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
.
3
4
9
~
3
5
6
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
6
7
3
4
349
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
Cluster
-
ba
sed in
f
o
rma
tion retr
iev
a
l by usi
ng
(K
-
m
ea
ns)
-
hiera
rchica
l para
llel genetic
alg
o
rit
hms
a
ppro
a
ch
Sa
ra
h H
us
s
ein
T
o
m
a
n
1
,
M
o
ha
m
m
ed
H
a
m
z
a
h Abed
2
,
Z
in
a
h H
us
s
ein
T
o
m
a
n
3
,
1
Ro
a
d
s a
n
d
Tran
sp
o
rt
De
p
a
rtme
n
t
,
Co
ll
e
g
e
o
f
E
n
g
in
e
e
rin
g
,
Un
iv
e
rs
it
y
o
f
Al
-
Qa
d
isi
y
a
h
,
A
d
-
Diwa
n
iah
,
Ira
q
2,
3
Co
m
p
u
ter S
c
ien
c
e
De
p
a
rtme
n
t,
Co
ll
e
g
e
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
In
f
o
rm
a
ti
o
n
Tec
h
n
o
lo
g
y
,
Un
iv
e
rsity
o
f
Al
-
Qa
d
isiy
a
h
,
Ad
-
Diwa
n
iah
,
Ira
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
1
8
,
2
0
2
0
R
ev
is
ed
Au
g
1
4
,
2
0
2
0
Acc
ep
ted
Au
g
2
9
,
2
0
2
0
Clu
ste
r
-
b
a
se
d
i
n
fo
rm
a
ti
o
n
re
tri
e
v
a
l
is
o
n
e
o
f
t
h
e
i
n
fo
rm
a
ti
o
n
re
tri
e
v
a
l
(IR)
to
o
ls
t
h
a
t
o
rg
a
n
ize
,
e
x
trac
t
fe
a
tu
re
s
a
n
d
c
a
teg
o
rize
th
e
we
b
d
o
c
u
m
e
n
ts
a
c
c
o
rd
in
g
t
o
th
e
ir
sim
il
a
rit
y
.
Un
l
i
k
e
trad
it
io
n
a
l
a
p
p
r
o
a
c
h
e
s,
c
lu
ste
r
-
b
a
se
d
IR
is
fa
st
in
p
ro
c
e
ss
in
g
larg
e
d
a
tas
e
ts
o
f
d
o
c
u
m
e
n
t.
T
o
imp
r
o
v
e
t
h
e
q
u
a
li
t
y
o
f
re
tri
e
v
e
d
d
o
c
u
m
e
n
ts,
in
c
re
a
se
th
e
e
fficie
n
c
y
o
f
IR
a
n
d
re
d
u
c
e
irrele
v
a
n
t
d
o
c
u
m
e
n
ts
fr
o
m
u
se
r
se
a
rc
h
.
I
n
th
is
p
a
p
e
r,
we
p
ro
p
o
se
d
a
(
K
-
m
e
a
n
s)
-
h
iera
rc
h
ica
l
p
a
ra
ll
e
l
g
e
n
e
ti
c
a
l
g
o
r
it
h
m
s
a
p
p
ro
a
c
h
(HPG
A)
th
a
t
c
o
m
b
in
e
s
t
h
e
K
-
m
e
a
n
s
c
lu
ste
rin
g
a
lg
o
r
it
h
m
wi
t
h
h
y
b
rid
PG
o
f
m
u
lt
i
-
d
e
m
e
a
n
d
m
a
ste
r/s
lav
e
PG
a
lg
o
rit
h
m
s.
K
-
m
e
a
n
s
u
se
s
to
c
lu
ste
r
t
h
e
p
o
p
u
latio
n
to
k
s
u
b
p
o
p
u
latio
n
s
th
e
n
ta
k
e
m
o
st
c
l
u
ste
rs
re
lev
a
n
t
to
th
e
q
u
e
ry
to
m
a
n
ip
u
late
i
n
a
p
a
ra
ll
e
l
wa
y
b
y
th
e
two
lev
e
l
s
o
f
g
e
n
e
t
ic
p
a
ra
ll
e
l
ism
,
th
u
s,
irrele
v
a
n
t
d
o
c
u
m
e
n
ts
will
n
o
t
b
e
in
c
lu
d
e
d
i
n
su
b
p
o
p
u
latio
n
s,
a
s
a
wa
y
to
imp
r
o
v
e
th
e
q
u
a
li
t
y
o
f
re
su
lt
s.
T
h
re
e
c
o
m
m
o
n
d
a
tas
e
ts
(NLP
,
CIS
I,
a
n
d
CACM)
a
re
u
se
d
to
c
o
m
p
u
te
th
e
re
c
a
ll
,
p
re
c
isio
n
,
a
n
d
F
-
m
e
a
su
re
a
v
e
ra
g
e
s.
F
in
a
ll
y
,
we
c
o
m
p
a
re
d
th
e
p
re
c
isio
n
v
a
lu
e
s
o
f
th
re
e
d
a
tas
e
ts
with
G
e
n
e
ti
c
-
IR
a
n
d
c
las
sic
-
IR.
Th
e
p
ro
p
o
se
d
a
p
p
ro
a
c
h
p
re
c
isio
n
imp
ro
v
e
m
e
n
ts
with
IR
-
G
A
we
r
e
4
5
%
in
th
e
CACM
,
2
7
%
in
th
e
CIS
I,
a
n
d
2
5
%
in
th
e
NLP
.
Wh
i
le,
b
y
c
o
m
p
a
rin
g
wit
h
Clas
sic
-
IR,
(
K
-
m
e
a
n
s)
-
HPG
A g
o
t
4
7
%
i
n
CACM,
2
8
%
i
n
CIS
I,
a
n
d
3
4
%
i
n
NLP
.
K
ey
w
o
r
d
s
:
C
lu
s
ter
-
b
ased
I
R
HPGA
I
n
f
o
r
m
atio
n
r
etr
ie
v
al
K
-
m
ea
n
s
W
eb
d
o
cu
m
en
t c
l
u
s
ter
in
g
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r
th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Sar
ah
Hu
s
s
ein
T
o
m
an
R
o
ad
s
an
d
T
r
an
s
p
o
r
t
Dep
ar
tm
en
t
C
o
lleg
e
o
f
E
n
g
i
n
ee
r
in
g
,
Un
iv
e
r
s
ity
o
f
Al
-
Qad
is
iy
ah
Ad
-
Diwan
iah
,
I
r
a
q
E
m
ail: sar
ah
.
to
m
an
@
q
u
.
ed
u
.
i
q
1.
I
NT
RO
D
UCT
I
O
N
In
th
e
r
ec
en
t
y
ea
r
s
,
th
e
in
f
o
r
m
atio
n
h
as
b
ee
n
o
v
er
lo
a
d
ed
b
ec
au
s
e
o
f
th
e
r
ap
id
g
r
o
wth
o
f
th
e
web
.
T
o
d
ea
l
with
th
is
in
f
o
r
m
atio
n
a
web
d
o
cu
m
en
t
in
f
o
r
m
atio
n
r
e
tr
iev
al
task
is
u
s
ed
to
r
etr
ie
v
e
th
e
m
o
s
t
r
elev
a
n
t
d
o
cu
m
e
n
ts
to
a
u
s
er
q
u
er
y
[
1
,
2
]
.
I
n
f
o
r
m
atio
n
r
etr
iev
al
n
ee
d
s
to
s
ca
n
all
d
o
cu
m
en
ts
th
at
a
r
e
f
o
u
n
d
in
a
d
atab
ase,
th
en
g
iv
e
s
co
r
es
ac
co
r
d
in
g
to
a
r
elev
an
ce
d
eg
r
ee
to
t
h
e
u
s
er
q
u
er
y
,
th
en
r
an
k
all
r
esu
lts
an
d
p
r
esen
t
th
em
to
t
h
e
u
s
er
[
3
,
4
]
.
T
h
u
s
,
in
f
o
r
m
atio
n
r
etr
iev
al
r
eq
u
ir
es
lo
n
g
r
u
n
tim
e
to
s
ca
n
all
d
o
c
u
m
en
t
s
.
T
h
e
clu
s
ter
an
aly
s
is
to
o
l
p
lay
s
a
b
asic
r
o
le
in
i
n
f
o
r
m
atio
n
r
etr
ie
v
al
to
im
p
r
o
v
e
th
e
in
f
o
r
m
atio
n
r
etr
iev
al
p
er
f
o
r
m
an
ce
b
y
r
ed
u
ci
n
g
th
e
s
ea
r
ch
tim
e
an
d
to
p
r
ev
en
t
ir
r
e
lev
an
t
r
esu
lts
f
r
o
m
th
e
r
etr
iev
e
d
d
o
cu
m
e
n
ts
.
T
h
e
id
ea
b
eh
in
d
th
e
web
d
o
cu
m
e
n
t
clu
s
ter
in
g
is
to
ass
ig
n
a
d
atas
et
o
f
web
d
o
cu
m
e
n
ts
to
a
s
et
o
f
clu
s
ter
s
th
at
d
ep
en
d
o
n
th
e
s
im
ilar
ity
’
s
d
eg
r
ee
am
o
n
g
th
em
.
