
D
6
0
.
0
9
3
2
3
0
.
1
3
9
6
7
0
.
1
1
1
8
1
0
.
1
3
6
7
0
.
1
0
6
5
5
0
.
1
1
9
3
7
I
D
7
0
.
1
3
8
5
5
0
.
2
0
1
5
1
0
.
1
6
4
2
0
0
.
0
9
6
1
2
0
.
0
9
6
6
2
0
.
0
9
6
3
5
I
D
8
0
.
2
2
2
1
3
0
.
3
6
5
4
8
0
.
2
7
6
3
1
0
.
1
2
2
9
8
0
.
1
2
1
4
4
0
.
1
2
2
1
9
I
D
9
0
.
2
0
3
7
4
0
.
1
4
9
1
2
0
.
1
7
2
2
0
0
.
1
0
8
4
1
0
.
1
0
9
6
2
0
.
1
0
9
I
D
1
0
0
.
1
9
3
0
3
0
.
2
7
2
5
2
0
.
2
2
5
9
8
0
.
2
4
8
3
0
.
2
5
1
7
7
0
.
2
5
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6
9
3
0
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
,
Vo
l.
19
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
1
:
8
9
-
95
94
T
ab
le
4
.
Av
e
r
ag
e
p
r
ec
is
io
n
,
r
e
ca
ll a
n
d
F
-
s
co
r
e
u
s
in
g
R
OUG
E
-
2
M
o
d
e
l
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F
-
S
c
o
r
e
P
r
o
p
o
se
d
m
o
d
e
l
0
.
1
9
3
5
2
4
0
.
2
9
0
5
8
2
0
.
2
3
1
9
8
4
[
1
2
]
0
.
1
4
9
5
0
7
0
.
1
4
3
8
8
2
0
.
1
4
6
3
4
2
6.
CO
NCLU
SI
O
N
T
h
e
n
ee
d
f
o
r
in
f
lu
en
tial
MD
S
ap
p
r
o
ac
h
es
to
ex
tr
ac
t
s
ig
n
if
ican
t
in
f
o
r
m
atio
n
f
r
o
m
a
d
o
cu
m
en
t
co
llectio
n
b
ec
o
m
es
o
f
n
ec
ess
ity
.
T
h
is
p
ap
er
u
s
ed
HSA
b
ased
MD
S
to
cr
ea
te
a
g
en
er
ic
ex
tr
ac
tiv
e
s
u
m
m
ar
y
.
T
h
e
Su
m
m
ar
izer
u
s
ed
a
b
en
ch
m
ar
k
d
ataset
ca
lled
T
AC
-
2
0
1
1
,
an
d
R
OUGE
p
ac
k
ag
e
was
ap
p
lied
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
s
u
m
m
ar
izer
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
is
b
ased
o
n
th
r
ee
im
p
o
r
tan
t
is
s
u
es
in
MD
S
th
at
in
clu
d
e
co
v
er
ag
e,
d
iv
er
s
ity
,
an
d
r
ea
d
ab
ilit
y
.
Go
o
d
r
esu
lts
wer
e
o
b
tain
ed
f
r
o
m
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
h
e
lim
itatio
n
o
f
th
is
m
eth
o
d
is
co
n
tr
o
llin
g
th
e
p
ar
am
eter
s
o
f
HM
C
R
an
d
PAR
th
at
r
eq
u
ir
e
s
p
ec
ial
tr
ea
tm
en
t.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
e
au
th
o
r
wo
u
ld
lik
e
to
th
an
k
Mu
s
tan
s
ir
iy
ah
Un
iv
er
s
ity
(
www.
u
o
m
u
s
tan
s
ir
iy
ah
.
ed
u
.
i
q
)
B
ag
h
d
ad
-
I
r
aq
f
o
r
its
s
u
p
p
o
r
t
in
th
e
p
r
esen
t
wo
r
k
.
Sp
ec
ial
th
an
k
s
to
Miss
Ma
wad
a
Al
Fais
al
f
o
r
h
er
v
alu
ab
le
s
cien
tific
ad
v
ice.
RE
F
E
R
E
NC
E
S
[1
]
A.
Ah
m
a
d
,
T
.
