I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
2
0
2
5
,
p
p
.
5
1
8
3
~
5
1
9
2
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
14
.i
6
.
p
p
5
1
8
3
-
5
1
9
2
5183
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Im
pa
ct
o
f
sm
o
o
th
ing
techni
ques for
t
e
x
t
cla
ss
ificatio
n
:
implem
entatio
n
i
n hidden
M
a
rko
v
mo
del
No
rsy
ela
M
uh
a
m
m
a
d No
o
r
M
a
t
hiv
a
na
n
1,
2,
3
,
Ro
zia
h M
o
hd
J
a
no
r
1
,
S
hu
k
o
r
Abd Ra
za
k
4
,
No
r
Azura
M
d.
G
ha
ni
1
1
S
c
h
o
o
l
o
f
M
a
t
h
e
ma
t
i
c
a
l
S
c
i
e
n
c
e
s,
F
a
c
u
l
t
y
o
f
C
o
m
p
u
t
e
r
a
n
d
M
a
t
h
e
m
a
t
i
c
a
l
S
c
i
e
n
c
e
s,
U
n
i
v
e
r
s
i
t
i
T
e
k
n
o
l
o
g
i
M
A
R
A
,
S
h
a
h
A
l
a
m
,
M
a
l
a
y
si
a
2
S
c
h
o
o
l
o
f
C
o
m
p
u
t
i
n
g
a
n
d
C
r
e
a
t
i
v
e
M
e
d
i
a
,
U
n
i
v
e
r
s
i
t
y
o
f
W
o
l
l
o
n
g
o
n
g
M
a
l
a
y
si
a
,
S
h
a
h
A
l
a
m,
M
a
l
a
y
s
i
a
3
U
O
W
M
a
l
a
y
si
a
K
D
U
P
e
n
a
n
g
U
n
i
v
e
r
si
t
y
C
o
l
l
e
g
e
,
G
e
o
r
g
e
t
o
w
n
,
M
a
l
a
y
s
i
a
4
F
a
c
u
l
t
y
o
f
I
n
f
o
r
m
a
t
i
c
s
a
n
d
C
o
mp
u
t
i
n
g
,
U
n
i
v
e
r
s
i
t
i
S
u
l
t
a
n
Z
a
i
n
a
l
A
b
i
d
i
n
,
K
u
a
l
a
Te
r
e
n
g
g
a
n
u
,
M
a
l
a
y
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
25
,
2
0
2
4
R
ev
is
ed
Au
g
18
,
2
0
2
5
Acc
ep
ted
Sep
7
,
2
0
2
5
A
h
i
d
d
e
n
M
a
r
k
o
v
m
o
d
e
l
(HMM
)
is
wi
d
e
ly
u
se
d
f
o
r
se
q
u
e
n
c
e
m
o
d
e
li
n
g
in
v
a
rio
u
s tex
t
c
las
sifica
ti
o
n
tas
k
s.
T
h
i
s
s
t
u
d
y
i
n
v
e
s
t
i
g
a
t
e
s
t
h
e
i
m
p
a
c
t
o
f
d
i
f
f
e
r
e
n
t
s
m
o
o
t
h
i
n
g
t
e
c
h
n
i
q
u
e
s
,
s
u
c
h
a
s
L
a
p
l
a
c
e
,
a
b
s
o
l
u
t
e
d
i
s
c
o
u
n
t
i
n
g
,
a
n
d
G
i
b
b
s
s
a
m
p
l
i
n
g
o
n
H
M
M
p
e
r
f
o
r
m
a
n
c
e
a
c
r
o
s
s
t
h
re
e
d
i
s
t
i
n
c
t
d
o
m
a
i
n
s
:
e
-
c
o
m
m
e
rc
e
p
ro
d
u
c
ts
,
s
p
a
m
fil
teri
n
g
,
a
n
d
o
c
c
u
p
a
ti
o
n
a
l
d
a
ta
m
in
in
g
.
T
h
r
o
u
g
h
t
h
e
c
o
m
p
a
ra
ti
v
e
a
n
a
ly
sis,
Lap
lac
e
sm
o
o
th
i
n
g
c
o
n
siste
n
tl
y
o
u
t
p
e
rfo
rm
s
o
th
e
r
tec
h
n
iq
u
e
s
in
h
a
n
d
li
n
g
z
e
ro
-
p
r
o
b
a
b
i
li
ty
issu
e
s,
d
e
m
o
n
stra
ti
n
g
su
p
e
rio
r
p
e
rfo
rm
a
n
c
e
in
th
e
e
-
c
o
m
m
e
rc
e
a
n
d
S
M
S
sp
a
m
d
a
tas
e
ts.
T
h
e
HM
M
with
o
u
t
a
n
y
sm
o
o
t
h
in
g
tec
h
n
iq
u
e
a
c
h
iev
e
d
th
e
b
e
st
re
su
lt
s
fo
r
jo
b
ti
tl
e
c
las
sifica
ti
o
n
.
Th
is
d
i
v
e
rg
e
n
c
e
u
n
d
e
rsc
o
re
s
t
h
e
d
a
tas
e
t
-
sp
e
c
ifi
c
n
a
tu
re
o
f
sm
o
o
th
i
n
g
re
q
u
irem
e
n
ts,
w
h
e
re
t
h
e
sim
p
li
c
it
y
o
f
p
a
ra
m
e
ter
e
stim
a
ti
o
n
p
r
o
v
e
s
e
ffe
c
ti
v
e
in
c
o
n
tex
ts
c
h
a
ra
c
teriz
e
d
b
y
a
li
m
it
e
d
a
n
d
re
p
e
ti
ti
v
e
v
o
c
a
b
u
lary
.
He
n
c
e
,
th
e
fin
d
i
n
g
s
s
u
g
g
e
st
th
a
t
tai
lo
re
d
sm
o
o
t
h
in
g
stra
teg
ies
a
re
c
ru
c
ial
fo
r
o
p
ti
m
izi
n
g
HMM
p
e
rfo
rm
a
n
c
e
i
n
d
iffere
n
t
te
x
tu
a
l
a
n
a
l
y
sis a
p
p
li
c
a
ti
o
n
s.
K
ey
w
o
r
d
s
:
E
-
co
m
m
er
ce
p
r
o
d
u
cts
J
o
b
titl
e
class
if
icatio
n
Occ
u
p
atio
n
al
d
ata
m
i
n
in
g
Pro
d
u
ct
class
if
icatio
n
Seq
u
en
tial d
ata
Sp
am
f
ilter
in
g
Su
p
er
v
is
ed
lear
n
in
g
m
o
d
el
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
:
No
r
Azu
r
a
Md
.
Gh
an
i
Sch
o
o
l o
f
Ma
th
e
m
atica
l Scie
n
ce
s
,
Facu
lty
o
f
C
o
m
p
u
te
r
an
d
Ma
th
em
atica
l Scie
n
ce
s
Un
iv
er
s
iti T
ek
n
o
lo
g
i M
AR
A
4
0
4
5
0
Sh
ah
Alam
,
Selan
g
o
r
Ma
lay
s
ia
E
m
ail: a
zu
r
a@
tm
s
k
.
u
itm
.
ed
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
T
ex
t
class
if
icatio
n
i
s
a
co
r
e
task
in
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
(
NL
P)
th
at
in
v
o
lv
es
as
s
ig
n
in
g
p
r
ed
ef
in
e
d
lab
els
to
tex
t
d
o
cu
m
en
ts
.
I
t
is
wid
ely
u
s
ed
in
ap
p
licatio
n
s
s
u
ch
as
in
f
o
r
m
atio
n
r
etr
iev
al
[
1
]
,
[
2
]
,
s
en
tim
en
t
an
aly
s
is
[
3
]
,
[
4
]
,
p
r
o
d
u
ct
class
if
icatio
n
[
5
]
,
[
6
]
,
s
p
am
d
etec
tio
n
[
7
]
,
[
8
]
,
a
n
d
d
o
cu
m
e
n
t
ca
teg
o
r
izatio
n
[
9
]
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
tex
t
class
if
icatio
n
m
eth
o
d
s
g
r
ea
tly
af
f
ec
ts
th
e
ef
f
ic
ien
cy
an
d
ac
cu
r
ac
y
o
f
m
a
n
y
au
to
m
ated
s
y
s
tem
s
,
m
ak
in
g
it
im
p
o
r
tan
t
to
im
p
r
o
v
e
an
d
ev
al
u
ate
d
if
f
er
en
t
tec
h
n
iq
u
es.
O
n
e
well
-
k
n
o
wn
m
eth
o
d
in
tex
t
class
if
icatio
n
is
th
e
h
id
d
en
Ma
r
k
o
v
m
o
d
el
(
HM
M)
[
1
0
]
,
w
h
ich
is
ef
f
ec
tiv
e
in
m
o
d
elin
g
s
eq
u
en
ce
s
d
u
e
to
its
p
r
o
b
ab
il
is
tic
s
tr
u
ctu
r
e.
Ho
wev
er
,
HM
Ms
o
f
ten
s
tr
u
g
g
le
with
s
p
ar
s
e
d
ata
an
d
u
n
s
ee
n
ev
en
ts
[
1
1
]
,
c
o
m
m
o
n
is
s
u
es
i
n
lar
g
e
tex
t
d
atasets
.
Sm
o
o
th
in
g
tech
n
iq
u
es
h
elp
a
d
d
r
ess
th
ese
p
r
o
b
lem
s
b
y
ad
ju
s
tin
g
p
r
o
b
ab
ilit
y
esti
m
ates f
o
r
r
a
r
e
o
r
m
is
s
in
g
d
ata
[
1
2
]
,
i
m
p
r
o
v
i
n
g
g
e
n
er
aliza
tio
n
a
n
d
m
o
d
el
r
eliab
ilit
y
.
Sev
er
al
s
m
o
o
th
in
g
m
eth
o
d
s
,
s
u
ch
as
L
ap
lace
s
m
o
o
t
h
in
g
,
Go
o
d
-
T
u
r
in
g
d
is
co
u
n
tin
g
,
an
d
b
ac
k
o
f
f
m
o
d
els
h
av
e
b
ee
n
ex
ten
s
iv
ely
s
tu
d
ied
in
NL
P.
T
h
ese
m
et
h
o
d
s
r
ed
u
ce
th
e
r
is
k
o
f
ass
ig
n
in
g
ze
r
o
p
r
o
b
ab
ilit
ies
to
u
n
s
ee
n
ev
en
ts
,
wh
ic
h
co
u
ld
o
th
e
r
wis
e
ca
u
s
e
er
r
o
r
s
d
u
r
in
g
class
if
icatio
n
.
R
ec
en
tly
,
m
o
r
e
ad
v
an
ce
d
tech
n
iq
u
es
h
a
v
e
b
ee
n
d
ev
el
o
p
ed
to
im
p
r
o
v
e
th
is
p
r
o
ce
s
s
.
Fo
r
ex
am
p
le,
R
en
et
a
l
.
[
1
3
]
i
n
tr
o
d
u
ce
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
1
8
3
-
5
1
9
2
5184
d
is
cr
im
in
atio
n
-
awa
r
e
lab
el
s
m
o
o
th
in
g
,
wh
ich
d
y
n
am
ically
le
ar
n
s
lab
el
d
is
tr
ib
u
tio
n
s
to
h
an
d
le
class
im
b
alan
ce
an
d
n
o
is
y
d
ata.
W
u
et
a
l
.