T
h
er
ef
o
r
e,
it
b
e
co
m
es
ea
s
y
f
o
r
s
ea
r
ch
en
g
in
e
s
to
q
u
er
y
in
th
e
s
am
e
clu
s
ter
if
ea
ch
web
p
ag
e
is
ass
ig
n
ed
to
a
s
im
ilar
g
r
o
u
p
[
5
,
6
]
.
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
:
3
4
9
-
3
5
6
350
An
ef
f
icien
t
clu
s
ter
in
g
alg
o
r
it
h
m
an
d
g
en
etic
alg
o
r
ith
m
s
h
o
u
ld
r
ep
r
esen
t
a
d
o
cu
m
e
n
t
as
s
tr
u
ctu
r
ed
d
ata
u
s
in
g
th
e
d
o
c
u
m
en
t
r
ep
r
e
s
en
tatio
n
m
o
d
el.
T
h
e
m
o
s
t
co
m
m
o
n
asp
ec
t
u
s
ed
in
d
o
cu
m
e
n
t
r
ep
r
esen
tatio
n
is
th
e
v
ec
to
r
s
p
ac
e
m
o
d
el
(
VSM)
[
7
]
.
B
esid
es,
a
s
im
ilar
ity
d
eg
r
ee
b
e
twee
n
two
d
o
cu
m
en
ts
o
r
clu
s
ter
s
s
h
o
u
ld
b
e
m
ea
s
u
r
ed
b
y
u
s
in
g
o
n
e
o
f
t
h
e
s
im
ilar
ity
m
ea
s
u
r
es
[
1
]
.
Hier
ar
c
h
ical
an
d
p
ar
titi
o
n
alg
o
r
ith
m
s
ar
e
th
e
m
ajo
r
k
in
d
s
o
f
clu
s
ter
in
g
alg
o
r
ith
m
s
h
av
e
b
ee
n
u
s
ed
[
8
]
.
A
h
ier
a
r
ch
ical
clu
s
ter
in
g
alg
o
r
ith
m
g
e
n
er
ates
a
tr
ee
o
f
c
lu
s
ter
s
(
g
r
o
u
p
s
)
d
ep
e
n
d
in
g
o
n
two
m
eth
o
d
s
.
T
h
e
f
ir
s
t
m
eth
o
d
s
tar
ts
with
o
n
e
clu
s
ter
th
en
m
er
g
e
s
ea
ch
two
s
im
ilar
clu
s
ter
s
,
wh
ich
is
k
n
o
wn
as
th
e
ag
g
lo
m
er
ativ
e
m
eth
o
d
.
T
h
e
s
ec
o
n
d
o
n
e
s
tar
ts
f
r
o
m
th
e
wh
o
le
d
ata
s
et
as
o
n
e
clu
s
ter
th
en
s
p
lit
it
in
to
clu
s
te
r
s
at
ea
ch
s
tag
e,
is
k
n
o
wn
as
th
e
d
iv
is
iv
e
m
eth
o
d
[
9
,
1
0
]
.
A
p
ar
titi
o
n
clu
s
ter
in
g
alg
o
r
ith
m
u
s
es
a
s
in
g
le
s
tep
to
d
iv
id
e
th
e
co
llectio
n
o
f
d
o
c
u
m
en
ts
in
to
p
r
ed
ef
in
ed
n
u
m
b
e
r
o
f
g
r
o
u
p
s
[
1
1
]
.
T
h
e
m
o
s
t
wid
ely
u
s
ed
p
ar
titi
o
n
clu
s
ter
in
g
alg
o
r
ith
m
is
th
e
K
-
m
e
an
s
alg
o
r
ith
m
[
1
2
]
.
I
t
is
an
u
n
s
u
p
er
v
is
ed
lear
n
in
g
alg
o
r
ith
m
th
at
r
elies
o
n
s
elec
tin
g
K
clu
s
ter
s
as
K
-
ce
n
tr
o
id
s
.
Af
ter
t
h
at,
th
e
s
im
ilar
ity
m
e
asu
r
e
is
ca
lcu
lated
b
etwe
en
ea
ch
d
o
cu
m
e
n
t
an
d
th
e
ce
n
tr
o
id
s
,
th
e
n
th
e
d
o
c
u
m
en
t
s
will
ass
ig
n
to
th
e
cl
o
s
est
ce
n
tr
o
id
af
ter
u
p
d
atin
g
o
f
ce
n
tr
o
id
s
m
u
ltip
le
tim
es [
1
3
]
.
I
n
th
e
p
r
esen
t
p
a
p
er
,
th
e
k
-
m
e
an
s
clu
s
ter
with
two
lev
els
o
f
g
en
etic
p
ar
allel
is
u
s
ed
f
o
r
in
f
o
r
m
atio
n
r
etr
iev
al.
Mu
lti
-
d
em
e
p
a
r
allel
g
en
etic
as
f
ir
s
t
lev
el
an
d
m
aster
-
s
lav
e
p
ar
allel
g
en
etic
as
s
ec
o
n
d
lev
el.
T
h
e
id
ea
b
eh
in
d
u
s
in
g
th
e
K
-
m
ea
n
clu
s
ter
in
g
alg
o
r
ith
m
is
to
g
r
o
u
p
a
s
et
o
f
d
o
cu
m
en
ts
to
clu
s
ter
s
ac
co
r
d
in
g
to
th
eir
s
im
ilar
ity
with
a
q
u
er
y
,
th
en
a
n
HPGA
alg
o
r
ith
m
will
p
er
f
o
r
m
a
s
ea
r
ch
in
t
h
e
m
o
s
t
r
elev
an
t
clu
s
ter
s
to
r
ed
u
ce
th
e
s
ea
r
ch
tim
e
an
d
to
p
r
o
v
id
e
o
p
tim
al
s
ea
r
ch
r
esu
lts
.
Nex
t,
at
ea
ch
s
u
b
p
o
p
u
la
tio
n
th
e
r
e
is
a
f
itn
ess
ev
alu
atio
n
p
ar
allelis
m
with
h
y
b
r
id
s
elec
tio
n
a
n
d
two
ch
r
o
m
o
s
o
m
es
cr
o
s
s
o
v
er
as
g
en
etic
o
p
er
ato
r
s
.
T
h
en
m
ig
r
atio
n
am
o
n
g
in
d
iv
id
u
als an
d
r
ep
ea
t H
PGA
s
tep
s
n
tim
e
u
n
til o
b
tain
in
g
th
e
o
p
tim
al
r
esu
lts
.
2.
T
E
R
M
F
R
E
Q
UE
N
CY
–
I
NV
E
RS
E
DO
CU
M
E
N
T
F
R
E
Q
UE
NCY
(
T
F
-
I
DF
)
Data
s
ets
in
m
o
s
t
clu
s
ter
in
g
al
g
o
r
ith
m
s
a
r
e
r
ep
r
esen
ted
b
y
a
s
et
o
f
v
ec
to
r
s
,
V
=
{
V1
,
V2
,
V3
…
Vn
},
wh
er
e,
Vi
is
th
e
f
ea
tu
r
e
v
ec
to
r
o
f
o
n
e
o
b
ject.
T
er
m
Fre
q
u
en
c
y
is
a
s
im
p
le
an
d
ef
f
ec
tiv
e
ter
m
s
elec
tio
n
m
eth
o
d
,
alik
e
wo
r
d
s
ar
e
u
s
ed
i
n
th
e
d
o
c
u
m
en
ts
th
at
b
el
o
n
g
to
th
e
s
am
e
s
u
b
ject,
th
u
s
,
te
r
m
f
r
eq
u
e
n
cy
c
an
b
e
a
r
esp
ec
tab
l
e
in
d
icato
r
f
o
r
a
ce
r
tain
s
u
b
ject.
T
F is
a
ter
m
o
cc
u
r
r
e
n
ce
f
r
eq
u
e
n
cy
in
th
e
d
o
cu
m
en
t a
s
s
h
o
wn
in
(
1
)
.
On
a
n
o
th
er
h
an
d
,
s
o
m
e
ter
m
s
s
h
o
u
ld
b
e
r
em
o
v
ed
s
u
c
h
as
wo
r
d
s
in
t
h
e
s
to
p
lis
t
co
r
r
esp
o
n
d
in
g
to
t
h
e
E
n
g
lis
h
lan
g
u
a
g
e,
b
ec
au
s
e
th
e
o
cc
u
r
r
en
ce
o
f
th
es
e
wo
r
d
s
is
n
o
t r
elev
a
n
t to
id
e
n
tify
th
e
s
u
b
ject
o
f
th
e
d
o
cu
m
e
n
t
[
1
4
]
.