Ah
m
a
d
,
“
A
g
a
m
e
t
h
e
o
ry
a
p
p
r
o
a
c
h
fo
r
m
u
lt
i
-
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
,”
Ar
a
b
i
a
n
J
o
u
r
n
a
l
fo
r
S
c
ie
n
c
e
a
n
d
En
g
i
n
e
e
rin
g
,
v
o
l.
4
4
,
p
p
.
3
6
5
5
-
3
6
6
7
,
2
0
1
9
.
[2
]
R.
M
.
Alg
u
li
e
v
,
R
.
M.
Ali
g
u
li
y
e
v
,
N
.
R.
Isa
z
a
d
e
,
“
M
u
lt
i
p
le
d
o
c
u
m
e
n
ts
su
m
m
a
riza
ti
o
n
b
a
se
d
o
n
t
h
e
e
v
o
lu
ti
o
n
a
ry
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
,
”
Exp
e
rt
S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
4
0
,
n
o
.
5
,
p
p
.
1
6
7
5
-
1
6
8
9
,
Ap
ril
2
0
1
3
.
[3
]
J.
M
.
S
a
n
c
h
e
z
-
G
o
m
e
z
,
e
t
a
l.
,
“
Ex
trac
ti
v
e
m
u
lt
i
-
d
o
c
u
m
e
n
t
tex
t
su
m
m
a
riza
ti
o
n
u
sin
g
a
m
u
lt
i
-
o
b
jec
ti
v
e
a
rti
ficia
l
b
e
e
c
o
lo
n
y
o
p
ti
m
iza
ti
o
n
a
p
p
ro
a
c
h
,”
K
n
o
wled
g
e
-
Ba
se
d
S
y
ste
ms
,
v
o
l.
1
5
9
,
p
p
.
1
-
8
,
No
v
e
m
b
e
r
2
0
1
8
.
[4
]
R.
M
.
Alg
u
li
e
v
,
e
t
a
l
.,
“
DES
A
M
C+
Do
c
S
u
m
:
Diffe
re
n
t
ial
e
v
o
l
u
ti
o
n
wit
h
se
lf
-
a
d
a
p
t
iv
e
m
u
tatio
n
a
n
d
c
ro
ss
o
v
e
r
p
a
ra
m
e
ters
fo
r
m
u
lt
i
-
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
,
”
Kn
o
wled
g
e
-
Ba
se
d
S
y
ste
ms
,
v
o
l.
3
6
,
p
p
.
2
1
-
3
8
,
De
c
e
m
b
e
r
2
0
1
2
.
[5
]
M
.
G
a
m
b
h
ir
,
V.
G
u
p
ta,
“
Re
c
e
n
t
a
u
to
m
a
ti
c
tex
t
su
m
m
a
riza
ti
o
n
tec
h
n
iq
u
e
s:
a
su
rv
e
y
,
”
Arti
f
icia
l
I
n
tell
ig
e
n
c
e
Rev
iew
,
v
o
l.
4
7
,
p
p
.
1
-
6
6
,
2
0
1
7
.
[6
]
R.
F
e
rre
ira
,
e
t
a
l
.,
“
As
se
ss
in
g
se
n
ten
c
e
sc
o
rin
g
tec
h
n
iq
u
e
s
fo
r
e
x
t
ra
c
ti
v
e
tex
t
su
m
m
a
riza
ti
o
n
,
”
Exp
e
rt
sy
ste
m
s
wit
h
a
p
p
li
c
a
ti
o
n
s
,
v
o
l.
4
0
,
p
p
.
5
7
5
5
-
5
7
6
4
,
Oc
t
o
b
e
r
2
0
1
3
.
[7
]
M
.
Yo
u
se
fi
-
Az
a
r,
L.
Ha
m
e
y
,
“
Tex
t
su
m
m
a
riza
ti
o
n
u
si
n
g
u
n
s
u
p
e
rv
ise
d
d
e
e
p
lea
rn
in
g
,”
Exp
e
rt
S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
.
6
8
,
p
p
.
93
-
1
0
5
,
F
e
b
ru
a
ry
2
0
1
7
.
[8
]
R.
Alg
u
li
y
e
v
,
e
t
a
l
.