[
1
4
]
p
r
o
p
o
s
ed
tex
t
s
m
o
o
th
i
n
g
,
wh
ich
u
s
es
p
r
e
-
tr
ain
ed
m
ask
ed
l
an
g
u
ag
e
m
o
d
els
to
co
n
v
er
t
o
n
e
-
h
o
t
v
ec
to
r
s
in
to
m
o
r
e
in
f
o
r
m
ativ
e
r
ep
r
esen
tatio
n
s
,
im
p
r
o
v
in
g
p
e
r
f
o
r
m
an
ce
in
l
o
w
-
r
eso
u
r
ce
s
ettin
g
s
.
Fettal
e
t
a
l.
[
1
5
]
e
x
p
lo
r
ed
s
em
an
tic
g
r
ap
h
s
m
o
o
th
in
g
,
u
s
in
g
s
em
an
tic
r
elati
o
n
s
h
ip
s
to
en
h
an
ce
s
en
ten
ce
em
b
ed
d
in
g
s
f
o
r
b
ett
er
class
if
icatio
n
an
d
clu
s
ter
in
g
.
I
n
HM
Ms,
s
m
o
o
th
in
g
e
n
h
a
n
ce
s
b
o
th
tr
an
s
itio
n
an
d
em
is
s
io
n
p
r
o
b
ab
ilit
y
esti
m
ates,
im
p
r
o
v
in
g
class
if
icatio
n
ac
cu
r
ac
y
.
W
u
et
a
l
.
[
1
4
]
s
h
o
wed
th
at
s
m
o
o
th
e
d
r
ep
r
esen
tatio
n
s
ca
n
o
u
tp
e
r
f
o
r
m
o
n
e
-
h
o
t
e
n
co
d
in
g
s
in
d
ata
au
g
m
e
n
tatio
n
,
s
u
g
g
esti
n
g
s
im
ilar
b
en
ef
its
f
o
r
HM
Ms.
Sm
o
o
th
in
g
em
is
s
io
n
p
r
o
b
ab
ilit
ies
is
also
ess
en
tial
to
a
v
o
id
ze
r
o
v
al
u
es,
as
s
h
o
wn
in
m
etad
at
a
ex
tr
ac
tio
n
f
r
o
m
b
ib
lio
g
r
a
p
h
ic
r
ef
er
en
ce
s
[
1
6
]
.
Usi
n
g
th
e
e
n
ti
r
e
v
o
ca
b
u
l
ar
y
wit
h
a
p
p
r
o
p
r
iate
s
m
o
o
t
h
i
n
g
h
as
o
u
t
p
e
r
f
o
r
m
ed
c
o
n
v
e
n
ti
o
n
al
f
ea
tu
r
e
s
el
ec
ti
o
n
in
e
n
s
u
r
i
n
g
r
e
lia
b
le
p
a
r
a
m
et
er
est
im
ati
o
n
[
1
7
]
.
Fu
r
th
e
r
m
o
r
e
,
f
u
zz
y
s
m
o
o
t
h
i
n
g
o
f
s
tat
e
t
r
a
n
s
i
ti
o
n
s
h
as
im
p
r
o
v
ed
cl
ass
if
ic
ati
o
n
r
at
es
i
n
u
n
ce
r
t
ai
n
e
n
v
i
r
o
n
m
e
n
ts
,
s
u
c
h
as
s
p
e
ec
h
r
e
co
g
n
it
io
n
,
wi
th
a
p
p
lic
ati
o
n
s
in
te
x
t
class
if
ica
ti
o
n
[
1
8
]
.
H
MM
s
h
a
v
e
als
o
s
h
o
wn
s
u
cc
e
s
s
in
d
o
m
ai
n
s
l
ik
e
b
i
o
m
e
d
ic
al
te
x
t
a
n
d
d
o
c
u
m
e
n
t
class
i
f
i
ca
t
io
n
,
p
a
r
t
ic
u
l
ar
l
y
w
h
e
n
e
n
h
an
ce
d
wit
h
s
m
o
o
t
h
i
n
g
t
e
ch
n
i
q
u
es
[
1
9
]
.
T
h
is
r
esear
ch
aim
s
to
an
aly
ze
th
e
im
p
ac
t
o
f
d
if
f
er
e
n
t
s
m
o
o
th
in
g
tech
n
iq
u
es
o
n
th
e
p
er
f
o
r
m
an
ce
o
f
HM
Ms
in
tex
t
class
if
icat
io
n
task
s
.
B
y
s
y
s
tem
atica
lly
im
p
lem
en
tin
g
an
d
co
m
p
ar
in
g
t
h
ese
m
eth
o
d
s
,
we
s
ee
k
to
id
en
tify
th
e
m
o
s
t
ef
f
ec
tiv
e
s
tr
ateg
ies
f
o
r
en
h
an
cin
g
th
e
ac
c
u
r
ac
y
a
n
d
r
o
b
u
s
tn
ess
o
f
HM
M
-
b
ased
class
if
ier
s
.
T
h
e
s
tu
d
y
also
ex
p
lo
r
es
th
e
tr
ad
e
-
o
f
f
s
ass
o
ciate
d
with
ea
c
h
tech
n
iq
u
e,
p
r
o
v
id
in
g
i
n
s
ig
h
ts
in
to
th
eir
p
r
ac
tical
ap
p
licatio
n
s
an
d
p
o
te
n
tial
ar
ea
s
f
o
r
f
u
r
th
e
r
im
p
r
o
v
em
en
t.
I
n
th
e
f
o
llo
win
g
s
ec
tio
n
s
,
we
will
r
ev
iew
th
e
th
eo
r
etica
l
f
o
u
n
d
atio
n
s
o
f
H
MM
s
an
d
s
m
o
o
th
in
g
tech
n
iq
u
es,
d
escr
ib
e
o
u
r
ex
p
e
r
im
en
t
al
s
etu
p
,
p
r
esen
t
th
e
r
esu
lts
o
f
o
u
r
c
o
m
p
ar
ativ
e
an
a
ly
s
is
,
an
d
d
is
cu
s
s
th
e
im
p
licatio
n
s
o
f
o
u
r
f
i
n
d
in
g
s
.
T
h
r
o
u
g
h
t
h
is
co
m
p
r
e
h
en
s
iv
e
ev
alu
atio
n
,
we
h
o
p
e
to
c
o
n
tr
i
b
u
te
to
t
h
e
o
n
g
o
in
g
ef
f
o
r
ts
in
o
p
tim
izin
g
tex
t
class
if
icatio
n
m
eth
o
d
o
lo
g
ies
an
d
ad
v
an
cin
g
th
e
f
ield
o
f
NL
P.
2.
M
E
T
H
O
D
2
.
1
.
Da
t
a
des
cr
iptio
n
Dep
ar
tm
en
t
o
f
s
tatis
tic
s
Ma
la
y
s
ia
(
DOSM)
h
as
co
l
lec
ted
p
r
o
d
u
ct
in
f
o
r
m
atio
n
f
r
o
m
o
n
e
o
f
th
e
m
ajo
r
o
n
lin
e
s
to
r
e
web
s
ites
th
r
o
u
g
h
th
e
STAT
SB
DA
p
r
o
ject
k
n
o
wn
as
p
r
ice
in
tellig
en
ce
(
PI)
u
s
in
g
its
p
r
o
to
ty
p
e
web
s
cr
ap
er
.
A
f
ew
leaf
n
o
d
es
wer
e
u
s
ed
to
r
ep
r
esen
t
th
e
ch
o
s
en
ca
teg
o
r
ies
f
r
o
m
th
e
b
r
o
wse
tr
ee
o
f
th
e
web
s
ite.
T
ab
le
1
p
r
esen
ts
th
e
d
escr
ip
tio
n
o
f
t
h
e
f
o
u
r
co
r
p
o
r
a
s
elec
ted
f
o
r
th
is
s
tu
d
y
wh
ich
in
co
r
p
o
r
ated
d
atasets
f
r
o
m
th
r
ee
d
if
f
er
en
t
d
o
m
ain
s
.
T
h
e
f
ir
s
t
d
o
m
ain
is
e
-
co
m
m
er
ce
p
r
o
d
u
cts
an
d
th
e
r
e
ar
e
two
d
atasets
u
s
ed
i.e
.
n
o
n
-
f
o
o
d
an
d
h
o
u
s
eh
o
ld
p
r
o
d
u
cts
u
n
d
er
th
is
d
o
m
ai
n
.
T
h
e
two
ca
teg
o
r
ies
u
n
d
er
t
h
e
n
o
n
-
f
o
o
d
d
ataset
ar
e
co
o
k
in
g
&
d
i
n
in
g
(
4
0
7
i
n
s
tan
ce
s
)
an
d
p
ar
ty
ac
ce
s
s
o
r
ies
(
8
0
in
s
tan
ce
s
)
.
On
th
e
o
t
h
er
h
an
d
,
th
e
f
i
v
e
ca
teg
o
r
ies
u
n
d
e
r
th
e
Fr
o
ze
n
d
ataset
ar
e
f
r
o
ze
n
f
o
o
d
(
2
9
1
in
s
tan
ce
s
)
,
y
o
g
u
r
t
(
1
6
2
in
s
tan
ce
s
)
,
ice
cr
ea
m
(
1
4
7
i
n
s
tan
ce
s
)
,
ch
ee
s
e
(
8
5
in
s
tan
ce
s
)
,
an
d
ju
ices (
8
7
in
s
tan
c
es
)
.
T
h
is
s
tu
d
y
also
u
tili
ze
d
two
ad
d
itio
n
al
d
atasets
f
r
o
m
d
if
f
e
r
en
t
d
o
m
ain
s
,
n
am
el
y
s
p
am
f
ilter
in
g
an
d
o
cc
u
p
atio
n
al
d
ata
m
in
i
n
g
.
T
h
e
d
ataset
r
elate
d
to
s
p
am
f
ilter
i
n
g
was
r
etr
iev
ed
f
r
o
m
th
e
UC
I
r
ep
o
s
ito
r
y
,
wh
ich
p
r
o
v
id
es
a
wid
ely
r
ec
o
g
n
ized
co
llectio
n
o
f
d
ata
f
o
r
m
ac
h
i
n
e
lear
n
in
g
a
p
p
licatio
n
s
.
T
h
is
d
ataset
co
m
p
r
is
es
lab
eled
in
s
tan
ce
s
o
f
em
ails
ca
teg
o
r
ize
d
as
s
p
am
o
r
n
o
n
-
s
p
am
,
allo
win
g
f
o
r
th
e
ev
alu
atio
n
o
f
tex
t
class
if
icatio
n
m
o
d
els
in
d
is
tin
g
u
is
h
in
g
b
etwe
en
u
n
s
o
licited
an
d
leg
itima
te
m
ess
ag
es.
Me
a
n
wh
ile,
th
e
d
ataset
f
r
o
m
Gith
u
b
was
u
s
ed
f
o
r
cl
ass
if
y
in
g
jo
b
titl
es
ac
co
r
d
in
g
to
th
ei
r
jo
b
ca
teg
o
r
ies.
T
h
is
d
ataset
co
n
s
is
ts
o
f
v
ar
io
u
s
jo
b
titl
es
with
c
o
r
r
es
p
o
n
d
in
g
ca
te
g
o
r
ies,
o
f
f
er
i
n
g
v
alu
ab
le
in
s
ig
h
ts
f
o
r
m
ac
h
i
n
e
lear
n
in
g
m
o
d
els
aim
ed
at
au
t
o
m
atin
g
j
o
b
class
if
icatio
n
.