TF(j
,
i
)
=
f
r
e
q
ue
n
c
y o
f
i
th te
r
m
i
n
d
oc
ume
n
t
j
(
1)
T
F
is
n
o
t
ef
f
ec
tiv
e
to
m
ea
s
u
r
e
th
e
f
r
eq
u
e
n
t
ter
m
s
in
a
s
et
o
f
d
o
cu
m
e
n
ts
.
T
h
u
s
,
i
n
v
er
s
e
d
o
cu
m
e
n
t
f
r
e
q
u
en
c
y
(
I
DF)
is
u
s
ed
.
T
DF is th
e
ter
m
f
r
eq
u
e
n
cy
ac
r
o
s
s
a
s
et
o
f
d
o
c
u
m
en
ts
as sh
o
wn
in
(
2
)
.
IDF
(
ti
)
=
l
og
|
D
|
|
D
ti
|
(
2
)
|D|
,
n
um
b
e
r
of
d
oc
ume
n
ts
.
|Dti
|,
n
u
mb
e
r
of
doc
ume
n
ts
tha
t
c
on
ta
in
the
te
r
m
ti.
T
o
d
eter
m
in
e
th
e
weig
h
t
f
o
r
ea
ch
ter
m
ti
in
ea
c
h
d
o
c
u
m
e
n
t
d
j,
T
F
an
d
I
DF
will
b
e
co
m
b
in
ed
b
y
m
u
ltip
licatio
n
o
f
th
e
r
esu
lted
v
alu
es,
T
F
-
I
DF
g
i
v
en
as
s
h
o
w
n
in
(
3
)
[
1
5
]
.
I
n
d
o
cu
m
en
t
clu
s
ter
in
g
,
ter
m
s
with
h
ig
h
er
T
D
-
I
DF h
av
e
b
etter
clu
s
ter
in
g
.
TF
-
IDF
(
ti,
dj)
=
TF (
j,
i
)
*
ID
F
(
ti)
(
3
)
3.
G
E
NE
T
I
C
A
L
G
O
RI
T
H
M
T
h
e
g
e
n
etic
alg
o
r
ith
m
(
GA)
i
s
a
p
r
o
b
ab
ilis
tic
m
eta
-
h
eu
r
is
tic
s
ea
r
ch
alg
o
r
ith
m
in
s
p
ir
ed
b
y
n
atu
r
al
g
en
etics
[
1
6
,
1
7
]
.
GA
g
i
v
es
a
g
o
o
d
s
o
lu
tio
n
in
m
a
n
y
life
f
i
eld
s
.
Fig
u
r
e
1
d
em
o
n
s
tr
at
e
s
th
e
f
lo
wch
a
r
t
o
f
th
e
g
en
etic
alg
o
r
ith
m
s
tep
s
.
T
h
e
b
asic
o
p
er
atio
n
s
o
f
a
g
en
etic
al
g
o
r
ith
m
a
r
e
[
1
8
,
19
]:
−
Gen
er
ate
r
an
d
o
m
s
o
lu
tio
n
s
th
a
t a
r
e
ca
lled
a
p
o
p
u
latio
n
.
−
Dete
r
m
in
e
Fit
n
ess
v
alu
e
to
ev
alu
ate
ea
ch
s
o
lu
tio
n
.
−
Select
th
e
b
est s
o
lu
tio
n
s
ac
co
r
d
in
g
to
t
h
e
f
itn
ess
.
−
Pro
d
u
ce
a
n
ew
p
o
p
u
latio
n
b
y
g
en
etic
o
p
er
at
o
r
s
(
cr
o
s
s
o
v
e
r
a
n
d
m
u
tatio
n
)
.
As
em
p
lo
y
th
e
p
ar
allelis
m
f
ea
tu
r
e
to
r
e
d
u
ce
th
e
p
r
o
ce
s
s
d
u
r
atio
n
.
T
h
er
e
ar
e
th
r
ee
m
o
d
els
o
f
p
ar
allel
g
en
etic
alg
o
r
ith
m
s
(
PGA)
as
ex
h
ib
ited
in
F
ig
u
r
e
2
:
a)
m
ast
er
/s
lav
e
PGA
wh
ich
d
ea
ls
wit
h
s
in
g
le
p
o
p
u
latio
n
an
d
p
ar
allel
f
itn
ess
ca
lcu
latio
n
;
b
)
m
u
lti
d
em
e
PGA
wh
ic
h
d
ea
ls
with
m
u
lti
-
p
o
p
u
latio
n
an
d
p
ar
allel
g
e
n
etic
o
p
er
atio
n
s
f
o
llo
we
d
b
y
m
ig
r
at
io
n
am
o
n
g
th
em
;
c)
ce
llu
lar
w
h
ich
d
ea
ls
with
a
s
in
g
le
p
o
p
u
l
atio
n
r
u
n
n
in
g
o
n
a
p
ar
allel
p
r
o
ce
s
s
in
g
s
y
s
tem
b
as
ed
clo
s
ely
-
lin
k
e
d
m
ass
iv
ely
.
T
h
e
p
r
e
v
io
u
s
m
o
d
els ca
n
b
e
h
y
b
r
id
ized
t
o
p
r
o
d
u
ce
h
ier
ar
ch
ical
PGA
(
HPGA)
m
o
d
els [
2
0
,
2
1
]
.
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
C
lu
s
ter
-
b
a
s
ed
in
fo
r
ma
tio
n
r
etri
ev
a
l b
y
u
s
in
g
(K
-
mea
n
s
)
-
h
iera
r
ch
ica
l p
a
r
a
llel..
.
(
S
a
r
a
h
H
u
s
s
ein
To
ma
n
)
351
Fig
u
r
e
1
.
Gen
etic
alg
o
r
ith
m
s
t
ep
s
(
a)
(
b
)
(
c)
Fig
u
r
e
2
.
(
a
)
Ma
s
ter
/s
lav
e
PGA,
(
b
)
Mu
lti
d
e
m
e
PGA,
(
c)
C
elu
llar
PGA
4.
T
H
E
P
RO
P
O
SE
D
AP
P
RO
A
CH
T
h
e
I
n
f
o
r
m
atio
n
R
etr
iev
al
s
y
s
tem
s
p
r
o
ce
s
s
a
lar
g
e
am
o
u
n
t
o
f
tex
t i
n
d
o
cu
m
e
n
ts
in
d
ex
an
d
u
s
er
q
u
e
r
y
s
tag
es.
Par
alleli
s
m
is
a
way
to
i
m
p
r
o
v
e
t
h
e
q
u
e
r
y
av
e
r
ag
e
tim
e
.
T
h
e
elab
o
r
ate
d
p
r
o
ce
d
u
r
e
u
s
e
s
a
p
ar
allel
g
en
etic
a
lg
o
r
ith
m
(
PGA)
with
K
-
m
ea
n
s
to
r
etr
iev
e
th
e
m
o
s
t r
elev
a
n
t d
o
cu
m
en
ts
to
a
u
s
er
q
u
er
y
th
at
r
elies o
n
th
e
s
tep
s
en
u
m
er
ated
b
elo
w,
Fig
u
r
e
3
p
r
esen
ts
th
e
p
r
o
p
o
s
ed
(K
-
m
ea
n
)
-
HPGA
ap
p
r
o
ac
h
:
4
.
1
.
Web
do
cum
ent
da
t
a
ex
t
ra
ct
io
n
W
eb
p
ag
e
ex
tr
ac
tio
n
r
ep
r
es
en
ts
th
e
in
ter
ac
tio
n
with
w
eb
p
ag
e
s
o
u
r
ce
(
HT
ML
)
to
s
cr
ap
th
e
in
f
o
r
m
atio
n
,
r
esp
ec
tiv
ely
t
o
id
en
tify
s
tr
u
ctu
r
e
d
d
ata
as
a
p
o
s
t
-
p
r
o
ce
s
s
in
g
s
tag
e
th
at
is
co
m
p
o
s
ed
o
f
two
s
tep
s
:
a.
T
r
ee
-
b
ased
ex
tr
ac
tio
n
W
eb
p
ag
es
h
av
e
a
s
em
i
-
s
tr
u
ctu
r
ed
f
ea
tu
r
e,
th
e
r
ef
o
r
e
,
th
is
f
ea
tu
r
e
is
co
n
s
id
er
ed
th
e
m
o
s
t
im
p
o
r
tan
t
f
ea
tu
r
e
to
r
ep
r
esen
t
th
e
HT
ML
tag
s
an
d
tex
t
as
a
lab
eled
tr
ee
,
wh
ich
is
ca
lled
a
d
o
cu
m
e
n
t
o
b
ject
m
o
d
el
(
DOM
)
[
2
2
]
,
an
d
a
d
d
r
e
s
s
in
g
th
e
elem
en
t
'
s
tag
in
th
e
tr
ee
v
ia
XPath
lan
g
u
a
g
e
.
b.
T
ex
t to
k
en
izer
I
ts
p
u
r
p
o
s
e
is
to
b
r
ea
k
t
h
e
tex
t
in
to
k
en
s
,
elim
in
atin
g
s
to
p
w
o
r
d
s
an
d
s
tem
m
er
f
r
o
m
to
k
e
n
s
.