,
“
A
m
o
d
e
l
f
o
r
t
e
x
t
su
m
m
a
riza
ti
o
n
,”
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
I
n
telli
g
e
n
t
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
lo
g
ies
,
v
o
l.
13
,
n
o
.
1
,
p
p
.
6
7
-
8
5
,
Ja
n
u
a
ry
2
0
1
7
.
[9
]
H.
P
.
L
u
h
n
,
“
Th
e
a
u
to
m
a
ti
c
c
re
a
ti
o
n
o
f
li
tera
tu
re
a
b
stra
c
t
s,”
IBM
J
o
u
rn
a
l
o
f
re
se
a
rc
h
a
n
d
d
e
v
e
lo
p
me
n
t
,
v
o
l.
2
,
n
o
.
2
,
p
p
.
1
5
9
-
1
6
5
,
Ap
ril
1
9
5
8
.
[1
0
]
D.
Wan
g
,
T.
Li
,
“
Do
c
u
m
e
n
t
u
p
d
a
te
su
m
m
a
riza
ti
o
n
u
sin
g
in
c
re
m
e
n
tal
h
iera
rc
h
ica
l
c
lu
ste
ri
n
g
,
”
Pro
c
e
e
d
in
g
s
o
f
th
e
1
9
t
h
AC
M
in
ter
n
a
ti
o
n
a
l
c
o
n
fer
e
n
c
e
o
n
In
fo
rm
a
t
io
n
a
n
d
k
n
o
wle
d
g
e
ma
n
a
g
e
me
n
t
,
p
p
.
2
7
9
-
2
8
8
,
Ja
n
u
a
ry
2
0
1
0
.
[1
1
]
D.
Wan
g
,
e
t
a
l
.
,
“
In
teg
ra
ti
n
g
d
o
c
u
m
e
n
t
c
lu
ste
rin
g
a
n
d
m
u
lt
i
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
,”
ACM
T
r
a
n
sa
c
ti
o
n
s
o
n
Kn
o
wled
g
e
Disc
o
v
e
ry
fro
m Da
ta
(T
KDD
)
,
v
o
l.
5
,
n
o
.
3
,
p
p
.
1
-
26
,
A
u
g
u
st
2
0
1
1
.
[1
2
]
M
.
El
-
Ha
j,
U
.
Kru
sc
h
witz
,
C.
F
o
x
,
“
M
u
lt
i
-
d
o
c
u
m
e
n
t
a
ra
b
ic
tex
t
s
u
m
m
a
riza
ti
o
n
,”
Co
n
fer
e
n
c
e
:
In
t
h
e
3
rd
Co
mp
u
ter
sc
ien
c
e
a
n
d
E
lec
tro
n
ic E
n
g
i
n
e
e
rin
g
C
o
n
fer
e
n
c
e
(CEE
C’
1
1
)
,
Ju
l
y
2
0
1
1
.
[1
3
]
K.
S
.
Th
a
k
k
a
r
,
e
t
a
l
.
,
“
G
ra
p
h
-
b
a
se
d
a
lg
o
rit
h
m
s
fo
r
tex
t
su
m
m
a
riza
ti
o
n
,”
2
0
1
0
3
rd
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Eme
rg
in
g
T
re
n
d
s i
n
E
n
g
in
e
e
rin
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
p
p
.
5
1
6
-
5
1
9
,
N
o
v
e
m
b
e
r
2
0
1
0
.
[1
4
]
S
.
Bri
n
,
L.
P
a
g
e
,
“
T
h
e
a
n
a
t
o
m
y
o
f
a
larg
e
-
sc
a
le
h
y
p
e
rtex
t
u
a
l
we
b
se
a
rc
h
e
n
g
i
n
e
,
”
C
o
mp
u
ter
n
e
t
wo
rk
s
a
n
d
IS
D
N
S
y
ste
ms
,
v
o
l.
3
0
,
n
o
.
1
-
7
,
p
p
.
1
0
7
-
1
1
7
,
Ap
r
il
1
9
9
8
.
[1
5
]
R.
M
i
h
a
lce
a
,
P
.