T
h
e
in
clu
s
io
n
o
f
t
h
e
s
e
d
atasets
en
s
u
r
es
th
e
r
o
b
u
s
tn
ess
o
f
th
e
s
tu
d
y
b
y
co
v
er
in
g
d
iv
er
s
e
d
o
m
ain
s
an
d
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
,
en
h
a
n
cin
g
th
e
g
e
n
er
aliza
b
ilit
y
o
f
th
e
f
in
d
in
g
s
.
T
ab
le
1
.
Su
m
m
a
r
y
d
escr
ip
tio
n
o
f
d
atasets
D
a
t
a
s
e
t
C
a
t
e
g
o
r
y
I
n
st
a
n
c
e
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
TF
TF
-
I
D
F
N
o
n
-
f
o
o
d
2
4
8
7
4
6
1
4
5
9
F
r
o
z
e
n
f
o
o
d
5
7
7
2
6
5
6
6
5
4
S
M
S
s
p
a
m
2
5
5
6
7
5
9
0
3
5
6
3
7
Jo
b
t
i
t
l
e
4
8
5
8
6
1
9
2
5
1
9
1
9
2
.
2
.
Da
t
a
cha
r
a
ct
er
is
t
ics
T
h
e
p
r
o
d
u
ct
titl
e
len
g
th
s
ac
r
o
s
s
th
e
d
atasets
s
h
o
w
a
r
elativ
ely
s
h
o
r
t
an
d
co
n
s
is
ten
t
d
is
tr
ib
u
tio
n
,
as
s
ee
n
in
Fig
u
r
e
1
.
T
h
e
n
o
n
-
f
o
o
d
p
r
o
d
u
cts
s
u
b
s
et
h
as
a
m
o
d
e
o
f
8
ch
ar
ac
ter
s
,
an
d
th
e
f
r
o
ze
n
f
o
o
d
p
r
o
d
u
cts
s
u
b
s
et
h
as
a
m
o
d
e
o
f
6
ch
ar
ac
ter
s
.
T
h
is
in
d
icate
s
th
at
th
e
p
r
o
d
u
ct
titl
es
in
th
ese
d
ataset
s
a
r
e
ty
p
ically
s
h
o
r
ter
,
wh
ich
m
ay
in
f
lu
en
ce
th
e
ef
f
ec
tiv
en
ess
o
f
d
if
f
er
e
n
t sm
o
o
th
in
g
tech
n
iq
u
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
mp
a
ct
o
f smo
o
th
in
g
tech
n
i
q
u
es fo
r
text
cla
s
s
if
ica
tio
n
…
(
N
o
r
s
ye
la
Mu
h
a
mma
d
N
o
o
r
Ma
th
iva
n
a
n
)
5185
F
i
g
u
r
e
2
s
h
o
w
s
t
h
e
d
i
s
t
r
i
b
u
ti
o
n
o
f
t
e
x
t
l
e
n
g
t
h
s
f
o
r
t
h
e
S
M
S
s
p
a
m
a
n
d
j
o
b
t
i
t
l
e
d
a
t
as
e
ts
.
T
h
e
S
M
S
s
p
a
m
t
e
x
t
l
e
n
g
t
h
d
i
s
t
r
i
b
u
t
i
o
n
s
h
o
w
s
a
m
o
d
e
o
f
7
c
h
a
r
a
c
t
e
r
s
,
w
i
t
h
t
h
e
m
a
j
o
r
i
t
y
o
f
m
e
s
s
a
g
e
s
b
ei
n
g
r
e
l
a
t
i
v
el
y
s
h
o
r
t
.
T
h
is
p
a
t
t
e
r
n
i
s
s
i
m
i
l
a
r
t
o
t
h
e
p
r
o
d
u
c
t
t
i
tl
e
l
e
n
g
t
h
s
s
ee
n
i
n
F
i
g
u
r
e
1
.
I
n
c
o
n
t
r
a
s
t
,
t
h
e
j
o
b
t
i
t
le
d
a
t
as
e
t
e
x
h
i
b
i
t
s
a
m
o
d
e
o
f
3
c
h
a
r
a
c
t
e
r
s
,
w
i
t
h
a
s
i
g
n
i
f
i
ca
n
t
p
o
r
t
i
o
n
o
f
t
h
e
d
a
t
a
c
o
n
s
is
t
i
n
g
o
f
v
e
r
y
s
h
o
r
t
a
n
d
r
e
p
e
t
i
t
i
v
e
t
i
tl
e
s
.
Fig
u
r
e
1
.
No
n
-
f
o
o
d
a
n
d
f
r
o
ze
n
f
o
o
d
p
r
o
d
u
cts titl
e
len
g
th
s
Fig
u
r
e
2
.
SMS
s
p
am
an
d
jo
b
ti
tle
tex
t le
n
g
th
s
2
.
3
.
Resea
rc
h
des
ig
n
T
h
er
e
ar
e
s
ev
e
r
al
s
tep
s
n
ee
d
e
d
b
ef
o
r
e
class
if
y
in
g
th
e
d
ata
,
as
s
h
o
wn
in
Fig
u
r
e
3
.
T
h
ese
s
tep
s
ar
e
ess
en
tial
in
tex
t
class
if
icatio
n
r
esear
ch
.
Af
ter
ex
tr
ac
tin
g
th
e
d
ata,
th
r
ee
p
r
e
p
r
o
ce
s
s
in
g
s
tep
s
ar
e
a
p
p
lied
i.e
.
,
to
k
en
izatio
n
,
s
to
p
w
o
r
d
r
em
o
v
al,
an
d
s
tem
m
in
g
[
2
0
]
.
T
o
k
en
izatio
n
s
p
lits
p
r
o
d
u
ct
d
escr
ip
tio
n
s
in
to
wo
r
d
s
,
s
to
p
s
wo
r
d
r
em
o
v
al
f
ilter
s
o
u
t
co
m
m
o
n
wo
r
d
s
,
an
d
s
tem
m
in
g
r
ed
u
ce
s
wo
r
d
s
to
th
eir
r
o
o
t
f
o
r
m
s
,
en
s
u
r
in
g
s
tan
d
ar
d
ized
d
ata.
T
h
e
s
tu
d
y
u
s
ed
ter
m
-
f
r
eq
u
e
n
cy
(
T
F)
a
n
d
ter
m
-
f
r
eq
u
en
cy
i
n
v
er
s
e
d
o
cu
m
en
t
-
f
r
eq
u
e
n
cy
(TF
-
I
DF)
f
o
r
f
ea
tu
r
e
e
x
tr
ac
ti
o
n
an
d
ap
p
lied
th
e
co
r
r
elatio
n
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
e.
T
h
ese
s
elec
ted
f
ea
tu
r
es
ar
e
th
en
u
s
ed
as in
p
u
ts
f
o
r
HM
Ms w
ith
d
if
f
er
en
t sm
o
o
th
in
g
tech
n
iq
u
es.
Acc
o
r
d
in
g
t
o
th
e
f
u
n
d
am
en
t
al
p
r
o
b
l
em
s
r
el
ate
d
t
o
H
M
M,
o
n
l
y
tw
o
s
t
e
p
s
s
h
o
u
ld
b
e
t
ak
e
n
t
o
im
p
le
m
e
n
t
a
s
u
p
er
v
is
e
d
H
MM
,
i
.
e
.
,
esti
m
at
in
g
t
h
e
p
a
r
a
m
et
er
s
b
y
le
ar
n
i
n
g
t
h
ei
r
in
iti
al
p
r
o
b
ab
i
lit
y
m
at
r
ic
es
an
d
d
e
co
d
i
n
g
t
h
e
s
e
q
u
e
n
c
e
t
o
f
i
n
d
t
h
e
b
est
-
h
i
d
d
en
s
e
q
u
e
n
c
e
[
2
1
]
.
B
o
t
h
s
t
ep
s
a
r
e
d
o
n
e
to
s
o
l
v
e
le
ar
n
i
n
g
a
n
d
d
e
co
d
i
n
g
p
r
o
b
l
em
s
,
r
es
p
e
cti
v
e
ly
.
T
h
e
li
k
el
ih
o
o
d
p
r
o
b
le
m
is
ty
p
ic
all
y
ig
n
o
r
e
d
w
h
e
n
u
s
i
n
g
s
u
p
e
r
v
is
e
d
HM
Ms
b
e
ca
u
s
e
t
h
e
al
g
o
r
i
th
m
u
s
ed
f
o
r
th
e
f
o
r
wa
r
d
al
g
o
r
it
h
m
is
em
p
l
o
y
e
d
f
o
r
li
k
eli
h
o
o
d
c
o
m
p
u
t
ati
o
n
i
n
s
e
m
i
-
s
u
p
e
r
v
is
e
d
o
r
u
n
s
u
p
e
r
v
is
e
d
le
ar
n
i
n
g
.
T
h
e
Vit
er
b
i
al
g
o
r
it
h
m
is
s
i
m
il
ar
to
t
h
e
f
o
r
wa
r
d
al
g
o
r
ith
m
,
b
u
t
i
n
s
te
a
d
o
f
u
s
i
n
g
th
e
s
u
m
o
f
p
r
e
v
i
o
u
s
p
at
h
p
r
o
b
a
b
il
ities
,
it
u
s
es
th
e
m
a
x
i
m
u
m
v
al
u
e
.
T
h
e
f
o
r
wa
r
d
al
g
o
r
it
h
m
l
ac
k
s
a
b
a
c
k
p
o
i
n
te
r
c
o
m
p
o
n
e
n
t
a
n
d
c
o
m
p
u
tes
t
h
e
o
b
s
er
v
ati
o
n
li
k
el
ih
o
o
d
,
w
h
il
e
t
h
e
Vi
te
r
b
i
al
g
o
r
it
h
m
f
i
n
d
s
th
e
m
o
s
t
l
ik
el
y
s
tate
s
e
q
u
e
n
c
e
b
y
tr
ac
k
i
n
g
th
e
p
at
h
o
f
h
i
d
d
en
s
t
ates
l
ea
d
i
n
g
t
o
ea
ch
s
t
ate
[
2
1
]
.
T
h
e
lear
n
i
n
g
p
r
o
b
lem
in
HM
Ms
in
v
o
lv
es
a
d
ju
s
tin
g
t
h
e
m
o
d
el
p
ar
a
m
eter
s
.
Usi
n
g
th
e
tr
ai
n
in
g
s
et
o
f
o
b
s
er
v
atio
n
s
,
t
h
e
aim
is
to
f
i
n
d
th
e
b
est
way
to
p
r
ed
ict
t
h
e
s
tates.
B
y
co
m
p
ar
in
g
th
e
p
r
ed
icted
s
tates
to
th
e
k
n
o
wn
s
tates,
th
e
p
r
e
d
ictio
n
a
cc
u
r
ac
y
ca
n
b
e
esti
m
ated
b
ased
o
n
th
e
co
r
r
ec
tly
d
ec
o
d
e
d
s
tates
in
th
e
test
s
et.
Sin
ce
th
e
s
tates
ar
e
k
n
o
wn
,
m
ax
im
u
m
lik
elih
o
o
d
esti
m
ates
(
ML
E
s
)
ar
e
u
s
ed
to
m
ax
im
ize
th
e
co
m
p
lete
-
d
ata
lik
elih
o
o
d
an
d
o
b
tain
th
e
HM
M
p
ar
am
eter
s
.