T
h
e
Sto
p
W
o
r
d
lis
t
th
at
we
u
s
ed
,
c
o
n
tai
n
s
1
3
0
0
wo
r
d
s
wh
ich
in
clu
d
e
ar
ticles
(
a,
an
,
th
e)
,
p
r
e
p
o
s
itio
n
s
(
in
,
i
n
to
,
o
n
,
at)
,
co
n
ju
n
ctio
n
s
(
an
d
,
o
r
,
b
u
t,
an
d
s
o
o
n
)
,
p
r
o
n
o
u
n
s
(
s
h
e,
h
e,
I
,
m
e)
,
an
d
o
th
e
r
wo
r
d
s
ir
r
ele
v
an
t
f
o
r
th
e
q
u
e
r
y
p
r
o
ce
s
s
.
Po
r
ter
Stem
m
in
g
is
u
s
ed
in
o
u
r
ap
p
r
o
ac
h
to
en
h
an
c
e
ac
cu
r
ac
y
v
ia
d
r
o
p
p
in
g
m
o
r
p
h
o
lo
g
ical
v
a
r
ian
ts
o
f
wo
r
d
s
.
T
h
u
s
,
to
k
en
s
with
co
m
m
o
n
s
tem
s
s
u
ch
as
-
E
D,
-
I
NG,
-
I
ON,
an
d
-
I
ONS
will h
av
e
s
im
ilar
m
ea
n
in
g
s
.
4
.
2
.
Do
cu
m
ent
a
nd
qu
er
y
re
presenta
t
io
n
I
n
th
is
a
p
p
r
o
a
ch
,
v
ec
to
r
s
p
a
ce
m
o
d
el
(
VSM)
is
u
s
ed
,
a
f
ea
tu
r
es
v
ec
to
r
is
g
en
er
ated
f
r
o
m
ea
ch
d
o
cu
m
e
n
t
co
n
ten
t
an
d
t
h
e
g
i
v
en
q
u
er
y
,
d
ep
e
n
d
in
g
o
n
th
e
o
cc
u
r
r
e
n
ce
o
f
wo
r
d
s
in
th
e
d
o
cu
m
e
n
t
b
y
u
s
in
g
TF
-
I
DF
f
u
n
ctio
n
(
th
e
f
r
e
q
u
en
c
y
o
cc
u
r
r
e
n
ce
o
f
th
e
ter
m
in
th
e
d
o
cu
m
en
t
(
T
F)
w
ith
th
e
f
r
e
q
u
en
cy
o
f
o
cc
u
r
r
en
ce
o
f
th
e
ter
m
in
th
e
d
ata
s
et
o
f
d
o
cu
m
en
ts
(
T
F
-
I
DF)
,
as sh
o
wn
in
(
3
)
.
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
:
3
4
9
-
3
5
6
352
Fig
u
r
e
3
.
(
K
-
m
ea
n
s
)
-
HPGA
ap
p
r
o
ac
h
4
.
3
.
K
-
m
e
a
ns
-
hiera
rc
hica
l p
a
ra
llel g
enet
ic
a
lg
o
ri
t
hm
a
p
pro
a
ch
T
h
e
id
ea
b
eh
in
d
u
s
in
g
th
e
Par
allel
alg
o
r
ith
m
is
to
s
p
lit
th
e
ta
s
k
in
to
a
s
et
o
f
s
u
b
task
s
th
at
w
ill
ex
h
ib
it
a
d
iv
id
e
-
a
n
d
-
c
o
n
q
u
er
b
eh
a
v
io
r
.
I
n
o
u
r
a
p
p
r
o
ac
h
we
u
s
e
m
u
lti
-
d
em
e
p
ar
allel
g
e
n
etic
(
m
u
ltip
le
p
o
p
u
latio
n
)
with
k
-
m
ea
n
s
clu
s
ter
in
g
.
Step
s
b
ello
w
ex
p
lain
th
e
alg
o
r
ith
m
o
p
e
r
atio
n
in
d
etails:
4
.
3
.
1
.
G
ener
a
t
e
po
pu
la
t
io
n
C
r
ea
te
th
e
s
u
b
p
o
p
u
latio
n
s
f
r
o
m
th
e
web
d
o
c
u
m
en
t d
ataset
v
ia
th
e
K
-
m
ea
n
s
alg
o
r
ith
m
.
K
-
m
ea
n
s
s
p
lit
th
e
d
o
cu
m
en
ts
to
b
e
in
d
ex
ed
i
n
to
k
cl
u
s
ter
s
th
en
ev
alu
ate
t
h
e
last
ce
n
tr
o
id
with
a
q
u
er
y
an
d
s
elec
t
ju
s
t c
lu
s
ter
s
th
at
ar
e
n
ea
r
f
r
o
m
th
e
q
u
er
y
.
T
h
e
K
-
m
ea
n
s
s
tep
s
ar
e
d
escr
ip
e
d
b
y
th
e
f
o
llo
win
g
alg
o
r
ith
m
:
K
-
m
ea
n
s
a
lg
o
r
ith
m
Input: D = {d
1
, d
2
, d
3
,…,d
n
}, set of documents.
K: number of clusters.
Output: C = {C
1
, C
2
, C
3
,…,C
k
}, set of
clusters.
Step1: Let centroid c
j
= random number // j= 1,…,k
Step2: Foreach (d
i
in D)
Calculate
CosDistance (
d
i
, c
j
), i = 1,…, n, j = 1,…,k
end
Step3: Assign each document d
i
with
minCosDistance (
d
i
, c
j
) to cluster C
j
Step4: Update centroid c
j
, for all j
Step5: Repeat
(
step2 and step 3) Until (no change in cluster Cj)
Step6: End.
4
.
3
.
2
.
F
it
nes
s
ev
a
lua
t
io
n
T
h
e
s
ec
o
n
d
lev
el
o
f
th
e
p
ar
all
el
alg
o
r
ith
m
is
ap
p
lied
t
o
ev
al
u
ate
th
e
f
itn
ess
f
u
n
ctio
n
in
ea
ch
clu
s
ter
(
s
u
b
p
o
p
u
lati
o
n
)
,
i.e
all
d
o
c
u
m
en
ts
in
th
e
clu
s
ter
will
b
e
e
v
alu
ated
at
th
e
s
am
e
tim
e
u
n
d
er
th
e
s
lav
e/m
aster
p
ar
allel
co
n
ce
p
t.
T
h
is
ev
alu
atio
n
s
tar
ts
b
y
f
o
r
war
d
i
n
g
u
s
er
q
u
er
y
to
ea
c
h
clu
s
ter
th
en
c
alcu
late
th
e
f
itn
ess
f
u
n
ctio
n
to
ea
c
h
d
o
cu
m
en
t
o
f
th
e
clu
s
ter
.
I
n
th
e
p
r
esen
t
ap
p
r
o
ac
h
,
a
c
o
s
in
e
s
im
ilar
ity
f
u
n
ctio
n
is
u
s
ed
as
a
f
itn
ess
f
u
n
ctio
n
[
2
3
]
.
T
h
e
co
s
i
n
e
s
im
ilar
ity
f
u
n
ctio
n
is
g
iv
en
in
(
4
)
.
Sco
s
=
∑
Pi
Qi
n
i
=
1
√
∑
Pi
2
n
i
=
1
√
∑
Qi
2
n
i
=
1
(
4
)
4
.
3
.
3
.
G
enet
ic
o
pera
t
o
rs
Gen
er
ate
a
n
ew
p
o
p
u
latio
n
b
y
ap
p
ly
in
g
g
en
etic
o
p
er
ato
r
s
(
s
elec
tio
n
an
d
cr
o
s
s
o
v
er
)
.
T
o
im
p
r
o
v
e
g
en
etic
p
er
f
o
r
m
a
n
ce
,
we
m
o
v
e
4
%
o
f
c
h
r
o
m
o
s
o
m
es
with
th
e
h
ig
h
est
p
r
o
b
ab
ilit
y
in
th
e
n
e
x
t
g
en
er
atio
n
with
o
u
t
ch
an
g
e
(
i.e
.
ap
p
l
y
elitis
m
f
ea
tu
r
e
)
.
Gen
etic
o
p
er
ato
r
s
in
(
K
-
m
ea
n
s
)
-
HPGA
f
lo
w
th
e
f
o
llo
win
g
s
tep
s
:
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
C
lu
s
ter
-
b
a
s
ed
in
fo
r
ma
tio
n
r
etri
ev
a
l b
y
u
s
in
g
(K
-
mea
n
s
)
-
h
iera
r
ch
ica
l p
a
r
a
llel..
.
(
S
a
r
a
h
H
u
s
s
ein
To
ma
n
)
353
a.
C
alcu
late
th
e
p
r
o
b
ab
ilit
y
f
o
r
e
ac
h
ch
r
o
m
o
s
o
m
e,
wh
er
e
P[i]
=
Fit
n
ess
[
i]
/To
tal
b.