Tara
u
,
“
Tex
tra
n
k
:
b
rin
g
in
g
o
r
d
e
r
i
n
to
tex
t
,”
Pr
o
c
e
e
d
in
g
s
o
f
t
h
e
2
0
0
4
c
o
n
fer
e
n
c
e
o
n
e
mp
irica
l
me
th
o
d
s
in
n
a
t
u
ra
l
l
a
n
g
u
a
g
e
p
ro
c
e
ss
in
g
,
p
p
.
4
0
4
-
4
1
1
,
J
u
ly
2
0
0
4
.
[1
6
]
H.
N.
T.
Th
u
,
“
An
o
p
ti
m
iza
ti
o
n
tex
t
su
m
m
a
riza
ti
o
n
m
e
th
o
d
b
a
se
d
o
n
n
a
iv
e
Ba
y
e
s a
n
d
t
o
p
ic wo
r
d
fo
r
sin
g
le sy
ll
a
b
l
e
lan
g
u
a
g
e
,”
Ap
p
li
e
d
M
a
th
e
ma
ti
c
a
l
S
c
ien
c
e
s
,
v
o
l.
8
,
p
p
.
99
-
1
1
5
,
Ja
n
u
a
a
ry
2
0
1
4
.
[1
7
]
K.
S
v
o
re
,
e
t
a
l
.
,
“
E
n
h
a
n
c
i
n
g
si
n
g
le
-
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
b
y
c
o
m
b
in
i
n
g
Ra
n
k
Ne
t
a
n
d
th
ir
d
-
p
a
rty
so
u
rc
e
s
,”
Pro
c
e
e
d
in
g
s
o
f
th
e
2
0
0
7
j
o
in
t
c
o
n
fer
e
n
c
e
o
n
e
mp
irica
l
me
th
o
d
s
i
n
n
a
tu
ra
l
l
a
n
g
u
a
g
e
p
ro
c
e
ss
in
g
a
n
d
c
o
m
p
u
t
a
t
io
n
a
l
n
a
t
u
ra
l
la
n
g
u
a
g
e
lea
rn
i
n
g
(EM
N
L
P
-
Co
NL
L
)
,
p
p
.
4
4
8
-
4
5
7
,
Ja
n
u
a
ry
2
0
0
7
.
[1
8
]
P
.
M
.
S
a
b
u
n
a
,
D.
B.
S
e
ty
o
h
a
d
i,
“
S
u
m
m
a
rizin
g
I
n
d
o
n
e
sia
n
tex
t
a
u
t
o
m
a
ti
c
a
ll
y
b
y
u
si
n
g
se
n
ten
c
e
sc
o
r
in
g
a
n
d
d
e
c
isi
o
n
tree
,
”
2
0
1
7
2
n
d
In
ter
n
a
ti
o
n
a
l
c
o
n
fer
e
n
c
e
s
o
n
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
,
I
n
fo
rm
a
ti
o
n
S
y
ste
ms
a
n
d
El
e
c
trica
l
En
g
i
n
e
e
rin
g
(
ICIT
IS
E
E)
,
p
p
.
1
-
6
,
No
v
e
m
b
e
r
2
0
1
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
KA
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l Co
n
tr
o
l
E
xtra
ctive
mu
lti d
o
cu
men
t su
mma
r
iz
a
tio
n
u
s
in
g
h
a
r
mo
n
y
s
ea
r
ch
a
lg
o
r
ith
m
(
Zu
h
a
ir
Hu
s
s
ein
A
li
)
95
[1
9
]
A.
F
a
n
a
n
i,
e
t
a
l
.
,
“
Re
g
re
ss
io
n
m
o
d
e
l
fo
c
u
se
d
o
n
q
u
e
r
y
fo
r
m
u
lt
i
d
o
c
u
m
e
n
ts
su
m
m
a
riza
ti
o
n
b
a
se
d
o
n
sig
n
ifi
c
a
n
c
e
o
f
th
e
se
n
ten
c
e
p
o
sit
io
n
,
”
T
EL
KO
M
NIKA
T
e
lec
o
mm
u
n
ica
ti
o
n
C
o
mp
u
ti
n
g
E
lec
tro
n
ics
a
n
d
C
o
n
tr
o
l,
v
o
l.
1
7
,
n
o
.
6
,
pp.
3
0
5
0
-
3
0
5
6
,
De
c
e
m
b
e
r
2
0
1
9
.