T
h
er
e
a
r
e
t
h
r
ee
HM
M
p
a
r
am
ete
r
s
w
h
i
ch
a
r
e
A
,
B
,
a
n
d
π
t
h
at
ca
n
b
e
le
ar
n
ed
g
i
v
e
n
an
o
b
s
er
v
ati
o
n
s
e
q
u
e
n
c
e
o
a
n
d
t
h
e
s
e
t
o
f
s
tat
es
in
t
h
e
HM
M.
T
h
e
i
n
it
ial
p
r
o
b
a
b
ili
ties
a
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
1
8
3
-
5
1
9
2
5186
d
e
n
o
te
d
b
y
ar
e
t
h
e
n
u
m
b
er
o
f
tim
es
f
o
r
s
t
at
e
s
l
ab
ele
d
in
t
h
e
d
atas
et
.
T
h
e
i
n
it
ial
p
ar
am
ete
r
c
o
m
p
u
t
ati
o
n
is
p
r
ese
n
t
ed
b
y
,
I
n
itial p
r
o
b
a
b
ilit
ies,
=
∑
(
′
)
′
Nex
t,
th
e
tr
a
n
s
itio
n
p
r
o
b
ab
ilit
ies
ar
e
d
en
o
ted
b
y
ar
e
th
e
n
u
m
b
er
o
f
tim
es
tr
an
s
itio
n
to
was
tak
en
am
o
n
g
t
h
e
s
eq
u
e
n
ce
s
.
T
h
e
tr
a
n
s
iti
o
n
p
a
r
a
m
et
er
c
o
m
p
u
tat
io
n
i
s
p
r
es
e
n
te
d
b
y
,
T
r
an
s
itio
n
p
r
o
b
ab
ilit
ies,
=
∑
(
′
)
′
T
h
e
e
m
is
s
i
o
n
p
r
o
b
ab
ilit
ies
a
r
e
d
e
n
o
te
d
b
y
(
)
a
r
e
th
e
n
u
m
b
er
o
f
ti
m
es
k
w
as
e
m
itt
e
d
wh
ile
i
n
s
t
ate
s
.
T
h
e
em
is
s
io
n
p
ar
am
et
er
c
o
m
p
u
t
ati
o
n
is
p
r
ese
n
t
ed
b
y
,
E
m
is
s
io
n
p
r
o
b
ab
ilit
ies,
(
)
=
(
)
∑
(
(
′
)
)
′
Fig
u
r
e
3
.
R
esear
ch
f
r
am
ewo
r
k
I
n
a
s
u
p
e
r
v
is
e
d
HM
M,
em
is
s
i
o
n
p
r
o
b
a
b
ili
ties
a
r
e
co
m
p
u
te
d
to
ca
t
eg
o
r
iz
e
a
ll
w
o
r
d
s
i
n
t
h
e
d
atas
et,
a
n
d
tr
a
n
s
i
ti
o
n
s
b
etwe
e
n
c
ate
g
o
r
ies
f
o
r
a
s
e
q
u
e
n
c
e
o
f
o
b
s
e
r
v
ati
o
n
s
a
r
e
c
o
u
n
te
d
.
T
h
ese
tw
o
a
d
j
u
s
t
ab
l
e
p
a
r
a
m
et
e
r
s
a
r
e
k
e
y
f
o
r
ac
h
i
e
v
i
n
g
a
h
i
g
h
-
p
e
r
f
o
r
m
a
n
ce
class
if
i
ca
ti
o
n
m
o
d
el
a
n
d
ar
e
o
f
t
en
t
h
e
f
o
c
u
s
o
f
r
es
ea
r
c
h
e
r
s
.
T
h
e
p
s
e
u
d
o
c
o
d
e
to
c
o
m
p
u
t
e
all
t
h
r
ee
p
ar
am
ete
r
s
is
s
h
o
w
n
i
n
Fi
g
u
r
e
4
.
I
n
class
if
icatio
n
m
o
d
els
lik
e
HM
Ms
with
h
id
d
en
v
ar
iab
l
es,
th
e
d
ec
o
d
in
g
task
aim
s
t
o
f
in
d
th
e
o
p
tim
al
s
tate
s
eq
u
en
ce
f
o
r
a
g
iv
en
o
b
s
er
v
atio
n
s
eq
u
en
ce
,
r
ev
ea
lin
g
th
e
h
i
d
d
en
s
tr
u
ctu
r
e
o
f
th
e
HM
M
.
T
y
p
ically
,
t
h
is
in
v
o
lv
es
r
u
n
n
i
n
g
th
e
f
o
r
war
d
alg
o
r
ith
m
to
c
o
m
p
u
te
th
e
li
k
elih
o
o
d
o
f
t
h
e
o
b
s
er
v
atio
n
s
eq
u
en
ce
f
o
r
ea
ch
p
o
s
s
ib
le
h
id
d
en
s
tate
s
eq
u
en
ce
.
Ho
wev
e
r
,
d
u
e
to
t
h
e
ex
p
o
n
en
tial
n
u
m
b
er
o
f
s
tate
s
eq
u
en
ce
s
,
d
ir
ec
tly
u
s
in
g
th
e
f
o
r
war
d
alg
o
r
ith
m
b
ec
o
m
es
im
p
r
ac
tical.
I
n
s
tead
,
th
e
Viter
b
i
alg
o
r
ith
m
is
co
m
m
o
n
ly
em
p
lo
y
ed
.
I
t
ef
f
icien
tly
f
in
d
s
th
e
o
p
tim
al
s
t
ate
s
eq
u
en
ce
u
s
in
g
d
y
n
a
m
ic
p
r
o
g
r
am
m
in
g
an
d
r
ec
u
r
s
io
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
mp
a
ct
o
f smo
o
th
in
g
tech
n
i
q
u
es fo
r
text
cla
s
s
if
ica
tio
n
…
(
N
o
r
s
ye
la
Mu
h
a
mma
d
N
o
o
r
Ma
th
iva
n
a
n
)
5187
Fig
u
r
e
4.
A
p
s
eu
d
o
co
d
e
f
o
r
p
a
r
am
eter
esti
m
atio
n
s
in
HM
M
2
.
4
.
S
m
o
o
t
hin
g
t
ec
hn
iqu
es
A
s
m
o
o
th
in
g
tech
n
iq
u
e
is
a
f
latter
in
g
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
p
r
o
ce
s
s
to
en
s
u
r
e
all
wo
r
d
s
eq
u
en
ce
s
ca
n
o
cc
u
r
with
s
o
m
e
p
r
o
b
a
b
ilit
ies
r
ath
er
th
an
h
av
i
n
g
c
er
tain
wo
r
d
s
with
ze
r
o
p
r
o
b
ab
ilit
ies.
T
h
e
ter
m
s
m
o
o
th
in
g
r
ef
er
s
tech
n
iq
u
e
f
o
r
ad
ju
s
tin
g
th
e
m
ax
im
u
m
lik
e
lih
o
o
d
esti
m
ate
o
f
p
r
o
b
ab
ilit
ies
to
p
r
o
v
i
d
e
m
o
r
e
ac
cu
r
ate
p
r
o
b
ab
ilit
ies.
T
h
er
e
ar
e
th
r
ee
s
m
o
o
th
in
g
tech
n
i
q
u
es
u
s
ed
in
th
e
s
tu
d
y
i.e
.
L
ap
lace
,
ab
s
o
lu
te
d
is
co
u
n
tin
g
,
an
d
Gib
b
s
s
am
p
lin
g
s
m
o
o
t
h
in
g
tech
n
iq
u
es.
F
ig
u
r
e
5
s
h
o
ws
th
e
p
s
eu
d
o
c
o
d
es
f
o
r
p
a
r
am
eter
co
m
p
u
tatio
n
s
f
o
r
ea
ch
s
m
o
o
th
in
g
tech
n
iq
u
e.
S
p
e
c
i
f
i
ca
l
l
y
,
F
i
g
u
r
e
5
(
a
)
il
l
u
s
t
r
a
t
e
s
t
h
e
p
a
r
a
m
e
te
r
c
o
m
p
u
t
a
t
i
o
n
s
f
o
r
t
h
e
L
a
p
la
c
e
s
m
o
o
t
h
i
n
g
t
e
c
h
n
i
q
u
e
,
w
h
i
c
h
is
t
h
e
s
i
m
p
le
s
t
a
n
d
m
o
s
t
f
r
e
q
u
e
n
t
l
y
u
s
e
d
m
et
h
o
d
f
o
r
a
d
d
r
e
s
s
i
n
g
d
a
t
a
s
p
a
r
s
i
t
y
.
T
h
e
L
a
p
l
a
c
e
s
m
o
o
t
h
i
n
g
t
e
c
h
n
i
q
u
e
i
s
m
o
s
t
f
r
e
q
u
e
n
t
l
y
u
s
e
d
b
y
p
r
e
v
i
o
u
s
r
e
s
e
a
r
c
h
e
r
s
d
e
s
p
i
t
e
v
a
r
i
o
u
s
s
m
o
o
t
h
i
n
g
t
e
ch
n
i
q
u
e
s
h
a
v
e
b
e
e
n
p
r
o
p
o
s
e
d
[
1
2
]
.
I
t
i
s
t
h
e
s
i
m
p
l
es
t
a
n
d
o
l
d
e
s
t
te
c
h
n
i
q
u
e
t
o
s
o
lv
e
d
a
t
a
s
p
a
r
s
e
n
es
s
p
r
o
b
l
e
m
s
.
T
h
i
s
t
e
c
h
n
i
q
u
e
a
ls
o
s
e
r
v
e
s
as
a
f
u
n
d
a
m
e
n
t
al
b
a
s
el
in
e
c
o
n
c
e
p
t
f
o
r
o
t
h
e
r
s
m
o
o
t
h
i
n
g
t
e
c
h
n
i
q
u
e
s
wi
t
h
t
h
e
s
a
m
e
p
a
r
am
e
t
e
r
s
.
Fig
u
r
e
5
(
b
)
d
ep
icts
t
h
e
p
ar
am
eter
co
m
p
u
tatio
n
s
f
o
r
th
e
ab
s
o
lu
te
d
is
co
u
n
tin
g
s
m
o
o
th
in
g
t
ec
h
n
iq
u
e,
wh
ich
ad
ju
s
ts
tr
an
s
itio
n
an
d
em
is
s
io
n
p
r
o
b
ab
ilit
ies
b
y
d
is
co
u
n
tin
g
o
b
s
er
v
ed
co
u
n
ts
.
I
t
is
a
m
et
h
o
d
c
o
m
m
o
n
l
y
u
s
e
d
i
n
la
n
g
u
a
g
e
m
o
d
eli
n
g
c
o
n
te
x
ts
t
o
a
d
j
u
s
t
p
r
o
b
a
b
il
it
y
est
im
a
tes
b
y
d
is
co
u
n
tin
g
o
b
s
er
v
ed
co
u
n
ts
o
f
ev
e
n
ts
[
2
2
]
.
I
n
th
e
co
n
te
x
t
o
f
HM
M
s
,
ab
s
o
lu
te
d
is
co
u
n
tin
g
ad
ju
s
t
s
b
o
th
tr
an
s
itio
n
p
r
o
b
ab
ilit
ies
(
th
e
lik
elih
o
o
d
o
f
m
o
v
in
g
f
r
o
m
o
n
e
h
id
d
e
n
s
tate
to
an
o
th
er
)
an
d
em
is
s
io
n
p
r
o
b
ab
ilit
ies
(
th
e
lik
elih
o
o
d
o
f
em
itti
n
g
o
b
s
er
v
ab
le
s
y
m
b
o
ls
g
iv
e
n
a
h
id
d
en
s
tate)
.