R
an
k
th
e
Pro
b
ab
ilit
y
v
alu
es
a
n
d
tak
e
th
e
to
p
4
%
E
liti
s
m
to
av
o
id
th
e
l
o
s
s
o
f
f
ittes
t
ch
r
o
m
o
s
o
m
es
in
th
e
n
ew
p
o
p
u
latio
n
.
c.
Hy
b
r
id
r
o
u
lette
-
to
u
r
n
am
en
t
s
elec
tio
n
(
HR
T
S):
I
t
is
th
e
p
r
o
ce
s
s
o
f
s
elec
tin
g
a
p
air
o
f
p
ar
en
ts
f
r
o
m
th
e
p
o
p
u
latio
n
to
em
p
h
asize
f
itter
o
f
f
s
p
r
in
g
s
in
a
n
ew
p
o
p
u
latio
n
.
I
n
o
u
r
a
p
p
r
o
ac
h
we
u
s
ed
a
h
y
b
r
id
m
et
h
o
d
to
tak
e
ad
v
an
tag
e
o
f
b
o
th
s
elec
tio
n
m
e
th
o
d
s
(
r
o
u
lette
wh
ee
l
an
d
to
u
r
n
a
m
en
t
)
.
T
h
e
s
elec
tio
n
p
r
o
ce
s
s
is
ex
p
lain
ed
b
y
th
e
f
o
llo
win
g
al
g
o
r
ith
m
:
HR
T
S
a
lg
o
r
ith
m
Output: parent1,
parent2
End
d.
C
r
o
s
s
o
v
er
o
p
er
ati
o
n
aim
s
t
o
g
et
b
etter
o
f
f
s
p
r
in
g
b
y
g
en
er
ati
n
g
a
n
ew
c
h
ild
f
r
o
m
two
s
ele
cted
p
ar
e
n
ts
.
I
n
th
is
ap
p
r
o
ac
h
,
we
p
r
o
p
o
s
ed
to
r
ep
r
esen
t
th
e
p
o
p
u
latio
n
as
a
m
atr
ix
,
ea
ch
ch
r
o
m
o
s
o
m
e
v
ec
to
r
r
ep
r
esen
tin
g
a
r
o
w
in
t
h
e
m
atr
ix
,
th
en
s
elec
t
two
r
an
d
o
m
p
o
s
itio
n
s
in
th
e
r
an
g
e
[
1
,
v
ec
to
r
_
len
g
th
]
.
T
h
e
cr
o
s
s
o
v
er
is
d
escr
ib
ed
b
y
th
e
f
o
llo
win
g
a
lg
o
r
ith
m
:
T
wo
ch
r
o
m
o
s
o
m
es c
r
o
s
s
o
v
er
alg
o
r
ith
m
Input: subP = subP
-
Elite Count.
Output: offsprings
Begin
subP_length = length(subP);
repeat
Call selection function to select two parents;
10 Call pickTwoPosition (subP_length);
Exchange two positions betweentwo selected parents;
until (index <= subPsize) Goto 10;
End
function [ position1, position2 ] = pickTwoPosition (subP_length)
r = randi([1, subP_length],2)// generate 2 random integer nu
mbers to vector r
position1 = r(1);
position2 = r(2);
end
4
.
3
.
4
.
M
ig
ra
t
io
n
Mig
r
atio
n
i
s
th
e
s
y
n
ch
r
o
n
o
u
s
p
r
o
ce
s
s
th
at
m
ea
n
s
th
e
ex
c
h
an
g
in
g
o
f
m
e
m
eb
er
s
.
It
wai
ts
f
o
r
th
e
ev
alu
atio
n
o
f
all
ch
r
o
m
o
s
o
m
es
in
all
s
u
b
p
o
p
u
latio
n
s
to
ex
c
h
a
n
g
e
th
e
in
d
iv
i
d
u
als.
Mig
r
at
io
n
h
as
an
in
ter
v
al
th
at
is
s
et
to
5
in
o
u
r
a
p
p
r
o
ac
h
.
4
.
4
.
5
.
T
er
m
ina
t
e
T
h
e
last
s
tep
in
o
u
r
ap
p
r
o
ac
h
is
r
ep
ea
t
in
g
th
e
p
r
ev
i
o
u
s
s
t
ep
s
(
f
r
o
m
f
itn
ess
to
m
ig
r
atio
n
)
.
T
h
ese
s
tep
s
will
r
ea
p
ea
t
n
tim
es,
wh
e
r
e
n
is
th
e
s
ize
o
f
th
e
p
o
p
u
latio
n
.
Af
ter
co
m
p
lete
t
h
e
r
e
p
ea
tio
n
,
th
e
d
o
c
u
m
en
ts
will
r
an
k
ac
c
o
r
d
in
g
to
f
itn
ess
p
r
o
b
ab
ilit
y
v
alu
es.
T
h
en
th
e
b
est
r
esu
lts
will
h
av
e
r
etr
iev
ed
f
r
o
m
th
e
d
o
cu
m
en
ts
th
at
h
av
e
h
ieg
h
er
r
a
n
k
.
5.
E
XP
E
R
I
M
E
N
T
A
L
RE
SUL
T
S
T
h
r
ee
d
atasets
wer
e
u
s
ed
f
o
r
ex
p
er
im
en
tal
r
esu
lts
.
N
PL
d
at
aset
(
DS1
)
co
n
tain
in
g
1
1
,
4
2
9
elec
tr
o
n
ic
en
g
in
ee
r
in
g
d
o
cu
m
en
ts
,
C
I
SI
d
ataset
(
DS2
)
with
1
,
4
6
0
co
m
p
u
ter
s
cien
ce
d
o
cu
m
e
n
ts
an
d
C
AC
M
d
ataset
(
D
S3
)
co
n
s
is
tin
g
o
f
3204
c
o
m
m
u
n
ica
tio
n
s
d
o
cu
m
en
ts
.
T
o
ev
alu
ate
t
h
e
web
d
o
cu
m
en
ts
r
etr
iev
al
,
th
e
r
ec
all
,
p
r
ec
is
io
n
,
an
d
F
-
m
ea
s
u
r
e
a
r
e
u
s
ed
f
o
r
1
0
0
q
u
er
ies in
t
h
r
ee
d
atasets
as d
ef
in
ed
in
th
e
f
o
llo
win
g
eq
u
ati
o
n
s
[
2
4
,
25
]:
R
ec
a
ll (
R
)
=
r
el
ev
a
n
t
i
t
em
s
r
et
r
i
ev
ed
r
el
ev
an
t
i
t
em
s
(
5
)
Begin
for j = 1 : 2
r = randi[1, pop
size
] //Select random number for subpopulation
for i = 1 : r
sum
fitness =
sum (fitness)
P
sum
= randi[1, sum
fitness
];
S = 0; index = 1;
S = S + fitness[i]; index++;
if (s < P
sum)
goto 10, else subPop[i] = current chromosome
end
Parent[j] = maxFitness(subPop)
// select parent
end
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
:
3
4
9
-
3
5
6
354
P
r
ec
is
io
n
(
P
)
=
r
el
ev
an
t
i
t
em
s
r
et
r
i
ev
ed
r
et
r
i
ev
ed
i
t
em
s
(
6
)
F
-
m
ea
s
u
r
e
=
2
.
.
+
(
7
)
T
h
e
r
esu
lts
ar
e
s
h
o
w
n
in
T
ab
l
es 1
-
3
.
Fo
r
th
e
NPL
d
ataset
(
DS1
)
wh
er
e
p
r
ec
is
io
n
a
v
er
ag
e
is
0
.
6
8
8
8
8
9
an
d
F
-
m
ea
s
u
r
e
a
v
er
ag
e
is
2
.
0
6
6
7
,
wh
ile
in
t
h
e
C
I
SI
d
ataset
(
DS2
)
,
th
e
p
r
ec
is
io
n
a
v
er
ag
e
was
0
.
6
5
8
8
9
an
d
th
e
F
-
m
ea
s
u
r
e
av
er
a
g
e
was
1
.
9
7
6
6
7
.
Fin
ally
,
th
e
C
AC
M
d
ataset
(
DS3
)
th
e
av
e
r
ag
e
f
o
r
p
r
ec
is
io
n
an
d
F
-
m
ea
s
u
r
e
w
er
e
0
.
7
4
8
8
8
9
an
d
4
5
.
2
2
2
2
2
r
esp
ec
tiv
ely
.
Af
ter
t
h
e
an
aly
s
is
of
th
e
p
r
ev
io
u
s
r
esu
lts
,
th
e
th
ir
d
d
ataset
g
av
e
h
ig
h
er
r
esu
lts
in
b
o
th
m
e
asu
r
es.
T
ab
le
1
.
T
h
e
r
esu
lts
o
f
r
ec
all,
p
r
ec
is
io
n
an
d
F
-
m
ea
s
u
r
e
f
o
r
1
0
0
q
u
er
y
in
N
PL
d
ataset
(
DS1
)
b
y
u
s
in
g
(K
-
m
ea
n
s
)
-
HPGA
ap
p
r
o
ac
h
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
F
-
mea
su
r
e
%
0
.
1
0
.
9
2
.