[2
0
]
R.
M
.
Al
g
u
l
iev
,
e
t
a
l
.
,
“
S
e
n
te
n
c
e
se
lec
ti
o
n
fo
r
g
e
n
e
ric
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
u
si
n
g
a
n
a
d
a
p
ti
v
e
d
iffere
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
,
”
S
wa
rm
a
n
d
Evo
l
u
ti
o
n
a
ry
Co
mp
u
t
a
ti
o
n
,
v
o
l.
1
,
n
o
.
1
,
p
p
.
2
1
3
-
2
2
2
,
De
c
e
m
b
e
r
2
0
1
1
.
[2
1
]
R.
Ra
u
tray
,
e
t
a
l
.
,
“
Do
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
u
sin
g
se
n
ten
c
e
fe
a
tu
re
s
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
In
f
o
rm
a
ti
o
n
Retrie
v
a
l
Res
e
a
rc
h
(IJ
IRR
)
,
v
o
l
.
5
,
n
o
.
1
,
p
p
.
36
-
4
7
,
Ja
n
u
a
ry
2
0
1
5
.
[2
2
]
M
.
A.
F
a
tt
a
h
a
n
d
F
.
Re
n
,
“
G
A,
M
R,
F
F
NN
,
P
NN
a
n
d
G
M
M
b
a
se
d
m
o
d
e
ls
fo
r
a
u
to
m
a
ti
c
tex
t
su
m
m
a
riza
ti
o
n
,
”
Co
mp
u
ter
S
p
e
e
c
h
&
L
a
n
g
u
a
g
e
,
v
o
l.
2
3
,
n
o
.
1
,
p
p
.
1
2
6
-
1
4
4
,
Ja
n
u
a
ry
2
0
0
9
.
[2
3
]
M
.
P
e
y
ra
rd
,
J.
Eck
le
-
Ko
h
ler,
“
A
g
e
n
e
ra
l
o
p
ti
m
iza
ti
o
n
fra
m
e
wo
rk
f
o
r
m
u
l
ti
-
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
u
sin
g
g
e
n
e
ti
c
a
lg
o
rit
h
m
s
a
n
d
sw
a
rm
in
telli
g
e
n
c
e
,”
Pro
c
e
e
d
in
g
s
o
f
COLING
2
0
1
6
,
t
h
e
2
6
t
h
I
n
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
o
n
Co
mp
u
t
a
ti
o
n
a
l
L
in
g
u
isti
c
s: T
e
c
h
n
ica
l
P
a
p
e
rs
,
p
p
.
2
4
7
-
2
5
7
,
2
0
1
6
.
[2
4
]
R.
M
.
Al
g
u
l
iev
,
e
t
a
l
.
,
“
CDD
S
:
C
o
n
stra
in
t
-
d
ri
v
e
n
d
o
c
u
m
e
n
t
su
m
m
a
ri
z
a
ti
o
n
m
o
d
e
ls
,”
Exp
e
rt
S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
.
4
0
,
n
o
.
2
,
p
p
.
4
5
8
-
4
6
5
,
F
e
b
r
u
a
ry
2
0
1
3
.
[2
5
]
R.
Ra
u
tra
y
,
R.
C
.
Ba
lab
a
n
tara
y
,
“
An
e
v
o
lu
ti
o
n
a
ry
fra
m
e
wo
rk
f
o
r
m
u
lt
i
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
u
si
n
g
C
u
c
k
o
o
se
a
rc
h
a
p
p
ro
a
c
h
:
M
DSCS
A
,”
A
p
p
li
e
d
c
o
mp
u
ti
n
g
a
n
d
i
n
fo
rm
a
ti
c
s
,
v
o
l.
1
4
,
n
o
.
2
,
p
p
.
1
3
4
-
1
4
4
,
J
u
ly
2
0
1
8
.
[2
6
]
R.
Ra
u
tray
,
R.
C.