T
h
e
co
r
e
id
ea
b
eh
in
d
a
b
s
o
lu
te
d
is
co
u
n
tin
g
is
s
tr
aig
h
tf
o
r
w
ar
d
y
et
ef
f
ec
tiv
e:
it
en
s
u
r
es
th
at
ev
en
if
ce
r
tain
s
tate
tr
an
s
itio
n
s
o
r
em
is
s
io
n
s
wer
e
n
o
t
o
b
s
er
v
e
d
d
u
r
i
n
g
tr
ain
in
g
,
th
ey
s
till
r
etain
a
non
-
ze
r
o
p
r
o
b
a
b
ilit
y
in
t
h
e
m
o
d
el.
T
h
is
is
ac
h
iev
ed
b
y
s
u
b
tr
ac
tin
g
a
f
ix
e
d
d
is
co
u
n
t
d
f
r
o
m
th
e
o
b
s
er
v
e
d
co
u
n
ts
o
f
e
v
en
ts
,
an
d
r
ed
is
tr
ib
u
tin
g
th
is
d
is
co
u
n
t
m
ass
am
o
n
g
all
p
o
s
s
ib
le
ev
en
ts
f
o
r
a
g
iv
e
n
co
n
tex
t.
Fig
u
r
e
5
(
c)
p
r
esen
ts
th
e
p
s
eu
d
o
co
d
e
f
o
r
th
e
Gib
b
s
Sam
p
lin
g
tech
n
iq
u
e,
a
M
ar
k
o
v
ch
ai
n
m
o
n
te
ca
r
l
o
(
MCMC
)
m
eth
o
d
u
s
ed
f
o
r
esti
m
atin
g
m
o
d
el
p
ar
am
eter
s
th
r
o
u
g
h
iter
ativ
e
s
am
p
lin
g
.
It
is
p
ar
ticu
lar
ly
well
-
s
u
ited
f
o
r
co
m
p
lex
m
o
d
els
an
d
lar
g
e
d
atasets
[
2
3
]
.
I
t
is
a
M
ar
k
o
v
c
h
ain
m
o
n
te
ca
r
lo
(
MCMC
)
m
eth
o
d
th
at
g
en
er
ates
s
am
p
les
f
r
o
m
a
jo
in
t
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
b
y
iter
ativ
e
s
am
p
lin
g
f
r
o
m
th
e
co
n
d
itio
n
a
l
d
is
tr
ib
u
tio
n
s
o
f
ea
ch
v
ar
ia
b
le.
I
n
th
e
c
o
n
tex
t
o
f
HM
Ms,
Gib
b
s
Sam
p
lin
g
ca
n
b
e
u
s
ed
t
o
esti
m
ate
th
e
h
id
d
e
n
s
tates
g
iv
en
th
e
o
b
s
er
v
ed
d
ata
an
d
th
en
to
u
p
d
ate
th
e
m
o
d
el
p
ar
am
eter
s
b
ased
o
n
th
ese
s
a
m
p
led
s
tates.
T
h
is
iter
ativ
e
p
r
o
ce
s
s
allo
ws
f
o
r
th
e
ex
p
lo
r
atio
n
o
f
th
e
p
o
s
ter
io
r
d
is
tr
ib
u
tio
n
o
f
th
e
m
o
d
el
p
a
r
a
m
eter
s
,
p
r
o
v
id
in
g
a
r
o
b
u
s
t m
ea
n
s
o
f
in
co
r
p
o
r
ati
n
g
th
e
v
ar
iab
ilit
y
in
t
h
e
d
ata
an
d
av
o
id
in
g
o
v
er
f
itti
n
g
to
s
p
ar
s
e
o
b
s
er
v
atio
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
1
8
3
-
5
1
9
2
5188
(
a)
(
b
)
(
c)
Fig
u
r
e
5.
P
s
eu
d
o
c
o
d
e
s
f
o
r
p
ar
am
eter
esti
m
atio
n
s
in
HM
M
u
s
in
g
:
(
a)
L
ap
lace
s
m
o
o
th
in
g
te
ch
n
iq
u
e,
(
b
)
ab
s
o
lu
te
d
is
co
u
n
tin
g
s
m
o
o
th
in
g
tech
n
i
q
u
e
,
a
n
d
(
c
)
Gib
b
s
s
am
p
lin
g
s
m
o
o
th
i
n
g
tech
n
iq
u
e
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
c
o
m
p
ar
ati
v
e
a
n
a
ly
s
is
was
d
o
n
e
b
y
o
b
s
e
r
v
i
n
g
t
h
e
H
M
M
p
er
f
o
r
m
a
n
ce
s
w
h
e
n
a
p
p
ly
i
n
g
d
if
f
e
r
e
n
t
s
m
o
o
t
h
i
n
g
t
ec
h
n
i
q
u
es.
T
a
b
l
e
2
s
h
o
ws t
h
e
cl
ass
i
f
ic
ati
o
n
r
es
u
lt
s
f
o
r
HM
M
v
ar
iati
o
n
s
ap
p
l
ie
d
to
t
wo
e
-
c
o
m
m
e
r
ce
d
atasets
:
n
o
n
-
f
o
o
d
p
r
o
d
u
cts
an
d
f
r
o
ze
n
f
o
o
d
p
r
o
d
u
cts.
T
h
e
ev
alu
atio
n
m
etr
ic
u
s
ed
is
th
e
F1
-
s
co
r
e,
an
d
two
d
if
f
er
en
t
em
b
ed
d
in
g
tech
n
iq
u
es,
T
F
an
d
TF
-
I
DF
ar
e
ap
p
lied
.
Fo
r
th
e
n
o
n
-
f
o
o
d
p
r
o
d
u
cts
d
ataset,
th
e
s
tan
d
ar
d
HM
M
with
o
u
t
s
m
o
o
th
in
g
ac
h
iev
es
an
F1
-
s
co
r
e
o
f
8
2
.
2
7
%
with
T
F
em
b
ed
d
in
g
,
wh
ic
h
is
s
lig
h
tly
b
etter
co
m
p
ar
ed
to
wh
en
u
s
in
g
TF
-
I
DF
em
b
ed
d
in
g
.
W
h
en
ab
s
o
lu
t
e
d
is
co
u
n
tin
g
is
ap
p
lied
,
th
e
F1
-
s
co
r
es
in
cr
ea
s
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
mp
a
ct
o
f smo
o
th
in
g
tech
n
i
q
u
es fo
r
text
cla
s
s
if
ica
tio
n
…
(
N
o
r
s
ye
la
Mu
h
a
mma
d
N
o
o
r
Ma
th
iva
n
a
n
)
5189
m
ar
g
in
ally
to
8
2
.
9
7
%
f
o
r
b
o
t
h
T
F
an
d
T
F
-
I
DF,
in
d
icatin
g
a
s
lig
h
t
b
en
ef
it
f
r
o
m
s
m
o
o
th
i
n
g
.
Gib
b
s
s
am
p
lin
g
s
m
o
o
th
in
g
r
esu
lts
in
F1
-
s
co
r
es
o
f
8
5
.
4
7
%
with
T
F
a
n
d
8
5
.
3
6
%
with
T
F
-
I
DF,
s
h
o
win
g
a
m
o
d
er
ate
im
p
r
o
v
em
e
n
t
o
v
er
th
e
s
tan
d
ar
d
HM
M.
L
ap
lace
s
m
o
o
th
in
g
s
h
o
ws
a
b
etter
im
p
r
o
v
em
en
t,
a
ch
iev
in
g
F1
-
s
co
r
es
o
f
8
7
.
0
1
%
with
T
F a
n
d
8
7
.
2
4
%
with
T
F
-
I
DF,
m
ak
in
g
it th
e
b
est
-
p
er
f
o
r
m
in
g
tech
n
iq
u
e
f
o
r
th
is
d
ataset.
Fo
r
th
e
f
r
o
ze
n
f
o
o
d
p
r
o
d
u
cts
d
ataset,
th
e
s
tan
d
ar
d
HM
M
a
ch
iev
es
an
F1
-
s
co
r
e
o
f
6
6
.
8
5
%
with
T
F,
im
p
r
o
v
in
g
to
6
6
.
9
6
%
with
T
F
-
I
DF.
L
ap
lace
s
m
o
o
th
in
g
ag
ai
n
ac
h
iev
es
th
e
h
ig
h
est
F1
-
s
co
r
e
6
7
.
5
0
%
with
T
F
an
d
6
7
.
8
7
%
with
T
F
-
I
DF,
m
ak
in
g
it
th
e
m
o
s
t
ef
f
ec
tiv
e
s
m
o
o
th
in
g
tech
n
i
q
u
e
f
o
r
th
is
d
ataset.
T
h
e
p
er
f
o
r
m
an
ce
s
o
f
HM
M
with
Ab
s
o
lu
te
d
is
co
u
n
t
an
d
Gib
b
s
s
am
p
lin
g
s
m
o
o
th
in
g
tech
n
iq
u
es
ar
e
lo
wer
co
m
p
ar
ed
to
s
tan
d
ar
d
HM
M
.
Ab
s
o
lu
te
d
is
co
u
n
t
s
m
o
o
t
h
in
g
s
u
b
tr
ac
ts
a
f
ix
ed
am
o
u
n
t
f
r
o
m
th
e
co
u
n
ts
o
f
o
b
s
er
v
ed
ev
en
ts
a
n
d
r
ed
is
tr
ib
u
tes
it
to
u
n
s
ee
n
e
v
en
ts
.
W
h
ile
th
is
h
elp
s
h
a
n
d
le
ze
r
o
p
r
o
b
ab
ilit
ies,
it
m
ay
o
v
er
-
s
m
o
o
th
p
r
o
b
ab
ilit
ies
wh
en
o
b
s
er
v
e
d
ev
en
ts
alr
ea
d
y
r
e
f
lect
th
e
u
n
d
e
r
ly
in
g
d
is
tr
ib
u
ti
o
n
well
[
2
2
]
.
Gib
b
s
s
am
p
lin
g
is
a
co
m
p
le
x
iter
ati
v
e
tech
n
iq
u
e
,
an
d
it
ca
n
in
tr
o
d
u
ce
n
o
is
e
if
n
o
t
p
r
o
p
er
ly
tu
n
ed
o
r
co
n
v
er
g
ed
.
T
h
ese
is
s
u
es c
an
r
esu
lt in
a
lo
wer
F1
-
s
co
r
e
co
m
p
ar
ed
t
o
u
s
in
g
a
s
tan
d
ar
d
HM
M
with
o
u
t s
m
o
o
th
in
g
[
2
4
]
.
T
h
e
b
est
F1
-
s
co
r
es
f
o
r
b
o
th
d
atasets
ar
e
ac
h
iev
ed
with
HM
M
u
s
in
g
L
ap
lace
s
m
o
o
th
i
n
g
a
n
d
T
F
-
I
DF
em
b
ed
d
in
g
.
I
n
ad
d
itio
n
,
T
ab
le
3
p
r
esen
ts
th
e
class
if
icati
o
n
r
esu
lts
f
o
r
HM
Ms
with
d
if
f
er
en
t
s
m
o
o
th
in
g
tech
n
iq
u
es
a
p
p
lied
t
o
two
n
ew
tex
t
class
if
icatio
n
d
o
m
ain
s
:
s
p
am
f
ilter
in
g
a
n
d
j
o
b
titl
e
class
if
icatio
n
.
T
h
e
b
est
F1
-
s
co
r
e
f
o
r
class
if
y
in
g
t
h
e
SMS
s
p
am
d
ataset
is
6
9
.