7
0
.
2
0
.
8
7
2
.
6
1
0
.
3
0
.
8
4
2
.
5
2
0
.
4
0
.
7
7
2
.
3
1
0
.
5
0
.
7
4
2
.
2
2
0
.
6
0
.
6
6
1
.
9
8
0
.
7
0
.
5
8
1
.
7
4
0
.
8
0
.
4
6
1
.
3
8
0
.9
0
.
3
8
1
.
1
4
AVG
0
.
6
8
8
8
2
.
0
6
6
6
T
ab
le
2
.
T
h
e
r
esu
lts
o
f
r
ec
all,
p
r
ec
is
io
n
an
d
F
-
m
ea
s
u
r
e
f
o
r
1
0
0
q
u
er
y
in
C
I
SI
d
ataset
(
DS2
)
b
y
u
s
in
g
(K
-
m
ea
n
s
)
-
HPGA
ap
p
r
o
ac
h
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
F
-
mea
su
r
e
%
0
.
1
0
.
8
9
2
.
6
7
0
.
2
0
.
8
4
2
.
5
2
0
.
3
0
.
7
8
2
.
3
4
0
.
4
0
.
7
6
2
.
2
8
0
.
5
0
.
6
9
2
.
0
7
0
.
6
0
.
5
5
1
.
6
5
0
.
7
0
.
5
1
1
.
5
3
0
.
8
0
.
4
7
1
.
4
1
0
.
9
0
.
4
4
1
.
3
2
AVG
0
.
6
5
8
8
19
.
766
T
ab
le
3
.
D
is
p
lay
s
th
e
r
esu
lts
o
f
r
ec
all,
p
r
ec
is
io
n
a
n
d
F
-
m
ea
s
u
r
e
f
o
r
1
0
0
q
u
er
y
in
C
AC
M
d
ataset
(
DS3
)
b
y
u
s
in
g
(K
-
m
ea
n
s
)
-
HPGA
ap
p
r
o
ac
h
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
F
-
mea
su
r
e
%
0
.
1
0
.
9
4
2
.
8
2
0
.
2
0
.
9
2
.
7
0
.
3
0
.
8
7
2
.
6
1
0
.
4
0
.
8
5
2
.
5
5
0
.
5
0
.
8
2
.
4
0
.
6
0
.
7
7
2
.
3
1
0
.
7
0
.
6
6
1
.
9
8
0
.
8
0
.
5
4
1
.
6
2
0
.
9
0
.
4
1
1
.
2
3
AVG
0
.
7
4
8
8
22
.
466
We
m
ea
s
u
r
ed
th
e
i
m
p
r
o
v
em
e
n
ts
t
h
at
wer
e
a
ch
iev
e
d
b
y
th
e
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
,
with
a
p
r
ec
is
io
n
o
f
in
f
o
r
m
atio
n
r
etr
ie
v
al
b
y
g
e
n
e
tic
alg
o
r
ith
m
(
GA
-
I
R
)
f
o
r
t
h
r
ee
d
atas
ets.
T
ab
les
4
-
6
p
r
esen
ts
a
co
m
p
a
r
is
o
n
b
etwe
en
o
u
r
ap
p
r
o
ac
h
an
d
GA
-
IR
.
I
m
p
r
o
v
em
e
n
t
av
er
ag
e
is
ca
lcu
lated
f
o
r
t
h
r
ee
d
atasets
an
d
th
e
r
esu
lts
wer
e
2
5
.
6
6
6
6
,
2
7
.
4
4
4
4
,
a
n
d
4
5
.
2
2
2
2
r
esp
ec
tiv
el
y
.
Fin
ally
,
we
co
m
p
a
r
ed
t
h
e
p
r
o
p
o
s
ed
ap
p
r
o
a
ch
with
class
ic
I
n
f
o
r
m
atio
n
R
etr
iev
al
(
class
ic
-
IR
)
p
r
ec
is
io
n
an
d
th
e
im
p
r
o
v
e
m
en
ts
wer
e
3
4
.
4
4
4
4
%
in
NL
P,
2
8
.
6
6
6
6
%
in
C
I
SI
,
an
d
4
7
% in
C
AC
M
as sh
o
wn
i
n
T
ab
les
7
-
9.
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
an
aly
s
is
o
f
(K
-
m
ea
n
s
)
-
HPGA
a
p
p
r
o
ac
h
an
d
GA
[
2
6
]
in
NPL
d
ataset
(
DS1
)
T
ab
le
5
.
C
o
m
p
a
r
is
o
n
an
aly
s
is
o
f
(K
-
m
ea
n
s
)
-
HPGA
ap
p
r
o
ac
h
an
d
GA
[
2
6
]
in
C
I
SI
d
ataset
(
DS2
)
R
e
c
a
l
l
H
P
G
A
-
(K
-
mea
n
s)
(
p
)
GA
-
I
R
(
P
)
I
mp
r
o
v
e
m
e
n
t
s
%
0
.
1
0
.
9
0
.
8
8
2
0
.
2
0
.
8
7
0
.
6
6
21
0
.
3
0
.
8
4
0
.
5
9
25
0
.
4
0
.
7
7
0
.
4
4
33
0
.
5
0
.
7
4
0
.
4
34
0
.
6
0
.
6
6
0
.
3
1
35
0
.
7
0
.
5
8
0
.
2
7
31
0
.
8
0
.
4
6
0
.
1
9
27
0
.
9
0
.
3
8
0
.
1
5
23
AVG
0
.
6
8
8
8
0
.
4
3
2
2
2
5
6
.
6
6
6
R
e
c
a
l
l
H
P
G
A
-
(K
-
mea
n
s)
(
p
)
GA
-
I
R
(
P
)
I
mp
r
o
v
e
m
e
n
t
s
%
0
.
1
0
.
8
9
0
.
8
9
0
.
2
0
.
8
4
0
.
5
5
29
0
.
3
0
.
7
8
0
.
4
8
30
0
.
4
0
.
7
6
0
.
3
9
37
0
.
5
0
.
6
9
0
.
3
6
33
0
.
6
0
.
5
5
0
.
2
8
27
0
.
7
0
.
5
1
0
.
2
4
27
0
.
8
0
.
4
7
0
.
2
27
0
.
9
0
.
4
4
0
.
1
6
28
AVG
0
.
6
5
8
8
0
.
3
8
4
4
2
7
4
.
4
4
4
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
C
lu
s
ter
-
b
a
s
ed
in
fo
r
ma
tio
n
r
etri
ev
a
l b
y
u
s
in
g
(K
-
mea
n
s
)
-
h
iera
r
ch
ica
l p
a
r
a
llel..
.
(
S
a
r
a
h
H
u
s
s
ein
To
ma
n
)
355
T
ab
le
6
.
C
o
m
p
a
r
is
o
n
an
aly
s
is
o
f
(K
-
m
ea
n
s
)
-
HPGA
a
p
p
r
o
ac
h
an
d
GA
[
2
6
]
i
n
C
AC
M
d
ataset
(
DS3
)
R
e
c
a
l
l
H
P
G
A
-
(K
-
mea
n
s)
(
p
)
GA
-
IR
(P)
I
mp
r
o
v
e
m
e
n
t
s
%
0
.
1
0
.
9
4
0
.
7
9
15
0
.
2
0
.
9
0
.
4
7
43
0
.
3
0
.
8
7
0
.
4
2
45
0
.
4
0
.
8
5
0
.
2
7
58
0
.
5
0
.
8
0
.
2
3
57
0
.
6
0
.
7
7
0
.
1
6
61
0
.
7
0
.
6
6
0
.
1
4
52
0
.
8
0
.
5
4
0
.
1
44
0
.
9
0
.
4
1
0
.
0
9
32
AVG
0
.
7
4
8
8
0
.
2
9
6
6
4
5
2
.
2
2
2
T
ab
le
7
.
C
o
m
p
a
r
is
o
n
an
aly
s
is
o
f
(K
-
m
ea
n
s
)
-
HPGA
a
p
p
r
o
ac
h
an
d
class
ic
I
R
[
2
0
]
i
n
NPL
d
ataset
(
DS1
)
R
e
c
a
ll
H
P
G
A
-
(K
-
mea
n
s)
(
p
)
C
l
a
s
si
c
I
R
(P)
I
mp
r
o
v
e
m
e
n
t
s
%
0
.
1
0
.
9
0
.
7
3
17
0
.
2
0
.
8
7
0
.
5
37
0
.
3
0
.
8
4
0
.
4
4
40
0
.
4
0
.
7
7
0
.
3
4
43
0
.
5
0
.
7
4
0
.
3
1
43
0
.
6
0
.
6
6
0
.
2
4
42
0
.
7
0
.
5
8
0
.
2
2
36
0
.
8
0
.
4
6
0
.
1
7
29
0
.
9
0
.
3
8
0
.
1
5
23
AVG
0
.
6
8
8
8
0
.
3
4
4
4
3
4
4
.
4
4
4
T
ab
le
8
.