Ba
lab
a
n
tara
y
,
“
Ca
t
sw
a
r
m
o
p
ti
m
iza
ti
o
n
b
a
se
d
e
v
o
l
u
ti
o
n
a
ry
fra
m
e
wo
rk
fo
r
m
u
lt
i
d
o
c
u
m
e
n
t
su
m
m
a
riza
ti
o
n
,
”
Ph
y
sic
a
A:
S
t
a
ti
stica
l
M
e
c
h
a
n
ics
a
n
d
it
s A
p
p
li
c
a
ti
o
n
s
,
v
o
l
.
4
7
7
,
p
p
.
1
7
4
-
1
8
6
,
F
e
b
r
u
a
ry
2
0
1
7
.
[2
7
]
K.
Um
a
m
,
e
t
a
l
.,
“
Co
v
e
ra
g
e
d
iv
e
r
sity
,
a
n
d
c
o
h
e
re
n
c
e
o
p
ti
m
iza
ti
o
n
f
o
r
m
u
lt
i
-
d
o
c
u
m
e
n
t
s
u
m
m
a
riza
ti
o
n
,”
J
u
rn
a
l
Ilm
u
Ko
mp
u
ter
d
a
n
In
fo
rm
a
si
,
v
o
l.
8
,
n
o
.
1
,
p
p
.
1
-
1
0
,
2
0
1
5
.
[2
8
]
S
.
H.
M
irs
h
o
jae
i,
a
n
d
B
.
M
a
so
o
m
i,
“
Tex
t
su
m
m
a
riza
ti
o
n
u
si
n
g
c
u
c
k
o
o
se
a
rc
h
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
,”
J
o
u
r
n
a
l
o
f
Co
mp
u
ter
&
Ro
b
o
t
ics
,
v
o
l.
8
,
p
p
.
1
9
-
2
4
,
2
0
1
5
.
[2
9
]
S
.
H.
M
irsh
o
jae
i
,
e
t
a
l
.
,
“
A
n
e
w
h
e
u
risti
c
o
p
ti
m
iza
ti
o
n
a
l
g
o
ri
th
m
:
h
a
rm
o
n
y
se
a
rc
h
,”
S
im
u
la
t
io
n
,
v
o
l
.
7
6
,
p
p
.
60
-
6
8
,
2
0
0
1
.
[3
0
]
B.
Wu
,
e
t
a
l
.
,
“
Hy
b
rid
h
a
rm
o
n
y
se
a
rc
h
a
n
d
a
rti
ficia
l
b
e
e
c
o
lo
n
y
a
lg
o
rit
h
m
fo
r
g
lo
b
a
l
o
p
ti
m
iza
t
io
n
p
r
o
b
lem
s
,
”
Co
mp
u
ter
s&
M
a
th
e
ma
ti
c
s wi
th
A
p
p
li
c
a
ti
o
n
s
,
v
o
l
.
6
4
,
n
o
.
8
,
p
p
.
2
6
2
1
-
2
6
3
4
,
Oc
to
b
e
r
2
0
1
2
.
[3
1
]
S
.
Ab
e
d
,
e
t
a
l
.,
“
Ha
rm
o
n
y
se
a
rc
h
a
lg
o
rit
h
m
f
o
r
w
o
rd
se
n
se
d
isa
m
b
i
g
u
a
ti
o
n
,”
P
lo
S
o
n
e
,
v
o
l.
1
0
,
2
0
1
5
.
[3
2
]
G
.
G
ian
n
a
k
o
p
o
u
l
o
s
,
e
t
a
l
.,
“
TA
C
2
0
1
1
M
u
lt
iL
i
n
g
p
il
o
t
o
v
e
rv
ie
w
,
”
Co
n
fer
e
n
c
e
:
T
e
x
t
An
a
lys
is
Co
n
fer
e
n
c
e
(T
AC
2
0
1
1
),
M
u
lt
iL
i
n
g
S
u
mm
a
risa
ti
o
n
Pi
lo
t
,
Ja
n
u
a
ry
2
0
1
1
.
[3
3
]
C.
Y.
Li
n
,
“
R
o
u
g
e
:
A
p
a
c
k
a
g
e
fo
r
a
u
to
m
a
ti
c
e
v
a
lu
a
t
io
n
o
f
su
m
m
a
ries
,
”
T
e
x
t
su
mm
a
riza
ti
o
n
b
ra
n
c
h
e
s
o
u
t
,
p
p
.
7
4
-
8
1
,
2
0
0
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.