1
7
%
an
d
it
is
ex
ec
u
ted
f
r
o
m
HM
M
u
s
in
g
L
ap
lace
s
m
o
o
th
in
g
a
n
d
T
F
-
I
DF
as
th
e
em
b
ed
d
i
n
g
tech
n
iq
u
e.
T
h
is
r
esu
lt
is
in
lin
e
with
th
e
b
est
m
o
d
el
o
b
tain
ed
f
o
r
class
if
y
in
g
e
-
co
m
m
er
ce
p
r
o
d
u
ct
d
atasets
u
s
ed
in
th
e
s
tu
d
y
.
T
ab
le
2
.
F1
-
s
co
r
es
o
f
HM
M
m
o
d
els f
o
r
e
-
co
m
m
er
ce
p
r
o
d
u
ct
d
atasets
D
a
t
a
C
l
a
s
si
f
i
c
a
t
i
o
n
m
o
d
e
l
Emb
e
d
d
i
n
g
t
e
c
h
n
i
q
u
e
F1
-
sc
o
r
e
(
%)
B
e
st
m
o
d
el
E
-
c
o
mm
e
r
c
e
N
o
n
-
f
o
o
d
p
r
o
d
u
c
t
s
H
M
M
TF
8
2
.
2
7
H
M
M
La
p
l
a
c
e
TF
-
I
D
F
TF
-
I
D
F
8
2
.
0
8
H
M
M
La
p
l
a
c
e
TF
8
7
.
0
1
TF
-
I
D
F
8
7
.
2
4
H
M
M
D
I
S
C
TF
8
2
.
9
7
TF
-
I
D
F
8
2
.
9
7
H
M
M
G
I
B
B
S
TF
8
5
.
4
7
TF
-
I
D
F
8
5
.
3
6
E
-
c
o
mm
e
r
c
e
F
r
o
z
e
n
f
o
o
d
p
r
o
d
u
c
t
s
H
M
M
TF
6
6
.
8
5
H
M
M
La
p
l
a
c
e
TF
-
I
D
F
TF
-
I
D
F
6
6
.
9
6
H
M
M
La
p
l
a
c
e
TF
6
7
.
5
0
TF
-
I
D
F
6
7
.
8
7
H
M
M
D
I
S
C
TF
6
5
.
0
0
TF
-
I
D
F
6
5
.
0
6
H
M
M
G
I
B
B
S
TF
6
6
.
0
1
TF
-
I
D
F
6
6
.
3
1
T
ab
le
3
.
F1
-
s
co
r
es
o
f
HM
M
m
o
d
els f
o
r
s
p
am
f
ilter
in
g
a
n
d
jo
b
titl
e
d
atasets
D
a
t
a
C
l
a
s
si
f
i
c
a
t
i
o
n
m
o
d
e
l
Emb
e
d
d
i
n
g
t
e
c
h
n
i
q
u
e
F1
-
sc
o
r
e
(
%)
B
e
st
m
o
d
e
l
S
M
S
S
p
a
m
C
o
r
p
u
s
H
M
M
TF
6
8
.
8
0
H
M
M
La
p
l
a
c
e
TF
-
I
D
F
TF
-
I
D
F
6
7
.
1
4
H
M
M
La
p
l
a
c
e
TF
6
9
.
0
8
TF
-
I
D
F
6
9
.
1
7
H
M
M
D
I
S
C
TF
6
8
.
3
0
TF
-
I
D
F
6
8
.
4
8
H
M
M
G
I
B
B
S
TF
6
8
.
8
0
TF
-
I
D
F
6
9
.
0
7
Jo
b
t
i
t
l
e
c
o
r
p
u
s
H
M
M
TF
6
8
.
6
4
H
M
M
TF
TF
-
I
D
F
6
7
.
1
7
H
M
M
La
p
l
a
c
e
TF
6
6
.
3
3
TF
-
I
D
F
6
6
.
1
2
H
M
M
D
I
S
C
TF
6
7
.
2
7
TF
-
I
D
F
6
7
.
1
5
H
M
M
G
I
B
B
S
TF
6
6
.
9
1
TF
-
I
D
F
6
6
.
6
2
Ho
wev
er
,
th
e
s
tan
d
ar
d
HM
M
with
o
u
t
s
m
o
o
th
in
g
tech
n
iq
u
e
s
ee
m
ed
to
f
it
th
e
b
est
f
o
r
class
if
y
in
g
jo
b
titl
es.
T
h
e
h
ig
h
est
F1
-
s
co
r
e
is
6
8
.
6
4
%
u
s
in
g
T
F
as
th
e
em
b
ed
d
in
g
tech
n
iq
u
e.
No
n
e
o
f
th
e
s
m
o
o
th
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
1
8
3
-
5
1
9
2
5190
tech
n
iq
u
es
h
elp
ed
in
im
p
r
o
v
i
n
g
th
e
HM
M
p
er
f
o
r
m
a
n
ce
s
.
J
o
b
titl
es
o
f
ten
u
s
e
a
n
ar
r
o
w
v
o
ca
b
u
lar
y
with
r
ep
etitiv
e
ter
m
s
,
ef
f
ec
tiv
ely
ca
p
tu
r
ed
b
y
T
F.
I
n
c
o
n
tr
ast,
TF
-
I
DF
d
im
in
is
h
es
co
m
m
o
n
ter
m
s
,
p
o
ten
tially
r
ed
u
cin
g
th
e
im
p
o
r
tan
ce
o
f
f
r
eq
u
en
t
jo
b
-
r
elate
d
wo
r
d
s
.
A
v
o
id
in
g
s
m
o
o
th
in
g
allo
ws
th
e
m
o
d
el
to
r
ely
s
o
lely
o
n
o
b
s
er
v
ed
d
ata,
b
en
ef
icial
f
o
r
s
m
all
an
d
s
p
ec
if
ic
d
atasets
lik
e
jo
b
titl
e
class
if
icatio
n
.
T
h
u
s
,
th
e
a
n
aly
s
is
r
ev
ea
ls
th
at
s
m
o
o
th
in
g
tech
n
iq
u
es
m
ay
n
o
t
alwa
y
s
im
p
r
o
v
e
HM
M
p
er
f
o
r
m
a
n
ce
s
,
as
t
h
eir
ef
f
ec
tiv
en
ess
d
ep
en
d
s
o
n
th
e
n
atu
r
e
a
n
d
co
m
p
lex
ity
o
f
th
e
d
ata.
No
n
eth
e
less
,
th
e
co
m
b
in
atio
n
o
f
L
ap
l
ac
e
s
m
o
o
th
in
g
a
n
d
TF
-
I
DF m
o
s
t
ly
p
r
o
v
i
d
es th
e
b
est r
esu
lts
a
m
o
n
g
th
e
test
ed
tech
n
iq
u
es.
T
h
e
r
esu
lts
s
h
o
wed
th
at
L
ap
lace
s
m
o
o
th
in
g
c
o
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
ed
o
th
er
tech
n
iq
u
es
in
th
e
e
-
co
m
m
er
ce
an
d
SMS
s
p
am
d
atasets
.
T
h
is
ca
n
b
e
attr
ib
u
te
d
to
t
h
e
s
p
ec
if
ic
c
h
ar
ac
ter
is
tics
o
f
th
ese
d
atasets
,
s
u
ch
as
th
eir
r
elativ
ely
lar
g
er
v
o
ca
b
u
lar
y
s
ize
an
d
h
ig
h
er
d
eg
r
ee
o
f
d
ata
s
p
ar
s
ity
.
As
s
h
o
wn
in
Fig
u
r
e
1
(
p
r
o
d
u
ct
titl
e
len
g
th
s
)
an
d
Fi
g
u
r
e
2
(
SMS
s
p
am
tex
t
len
g
th
s
)
,
th
e
r
esu
lts
alig
n
with
th
ese
d
ataset
f
ea
tu
r
es.
B
o
th
p
r
o
d
u
ct
titl
es
an
d
SM
S
s
p
am
co
n
s
is
t
o
f
r
elativ
ely
s
h
o
r
t
tex
t,
m
ak
in
g
th
em
s
u
itab
le
f
o
r
L
a
p
lace
s
m
o
o
th
in
g
.
I
n
th
ese
ca
s
es,
L
ap
lace
s
m
o
o
t
h
in
g
,
wh
ich
a
d
d
s
a
s
m
all
co
n
s
tan
t
t
o
all
o
b
s
er
v
ed
co
u
n
ts
,
is
b
en
ef
icial
in
ad
d
r
ess
in
g
ze
r
o
-
p
r
o
b
a
b
ilit
y
is
s
u
es.
I
t
en
s
u
r
es
t
h
at
ev
er
y
p
o
s
s
ib
le
ev
en
t,
ev
en
th
o
s
e
n
o
t
o
b
s
er
v
e
d
in
th
e
tr
ai
n
in
g
d
ata,
h
as
a
n
o
n
-
ze
r
o
p
r
o
b
a
b
ilit
y
[
1
2
]
.
T
h
is
ch
a
r
ac
ter
is
tic
m
ak
es
L
ap
lace
s
m
o
o
th
in
g
p
ar
ticu
lar
ly
ef
f
ec
tiv
e
f
o
r
d
atasets
with
a
m
o
r
e
d
iv
e
r
s
e
v
o
ca
b
u
lar
y
,
a
s
it
p
r
ev
e
n
ts
r
ar
e
ter
m
s
f
r
o
m
b
ein
g
c
o
m
p
letely
d
is
r
eg
ar
d
ed
,
th
u
s
im
p
r
o
v
in
g
th
e
m
o
d
el’
s
p
er
f
o
r
m
a
n
ce
.
I
n
co
n
tr
ast,
jo
b
titl
e
class
if
ic
atio
n
d
id
n
o
t
b
e
n
ef
it
f
r
o
m
L
ap
lace
s
m
o
o
th
in
g
,
p
r
im
ar
ily
d
u
e
to
th
e
ex
tr
em
ely
s
h
o
r
t
a
n
d
r
ep
etitiv
e
n
atu
r
e
o
f
th
e
titl
es
,
as
s
h
o
wn
in
Fig
u
r
e
2
(
jo
b
titl
e
tex
t
len
g
th
s
)
.
T
h
e
jo
b
titl
e
d
ataset
is
ch
ar
ac
ter
ized
b
y
a
lim
ited
v
o
ca
b
u
lar
y
,
with
a
h
ig
h
f
r
e
q
u
en
c
y
o
f
ter
m
r
e
p
et
itio
n
.
T
h
is
ty
p
e
o
f
d
ataset,
co
m
b
in
e
d
with
a
s
m
aller
s
am
p
le
s
ize,
r
ed
u
ce
s
t
h
e
n
ec
ess
ity
f
o
r
s
m
o
o
th
in
g
te
ch
n
iq
u
es.
L
ap
lace
s
m
o
o
th
in
g
,
i
n
th
is
in
s
tan
ce
,
i
n
tr
o
d
u
ce
d
u
n
n
ec
ess
ar
y
co
m
p
lex
ity
with
o
u
t
p
r
o
v
id
i
n
g
an
y
c
lear
ad
v
an
tag
e
.
T
h
e
s
tan
d
ar
d
HM
M
m
o
d
el,
r
ely
in
g
o
n
o
b
s
er
v
ed
f
r
eq
u
en
cies,
w
as
ab
le
to
ef
f
ec
tiv
ely
ca
p
tu
r
e
t
h
e
r
elev
an
t
p
atter
n
s
in
th
e
d
ataset.