C
o
m
p
a
r
is
o
n
b
etwe
en
(K
-
m
ea
n
s
)
-
HPGA
a
p
p
r
o
ac
h
an
d
class
ic
I
R
[
2
0
]
i
n
C
I
SI
d
ataset
(
DS2
)
R
e
c
a
ll
H
P
G
A
-
(K
-
mea
n
s)
(
p
)
C
l
a
s
si
c
I
R
(P)
I
mp
r
o
v
e
m
e
n
t
s
%
0
.
1
0
.
8
9
0
.
6
8
21
0
.
2
0
.
8
4
0
.
5
6
28
0
.
3
0
.
7
8
0
.
4
6
32
0
.
4
0
.
7
6
0
.
4
36
0
.
5
0
.
6
9
0
.
3
5
34
0
.
6
0
.
5
5
0
.
3
25
0
.
7
0
.
5
1
0
.
2
5
26
0
.
8
0
.
4
7
0
.
2
27
0
.
9
0
.
4
4
0
.
1
5
29
AVG
0
.
6
5
8
8
0
.
3
7
2
2
2
8
6
.
6
6
6
T
ab
le
9
.
C
o
m
p
a
r
is
o
n
an
aly
s
is
o
f
(K
-
m
ea
n
s
)
-
HPGA
a
p
p
r
o
ac
h
an
d
class
ic
I
R
[
2
0
]
i
n
C
AC
M
d
ataset
(
DS3
)
R
e
c
a
l
l
H
P
G
A
-
(K
-
mea
n
s)
(
p
)
C
l
a
s
si
c
I
R
(
P
)
I
mp
r
o
v
e
m
e
n
t
s
%
0
.
1
0
.
9
4
0
.
7
2
22
0
.
2
0
.
9
0
.
4
5
45
0
.
3
0
.
8
7
0
.
3
7
50
0
.
4
0
.
8
5
0
.
2
5
60
0
.
5
0
.
8
0
.
2
2
58
0
.
6
0
.
7
7
0
.
1
6
61
0
.
7
0
.
6
6
0
.
1
4
52
0
.
8
0
.
5
4
0
.
1
1
43
0
.
9
0
.
4
1
0
.
0
9
32
AVG
0
.
7
4
8
8
0
.
2
7
8
8
47
6.
CO
NCLU
SI
O
NS
Af
ter
th
e
test
s
an
d
r
esear
ch
f
o
r
th
is
p
a
p
er
,
we
c
o
n
clu
d
e
d
an
in
f
o
r
m
atio
n
r
etr
iev
al
p
e
r
f
o
r
m
an
ce
im
p
r
o
v
em
e
n
t:
(
K
-
m
ea
n
s
)
-
HPGA
ac
h
iev
ed
h
ig
h
er
p
r
ec
is
io
n
an
d
b
etter
q
u
ality
in
d
o
c
u
m
e
n
t
r
etr
iev
al.
Als
o
a
r
ed
u
ctio
n
o
f
ir
r
elev
a
n
t
r
esu
lts
in
u
s
er
s
ea
r
ch
was
o
b
s
er
v
e
d
.
Ou
r
r
esu
lts
wer
e
d
eter
m
in
ed
b
y
co
m
p
ar
in
g
th
r
ee
co
m
m
o
n
d
atasets
(
NL
P,
C
I
SI,
an
d
C
AC
M)
with
class
ic
I
R
an
d
GA.
T
h
e
r
an
g
e
o
f
p
r
ec
is
io
n
im
p
r
o
v
e
m
en
ts
f
o
r
th
r
ee
d
atasets
with
class
ic
-
I
R
was
(
28
-
47%
)
w
h
ile
with
GA
-
I
R
th
e
p
r
ec
is
io
n
was
(
25
-
45%
)
.
RE
F
E
R
E
NC
E
S
[1
]
C.
D.
M
a
n
n
i
n
g
,
P
.
Ra
g
a
h
v
a
n
,
a
n
d
H.
S
c
h
u
tze
,
“
An
in
t
ro
d
u
c
ti
o
n
t
o
in
fo
rm
a
ti
o
n
re
tri
e
v
a
l
,
”
Ca
m
b
ri
d
g
e
Un
ive
rs
it
y
Pre
ss
,
2
0
0
9
.
[2
]
J.
M
.
Ka
ss
im a
n
d
M
.
Ra
h
m
a
n
y
,
“
In
tro
d
u
c
ti
o
n
t
o
S
e
m
a
n
ti
c
S
e
a
rc
h
En
g
i
n
e
,
”
2
0
0
9
In
t.
C
o
n
f.
El
e
c
tr.
E
n
g
.
In
fo
rm
a
t
ics
,
v
o
l.
2
,
p
p
.
3
8
0
-
3
8
6
,
Au
g
u
st
2
0
0
9
.
[3
]
S
.
M
.
Weiss
,
N.
I
n
d
u
rk
h
y
a
,
T.
Z
h
a
n
g
,
a
n
d
F
.
J.
Da
m
e
ra
u
,
“
In
fo
r
m
a
ti
o
n
re
tri
e
v
a
l
a
n
d
tex
t
m
in
i
n
g
,
”
S
p
rin
g
e
r
Ber
li
n
He
id
e
lb
,
p
p
.
7
5
-
9
0
,
2
0
1
0
.
[4
]
Y.
Wan
g
,
“
De
sig
n
o
f
i
n
fo
rm
a
ti
o
n
re
tri
e
v
a
l
s
y
ste
m
u
sin
g
ro
u
g
h
fu
z
z
y
se
t
,
”
T
EL
KOM
NIKA
T
e
l
e
c
o
mm
u
n
ica
ti
o
n
Co
mp
u
t
in
g
El
e
c
tro
n
ics
a
n
d
C
o
n
tr
o
l
,
v
o
l
.
1
2
,
n
o
.
1
,
p
p
.
8
4
4
-
8
5
1
,
Ja
n
u
a
ry
2
0
1
4
.
[5
]
Y.
Dje
n
o
u
r
i
a
n
d
e
t
a
l
,
”
F
a
st
a
n
d
e
ffe
c
ti
v
e
c
lu
ste
r
-
b
a
se
d
in
f
o
rm
a
ti
o
n
re
tri
e
v
a
l
u
sin
g
fre
q
u
e
n
t
c
l
o
se
d
it
e
m
se
t
s
,
”
In
fo
rm
a
t
io
n
S
c
ien
c
e
s,
v
o
l
.
4
5
3
,
p
p
.
1
5
4
-
1
6
7
,
J
u
ly
2
0
1
8
.
[6
]
C.
Co
b
o
s
,
e
t
a
l
.
,
“
Web
d
o
c
u
m
e
n
t
c
lu
ste
rin
g
b
a
se
d
o
n
G
lo
b
a
l
-
Be
st
Ha
rm
o
n
y
S
e
a
rc
h
,
K
-
m
e
a
n
s
,
fre
q
u
e
n
t
term
se
ts
a
n
d
b
a
y
e
sia
n
in
fo
rm
a
ti
o
n
c
rit
e
rio
n
,
”
I
EE
E
,
Au
g
u
st
2
0
1
0
.
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
:
3
4
9
-
3
5
6
356
[7
]
S
.
E.
P
ra
tam
a
,
e
t
a
l,
“
Weig
h
ted
in
v
e
rse
d
o
c
u
m
e
n
t
fre
q
u
e
n
c
y
a
n
d
v
e
c
to
r
sp
a
c
e
m
o
d
e
l
fo
r
h
a
d
i
th
s
e
a
rc
h
e
n
g
in
e
,
”
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
i
n
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
1
8
,
n
o
.
2
,
p
p
.
1
0
0
4
-
1
0
1
4
,
M
a
y
2
0
2
0
.
[8
]
R.
F.
Ha
ss
a
n
a
n
d
e
t
a
l
.
,
“
Im
p
r
o
v
i
n
g
th
e
we
b
in
d
e
x
in
g
q
u
a
li
t
y
th
ro
u
g
h
a
we
b
site
-
se
a
rc
h
e
n
g
i
n
e
c
o
a
c
ti
o
n
s
,
”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
C
o
mp
u
ter
a
n
d
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l.
3
,
n
o
.
2
,
M
a
rc
h
2
0
1
4
.
[9
]
S
.
S
.
Tan
d
e
l,
A.
Ja
m
a
d
a
r,
a
n
d
S
.
Du
d
u
g
u
,
“
A
su
rv
e
y
o
n
te
x
t
m
i
n
in
g
tec
h
n
iq
u
e
s
,
”
5
t
h
I
n
t.
C
o
n
f.
Ad
v
.
Co
m
p
u
t
.
Co
mm
u
n
.
S
y
st
,
p
p
.
1
0
2
2
-
1
0
2
6
,
M
a
rc
h
2
0
1
9
.
[1
0
]
S.
H.
T
o
m
a
n
,
e
t
a
l
.