Fu
r
th
er
m
o
r
e,
t
h
e
T
F
em
b
ed
d
i
n
g
m
eth
o
d
,
w
h
ich
r
etain
s
t
h
e
f
r
eq
u
e
n
cy
o
f
ter
m
s
,
p
r
o
v
ed
m
o
r
e
s
u
itab
le
f
o
r
t
h
is
d
ataset.
I
ts
a
b
ilit
y
to
em
p
h
asize
f
r
e
q
u
en
t
t
er
m
s
co
n
tr
ib
u
ted
to
b
etter
m
o
d
el
p
er
f
o
r
m
a
n
ce
in
jo
b
titl
e
class
if
icatio
n
.
T
h
e
v
a
r
y
i
n
g
p
e
r
f
o
r
m
a
n
c
e
o
f
t
h
e
s
m
o
o
t
h
i
n
g
t
e
c
h
n
i
q
u
e
s
a
c
r
o
s
s
d
a
ta
s
et
s
c
a
n
a
ls
o
b
e
a
t
t
r
i
b
u
t
e
d
t
o
t
h
e
s
t
r
u
c
t
u
r
a
l
d
i
f
f
e
r
e
n
c
e
s
b
et
w
e
e
n
t
h
e
m
.
I
n
m
o
r
e
v
a
r
i
a
b
l
e
d
a
t
a
s
e
t
s
,
s
u
c
h
a
s
t
h
o
s
e
u
s
e
d
f
o
r
s
p
a
m
f
i
l
t
e
r
i
n
g
,
w
h
e
r
e
t
h
e
r
e
i
s
a
b
r
o
a
d
e
r
r
a
n
g
e
o
f
c
o
n
t
e
n
t
a
n
d
d
i
v
e
r
s
i
t
y
i
n
t
h
e
v
o
c
a
b
u
l
a
r
y
,
s
m
o
o
t
h
i
n
g
m
e
t
h
o
d
s
l
i
k
e
L
a
p
l
a
c
e
c
o
n
t
r
i
b
u
t
e
t
o
m
o
d
e
l
g
e
n
e
r
a
l
i
z
a
ti
o
n
.
B
y
p
r
e
v
e
n
t
i
n
g
o
v
e
r
f
i
t
t
i
n
g
t
o
s
p
e
ci
f
i
c
w
o
r
d
p
a
t
t
e
r
n
s
,
L
a
p
l
a
c
e
s
m
o
o
t
h
i
n
g
e
n
h
a
n
c
e
s
c
l
a
s
s
i
f
ic
a
t
i
o
n
a
c
c
u
r
a
c
y
.
H
o
w
e
v
e
r
,
i
n
h
i
g
h
l
y
s
t
r
u
ct
u
r
e
d
d
a
t
as
e
ts
li
k
e
j
o
b
t
it
l
es
,
w
h
e
r
e
t
h
e
v
o
c
a
b
u
l
a
r
y
i
s
r
e
p
e
ti
t
i
v
e
a
n
d
n
a
r
r
o
w
l
y
d
e
f
i
n
e
d
,
s
m
o
o
t
h
in
g
c
a
n
r
e
d
u
c
e
t
h
e
i
m
p
a
ct
o
f
f
r
eq
u
e
n
t
,
i
m
p
o
r
t
a
n
t
t
e
r
m
s
[
2
5
]
.
I
n
s
u
c
h
c
as
es
,
r
e
l
y
i
n
g
o
n
t
h
e
r
a
w
o
b
s
e
r
v
e
d
c
o
u
n
t
s
,
as
d
o
n
e
i
n
t
h
e
s
t
a
n
d
a
r
d
H
MM
a
p
p
r
o
a
c
h
,
p
r
o
v
e
s
m
o
r
e
e
f
f
e
c
t
i
v
e
.
4.
CO
NCLU
SI
O
N
T
h
e
s
tu
d
y
co
m
p
a
r
es
th
e
p
er
f
o
r
m
an
ce
o
f
HM
Ms
u
s
in
g
d
if
f
er
en
t
s
m
o
o
th
i
n
g
tech
n
iq
u
es
.
T
h
e
r
esu
lts
co
n
s
is
ten
tly
f
av
o
r
ed
L
ap
lace
s
m
o
o
th
in
g
f
o
r
th
e
e
-
c
o
m
m
e
r
ce
an
d
SMS
s
p
am
d
atasets
,
d
em
o
n
s
tr
atin
g
its
ef
f
icac
y
i
n
a
d
d
r
ess
in
g
ze
r
o
-
p
r
o
b
a
b
ilit
y
s
ce
n
ar
io
s
ac
r
o
s
s
d
iv
er
s
e
co
n
tex
ts
.
Ho
wev
e
r
,
a
n
o
ta
b
le
d
e
p
ar
tu
r
e
em
er
g
ed
with
th
e
j
o
b
titl
e
d
ataset,
wh
er
e
th
e
HM
M
wit
h
o
u
t
a
n
y
s
m
o
o
th
in
g
tech
n
i
q
u
e
y
ield
ed
s
u
p
er
i
o
r
p
er
f
o
r
m
an
ce
.
T
h
is
d
iv
er
g
en
ce
ca
n
b
e
attr
ib
u
ted
to
th
e
d
is
tin
ct
n
atu
r
e
o
f
jo
b
titl
es,
ch
ar
ac
ter
ized
b
y
a
r
estricte
d
an
d
r
ep
etitiv
e
v
o
ca
b
u
lar
y
.
T
h
e
m
o
d
el
b
e
n
ef
ited
f
r
o
m
ex
ac
t
m
atch
p
r
o
b
a
b
ilit
ies
with
o
u
t
th
e
ad
d
itio
n
al
co
m
p
lex
ity
in
tr
o
d
u
ce
d
b
y
s
m
o
o
th
in
g
m
eth
o
d
s
.
T
h
is
f
in
d
i
n
g
u
n
d
er
s
co
r
es
th
e
e
f
f
ec
tiv
en
e
s
s
o
f
s
im
p
licity
in
s
p
ec
if
ic
d
ata
s
ettin
g
s
,
s
u
g
g
e
s
tin
g
th
at
s
im
p
ler
p
ar
a
m
eter
esti
m
atio
n
ap
p
r
o
ac
h
es
m
ay
b
o
ls
ter
a
n
HM
M
p
er
f
o
r
m
an
ce
i
n
s
im
ilar
class
if
icatio
n
task
s
.
Fu
tu
r
e
r
esear
c
h
co
u
ld
f
o
cu
s
o
n
c
o
m
b
in
i
n
g
d
if
f
er
en
t
s
m
o
o
th
in
g
tech
n
iq
u
es
to
im
p
r
o
v
e
p
er
f
o
r
m
an
ce
,
s
u
ch
as
in
te
g
r
atin
g
L
a
p
lace
s
m
o
o
th
i
n
g
with
o
th
e
r
m
eth
o
d
s
lik
e
ab
s
o
lu
te
d
is
co
u
n
tin
g
.
E
x
p
lo
r
in
g
th
e
i
m
p
ac
t
o
f
s
m
o
o
th
in
g
o
n
ad
v
a
n
ce
d
m
o
d
els,
lik
e
d
ee
p
lear
n
in
g
class
if
ier
s
,
co
u
ld
also
p
r
o
v
id
e
v
alu
ab
le
in
s
ig
h
ts
.
Ad
d
itio
n
ally
,
au
t
o
m
atin
g
th
e
s
elec
tio
n
o
f
th
e
b
est
s
m
o
o
th
i
n
g
tech
n
iq
u
e
b
ased
o
n
d
ataset
ch
ar
ac
ter
is
tics
wo
u
ld
b
e
a
u
s
ef
u
l
d
ir
ec
tio
n
.
Fin
ally
,
ap
p
ly
in
g
th
ese
tech
n
iq
u
es
to
s
p
ec
ialized
f
ield
s
,
lik
e
m
ed
ical
o
r
le
g
al
tex
t c
lass
if
icatio
n
,
co
u
ld
f
u
r
t
h
er
en
h
an
c
e
th
eir
p
r
ac
tical
u
s
e.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
e
r
esear
ch
is
f
u
n
d
ed
b
y
th
e
Un
iv
er
s
ity
T
ek
n
o
lo
g
i
MA
R
A
an
d
Min
is
tr
y
o
f
E
d
u
ca
tio
n
Ma
lay
s
ia
u
n
d
er
t
h
e
Gr
an
t Sch
em
e
(
FR
GS/1
/2
0
1
8
/STG0
6
/UI
T
M/0
1
/
1
).
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
t
r
ib
u
to
r
R
o
les
T
a
x
o
n
o
m
y
(
C
R
ed
iT
)
to
r
ec
o
g
n
ize
in
d
iv
i
d
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
I
mp
a
ct
o
f smo
o
th
in
g
tech
n
i
q
u
es fo
r
text
cla
s
s
if
ica
tio
n
…
(
N
o
r
s
ye
la
Mu
h
a
mma
d
N
o
o
r
Ma
th
iva
n
a
n
)
5191
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
No
r
s
y
ela
Mu
h
am
m
a
d
No
o
r
Ma
th
v
ian
a
n
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
R
o
ziah
Mo
h
d
J
an
o
r
✓
✓
✓
✓
✓
✓
✓
Sh
u
k
o
r
A
b
d
R
az
ak
✓
✓
✓
✓
✓
✓
No
r
Azu
r
a
Md
.
Gh
an
i
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
e
-
c
o
m
m
er
ce
p
r
o
d
u
ct
d
atasets
ar
e
n
o
t
p
u
b
licly
av
ailab
le
as
th
ey
ar
e
p
r
o
p
r
iet
ar
y
to
th
e
Dep
ar
tm
en
t
o
f
Statis
tics
Ma
lay
s
ia.
T
h
e
SMS
Sp
am
C
o
llectio
n
is
p
u
b
l
icly
ac
ce
s
s
ib
le
at:
h
ttp
s
://ar
ch
iv
e.
ics.u
ci.
ed
u
/d
ata
s
et
/2
2
8
/s
m
s
+sp
am
+c
o
llectio
n
,
an
d
th
e
J
o
b
T
itle
C
o
r
p
u
s
ca
n
b
e
ac
ce
s
s
ed
v
ia:
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/k
s
h
itizr
eg
m
i/jo
b
s
-
an
d
-
jo
b
-
d
escr
ip
tio
n
.
RE
F
E
R
E
NC
E
S
[
1
]
T.
Y
a
n
g
,
L
.
H
u
,
C
.
S
h
i
,
H
.
Ji
,
X
.
Li
,
a
n
d
L.
N
i
e
,
“
H
G
A
T:
H
e
t
e
r
o
g
e
n
e
o
u
s
g
r
a
p
h
a
t
t
e
n
t
i
o
n
n
e
t
w
o
r
k
s
f
o
r
se
mi
-
s
u
p
e
r
v
i
se
d
s
h
o
r
t
t
e
x
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
,”
AC
M
T
r
a
n
s
a
c
t
i
o
n
s
o
n
I
n
f
o
rm
a
t
i
o
n
S
y
st
e
m
s
,
v
o
l
.
3
9
,
n
o
.
3
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
4
5
/
3
4
5
0
3
5
2
.
[
2
]
S
.
E
.
V
.
S
.
P
i
l
l
a
i
a
n
d
W
.
C
.