,
“
Co
n
ten
t
-
b
a
se
d
a
u
d
i
o
re
tri
e
v
a
l
b
y
u
sin
g
e
li
ti
s
m
GA
-
K
NN
a
p
p
ro
a
c
h
”
,
J
o
u
rn
a
l
o
f
AL
-
Q
a
d
isiy
a
h
fo
r c
o
mp
u
ter
sc
ien
c
e
a
n
d
ma
t
h
e
m
a
ti
c
s
,
v
o
l.
9,
n
o.
1,
M
a
y
2
0
1
7
.
[1
1
]
A.
Ko
n
a
r
,
“
Artifi
c
ial
i
n
telli
g
e
n
c
e
a
n
d
so
ft
c
o
m
p
u
ti
n
g
:
b
e
h
a
v
io
ra
l
a
n
d
c
o
g
n
i
ti
v
e
m
o
d
e
li
n
g
o
f
t
h
e
h
u
m
a
n
b
ra
in
,
”
J
a
d
a
v
p
u
r Un
ive
rs
it
y
,
CRC
Pre
ss
L
L
C
,
2
0
0
0
.
[1
2
]
T.
M
u
n
a
k
a
ta,
“
F
u
n
d
a
m
e
n
tals
o
f
t
h
e
n
e
w artifi
c
ial
i
n
telli
g
e
n
c
e
,
”
S
p
rin
g
e
r
,
2
0
0
8
.
[1
3
]
C.
C.
Ag
g
a
rwa
l
a
n
d
C
.
Xh
a
i,
“
A s
u
rv
e
y
o
f
te
x
t
c
lu
ste
ri
n
g
a
l
g
o
r
it
h
m
s
,
”
M
in
i
n
g
tex
t
d
a
ta
,
Au
g
u
st
2
0
1
2
.
[1
4
]
A.
M
.
S
ire
g
a
r
a
n
d
A.
P
u
sp
a
b
h
u
a
n
a
,
“
Im
p
ro
v
e
m
e
n
t
o
f
term
we
ig
h
t
re
su
lt
i
n
th
e
i
n
fo
rm
a
ti
o
n
re
t
riev
a
l
sy
ste
m
s,”
Pro
c
e
e
d
in
g
s
o
f
4
th
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Ne
w
M
e
d
ia
S
tu
d
ie
s
,
No
v
e
m
b
e
r
2
0
1
7
.
[1
5
]
T.
To
k
u
n
a
g
a
,
T.
T
o
k
u
n
a
g
a
,
I.
M
a
k
o
to
,
a
n
d
I
.
M
a
k
o
to
,
“
Tex
t
c
a
teg
o
riza
ti
o
n
b
a
se
d
o
n
we
ig
h
ted
in
v
e
rse
d
o
c
u
m
e
n
t
fre
q
u
e
n
c
y
,
”
S
p
e
c
.
I
n
ter
e
s.
Gr
o
u
p
s In
f.
Pro
c
e
ss
S
o
c
.
J
a
p
a
n
(
S
IG
-
IP
S
J,
1
9
9
4
.
[1
6
]
P
.
S
imo
n
a
n
d
S
.
S
.
S
a
t
h
y
a
,
“
Ge
n
e
ti
c
a
lg
o
rit
h
m
fo
r
i
n
fo
rm
a
t
io
n
re
tri
e
v
a
l
,”
2
0
0
9
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
In
telli
g
e
n
t
A
g
e
n
t
&
M
u
l
ti
-
Ag
e
n
t
S
y
ste
ms
,
Ju
ly
2
0
0
9
.
[1
7
]
Z.
Wan
g
a
n
d
B.
F
e
n
g
,
"
Op
ti
m
a
l
g
e
n
e
ti
c
q
u
e
ry
a
lg
o
rit
h
m
f
o
r
i
n
fo
rm
a
ti
o
n
re
tri
e
v
a
l
,”
S
p
rin
g
e
r
,
2
0
0
9
.
[1
8
]
H.
Im
ra
n
,
“
G
e
n
e
ti
c
a
lg
o
ri
th
m
b
a
se
d
m
o
d
e
l
fo
r
e
ffe
c
ti
v
e
d
o
c
u
m
e
n
t
re
tri
e
v
a
l
,
”
In
telli
g
e
n
t
C
o
n
tr
o
l
a
n
d
Co
mp
u
ter
En
g
i
n
e
e
rin
g
,
2
0
1
1
.
[1
9
]
P
.
P
a
th
a
k
,
M
.
G
o
rd
o
n
a
n
d
W.
F
a
n
,
“
Eff
e
c
ti
v
e
in
fo
rm
a
ti
o
n
re
tri
e
v
a
l
u
sin
g
g
e
n
e
ti
c
a
lg
o
rit
h
m
s
b
a
se
d
m
a
tch
in
g
.”
Ha
wa
i
i
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
S
y
ste
m S
c
ien
c
e
s
,
IEE
E,
F
e
b
ru
a
ry
2
0
0
0
.
[2
0
]
M
.
Eb
ra
h
imi
a
n
d
A.
Ja
h
a
n
g
ir
ian
,
“
A
h
iera
rc
h
ica
l
p
a
ra
ll
e
l
stra
teg
y
fo
r
a
e
ro
d
y
n
a
m
ic
sh
a
p
e
o
p
ti
m
iza
ti
o
n
with
g
e
n
e
ti
c
a
lg
o
rit
h
m
,”
S
c
ien
ti
a
Ir
a
n
ica
,
v
o
l.
22
,
n
o
.
6
,
p
p
.
2
3
7
9
-
2
3
8
8
,
Ja
n
u
a
r
y
2
0
1
5
.
[2
1
]
K.
I.
Ab
u
z
a
n
o
u
n
e
h
,
“
P
a
ra
ll
e
l
a
n
d
d
istri
b
u
ted
g
e
n
e
ti
c
a
lg
o
r
it
h
m
wit
h
m
u
lt
i
p
le
-
o
b
jec
ti
v
e
s
t
o
imp
r
o
v
e
a
n
d
d
e
v
e
lo
p
o
f
e
v
o
lu
ti
o
n
a
ry
a
lg
o
rit
h
m
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Ad
v
a
n
c
e
d
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
7
,
n
o
.
5
,
2
0
1
6
.
[2
2
]
J.
M
a
rin
i,
“
Th
e
d
o
c
u
m
e
n
t
o
b
jec
t
m
o
d
e
l,
p
ro
c
e
ss
in
g
str
u
c
tu
re
d
d
o
c
u
m
e
n
ts
,”
M
c
Gr
a
w
-
Hill
/Os
b
o
rn
e
,
2
0
0
2
.
[2
3
]
S
-
H.
Ch
a
,
“
Co
m
p
re
h
e
n
si
v
e
su
r
v
e
y
o
n
d
istan
c
e
/sim
il
a
rit
y
m
e
a
su
re
s
b
e
twe
e
n
p
r
o
b
a
b
il
it
y
d
e
n
sity
f
u
n
c
ti
o
n
s
”
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
M
a
t
h
e
ma
ti
c
a
l
M
o
d
e
ls a
n
d
M
e
t
h
o
d
s
in
Ap
p
li
e
d
S
c
ien
c
e
s
,
v
o
l.
1
,
n
o
.
4
,
Ja
n
u
a
ry
2
0
0
7
.
[2
4
]
L.
T.
S
u
,
“
Th
e
re
lev
a
n
c
e
o
f
re
c
a
ll
a
n
d
p
re
c
isio
n
in
u
se
r
e
v
a
lu
a
ti
o
n
,
”
J
o
u
rn
a
l
o
f
t
h
e
Am
e
rica
n
S
o
c
iet
y
f
o
r
In
f
o
rm
a
ti
o
n
S
c
i
e
n
c
e
,
v
o
l
.
4
5
,
n
o
.
3
,
1
9
9
4
.
[2
5
]
M
.
Ju
n
k
e
r,
e
t
a
l
.
,
“
On
t
h
e
e
v
a
lu
a
ti
o
n
o
f
d
o
c
u
m
e
n
t
a
n
a
ly
sis
c
o
m
p
o
n
e
n
ts
b
y
re
c
a
ll
,
p
re
c
isio
n
,
a
n
d
a
c
c
u
ra
c
y
,
”
Pro
c
e
e
d
in
g
s
o
f
t
h
e
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Do
c
u
me
n
t
An
a
lys
i
s a
n
d
Rec
o
g
n
it
i
o
n
,
ICDAR
,
S
e
p
te
m
b
e
r
1
9
9
9
.
[2
6
]
A.
A.
Ra
d
wa
n
,
e
t
a
l
.
,
“
Us
in
g
g
e
n
e
t
ic
a
lg
o
ri
th
m
t
o
im
p
ro
v
e
in
f
o
rm
a
ti
o
n
re
tr
iev
a
l
sy
ste
m
s
”
,
W
o
rld
Aca
d
e
my
o
f
S
c
ien
c
e
,
En
g
i
n
e
e
rin
g
a
n
d
T
e
c
h
n
o
lo
g
y
In
te
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
a
n
d
In
f
o
rm
a
ti
o
n
En
g
in
e
e
rin
g
,
v
ol
.
2
,
n
o
.
5
,
2
0
0
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.