H
u
,
“
M
o
b
i
l
e
t
e
x
t
m
i
si
n
f
o
r
m
a
t
i
o
n
d
e
t
e
c
t
i
o
n
u
s
i
n
g
e
f
f
e
c
t
i
v
e
i
n
f
o
r
ma
t
i
o
n
r
e
t
r
i
e
v
a
l
m
e
t
h
o
d
s
,
”
i
n
I
n
f
o
rm
a
t
i
o
n
S
e
c
u
ri
t
y
a
n
d
Pri
v
a
c
y
i
n
S
m
a
r
t
D
e
v
i
c
e
s:
T
o
o
l
s,
Me
t
h
o
d
s,
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
P
a
l
m
d
a
l
e
,
U
n
i
t
e
d
S
t
a
t
e
s
:
I
G
I
G
l
o
b
a
l
S
c
i
e
n
t
i
f
i
c
P
u
b
l
i
s
h
i
n
g
,
2
0
2
3
,
d
o
i
:
1
0
.
4
0
1
8
/
9
7
8
-
1
-
6
6
8
4
-
5
9
9
1
-
1
.
c
h
0
0
8
.
[
3
]
A
.
O
n
a
n
,
“
S
e
n
t
i
me
n
t
a
n
a
l
y
s
i
s
o
n
mas
si
v
e
o
p
e
n
o
n
l
i
n
e
c
o
u
r
se
e
v
a
l
u
a
t
i
o
n
s:
a
t
e
x
t
mi
n
i
n
g
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
C
o
m
p
u
t
e
r
Ap
p
l
i
c
a
t
i
o
n
s
i
n
E
n
g
i
n
e
e
r
i
n
g
Ed
u
c
a
t
i
o
n
,
v
o
l
.
2
9
,
n
o
.
3
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
2
/
c
a
e
.
2
2
2
5
3
.
[
4
]
L.
K
h
a
n
,
A
.
A
m
j
a
d
,
K
.
M
.
A
f
a
q
,
a
n
d
H
.
T.
C
h
a
n
g
,
“
D
e
e
p
se
n
t
i
m
e
n
t
a
n
a
l
y
si
s
u
s
i
n
g
C
N
N
-
LST
M
a
r
c
h
i
t
e
c
t
u
r
e
o
f
E
n
g
l
i
s
h
a
n
d
R
o
m
a
n
U
r
d
u
t
e
x
t
s
h
a
r
e
d
i
n
s
o
c
i
a
l
m
e
d
i
a
,
”
A
p
p
l
i
e
d
S
c
i
e
n
c
e
s
,
v
o
l
.
1
2
,
n
o
.
5
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
a
p
p
1
2
0
5
2
6
9
4
.
[
5
]
D
.
M
e
n
so
u
r
i
,
A
.
A
z
m
a
n
i
,
a
n
d
M
.
A
z
ma
n
i
,
“
T
o
w
a
r
d
s
a
n
e
-
c
o
mm
e
r
c
e
p
e
r
s
o
n
a
l
i
z
e
d
r
e
c
o
mm
e
n
d
a
t
i
o
n
s
y
s
t
e
m
w
i
t
h
K
N
N
c
l
a
ss
i
f
i
c
a
t
i
o
n
met
h
o
d
,
”
i
n
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
A
d
v
a
n
c
e
d
I
n
t
e
l
l
i
g
e
n
t
S
y
st
e
m
s
f
o
r
S
u
st
a
i
n
a
b
l
e
D
e
v
e
l
o
p
m
e
n
t
,
2
0
2
3
,
p
p
.
3
6
4
-
3
8
2
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
1
-
2
6
3
8
4
-
2
_
3
2
.
[
6
]
D
.
P
a
k
p
a
h
a
n
,
V
.
S
i
a
l
l
a
g
a
n
,
a
n
d
S
.
S
i
r
e
g
a
r
,
“
C
l
a
s
si
f
i
c
a
t
i
o
n
o
f
e
-
c
o
mm
e
r
c
e
p
r
o
d
u
c
t
d
e
scri
p
t
i
o
n
s
w
i
t
h
t
h
e
TF
-
IDF
a
n
d
S
V
M
met
h
o
d
s
,
”
S
i
n
k
r
o
n
,
v
o
l
.
7
,
n
o
.
4
,
p
p
.
2
1
3
0
–
2
1
3
7
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
3
9
5
/
s
i
n
k
r
o
n
.
v
8
i
4
.
1
2
7
7
9
.
[
7
]
S
.
K
a
d
d
o
u
r
a
,
O
.
A
l
f
a
n
d
i
,
a
n
d
N
.
D
a
h
ma
n
i
,
“
A
sp
a
m
e
ma
i
l
d
e
t
e
c
t
i
o
n
m
e
c
h
a
n
i
sm
f
o
r
E
n
g
l
i
s
h
l
a
n
g
u
a
g
e
t
e
x
t
e
mai
l
s
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
,
”
I
EEE
2
9
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
En
a
b
l
i
n
g
T
e
c
h
n
o
l
o
g
i
e
s:
I
n
f
r
a
st
ru
c
t
u
re
f
o
r
C
o
l
l
a
b
o
r
a
t
i
v
e
E
n
t
e
rp
r
i
ses
(
WE
T
I
C
E)
,
p
p
.
1
9
3
–
1
9
8
,
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
W
E
TI
C
E4
9
6
9
2
.
2
0
2
0
.
0
0
0
4
5
.
[
8
]
A
.
S
a
d
i
a
,
F
.
B
a
s
h
i
r
,
R
.
Q
.
K
h
a
n
,
A
.
B
a
s
h
i
r
,
a
n
d
A
.
K
h
a
l
i
d
,
“
C
o
m
p
a
r
i
so
n
o
f
mac
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
f
o
r
sp
a
m
d
e
t
e
c
t
i
o
n
,
”
J
o
u
rn
a
l
o
f
Ad
v
a
n
c
e
s
i
n
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
4
,
n
o
.
2
,
p
p
.
1
7
8
-
1
8
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
2
7
2
0
/
j
a
i
t
.
1
4
.
2
.
1
7
8
-
1
8
4
.
[
9
]
S
.
I
.
M
.
A
l
i
,
M
.
N
i
h
a
d
,
H
.
M
.
S
h
a
r
a
f
,
a
n
d
H
.
F
a
r
o
u
k
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
t
e
x
t
d
o
c
u
m
e
n
t
c
l
a
s
s
i
f
i
c
a
t
i
o
n
-
e
f
f
i
c
i
e
n
t
c
l
a
s
s
i
f
i
c
a
t
i
o
n
a
p
p
r
o
a
c
h
,
”
I
A
E
S
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
r
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
3
,
n
o
.
1
,
p
p
.
7
0
3
-
7
1
0
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
a
i
.
v
1
3
.
i
1
.
p
p
7
0
3
-
710.
[
1
0
]
P
.
F
r
a
sc
o
n
i
,
G
.
S
o
d
a
,
a
n
d
A
.
V
u
l
l
o
,
“
H
i
d
d
e
n
M
a
r
k
o
v
mo
d
e
l
s
f
o
r
t
e
x
t
c
a
t
e
g
o
r
i
z
a
t
i
o
n
i
n
m
u
l
t
i
-
p
a
g
e
d
o
c
u
m
e
n
t
s
,
”
J
o
u
r
n
a
l
o
f
I
n
t
e
l
l
i
g
e
n
t
I
n
f
o
rm
a
t
i
o
n
S
y
s
t
e
m
s
,
v
o
l
.
1
8
,
p
p
.
1
9
5
–
2
1
7
,
2
0
0
2
,
d
o
i
:
1
0
.
1
0
2
3
/
A
:
1
0
1
3
6
8
1
5
2
8
7
4
8
.
[
1
1
]
P
.
H
o
f
ma
n
n
a
n
d
Z.
Ta
s
h
ma
n
,
“
H
i
d
d
e
n
M
a
r
k
o
v
m
o
d
e
l
s
a
n
d
t
h
e
i
r
a
p
p
l
i
c
a
t
i
o
n
f
o
r
p
r
e
d
i
c
t
i
n
g
f
a
i
l
u
r
e
e
v
e
n
t
s,”
i
n
C
o
m
p
u
t
a
t
i
o
n
a
l
S
c
i
e
n
c
e
,
2
0
2
0
,
p
p
.
4
6
4
-
4
7
7
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
030
-
5
0
4
2
0
-
5
_
3
5
.
[
1
2
]
A
.
P
.
N
o
t
o
a
n
d
D
.
R
.
S
.
S
a
p
u
t
r
o
,
“
C
l
a
ssi
f
i
c
a
t
i
o
n
d
a
t
a
mi
n
i
n
g
w
i
t
h
La
p
l
a
c
i
a
n
sm
o
o
t
h
i
n
g
o
n
n
a
ï
v
e
B
a
y
e
s
met
h
o
d
,
”
i
n
AI
P
C
o
n
f
e
re
n
c
e
Pr
o
c
e
e
d
i
n
g
s
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
6
3
/
5
.
0
1
1
6
5
1
9
.
[
1
3
]
H
.
R
e
n
,
Y
.
Z
h
a
o
,
Y
.
Z
h
a
n
g
,
a
n
d
W
.
S
u
n
,
“
Le
a
r
n
i
n
g
l
a
b
e
l
smo
o
t
h
i
n
g
f
o
r
t
e
x
t
c
l
a
ssi
f
i
c
a
t
i
o
n
,
”
Pe
e
rJ
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
v
o
l
.
1
0
,
2
0
2
4
,
d
o
i
:
1
0
.
7
7
1
7
/
p
e
e
r
j
-
c
s.
2
0
0
5
.
[
1
4
]
X
.
W
u
,
C
.
G
a
o
,
M
.
Li
n
,
L.
Za
n
g
,
a
n
d
S
.
H
u
,
“
T
e
x
t
smo
o
t
h
i
n
g
:
e
n
h
a
n
c
e
v
a
r
i
o
u
s
d
a
t
a
a
u
g
me
n
t
a
t
i
o
n
met
h
o
d
s
o
n
t
e
x
t
c
l
a
s
si
f
i
c
a
t
i
o
n
t
a
s
k
s
,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
6
0
t
h
A
n
n
u
a
l
M
e
e
t
i
n
g
o
f
t
h
e
A
sso
c
i
a
t
i
o
n
f
o
r
C
o
m
p
u
t
a
t
i
o
n
a
l
L
i
n
g
u
i
st
i
c
s
,
2
0
2
2
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
2
0
2
2
.
a
c
l
-
s
h
o
r
t
.
9
7
.
[
1
5
]
C
.
F
e
t
t
a
l
,
L
.
La
b
i
o
d
,
a
n
d
M
.
N
a
d
i
f
,
“
M
o
r
e
d
i
scri
m
i
n
a
t
i
v
e
se
n
t
e
n
c
e
e
mb
e
d
d
i
n
g
s
v
i
a
s
e
ma
n
t
i
c
g
r
a
p
h
sm
o
o
t
h
i
n
g
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
1
8
t
h
C
o
n
f
e
r
e
n
c
e
o
f
t
h
e
E
u
r
o
p
e
a
n
C
h
a
p
t
e
r
o
f
t
h
e
As
so
c
i
a
t
i
o
n
f
o
r C
o
m
p
u
t
a
t
i
o
n
a
l
L
i
n
g
u
i
st
i
c
s
,
p
p
.
8
–
13
,
2
0
2
4
.
[
1
6
]
B.
A
.
O
j
o
k
o
h
,
O
.
S
.
A
d
e
w
a
l
e
,
S
.
O
.
F
a
l
a
k
i
,
“
I
mp
r
o
v
i
n
g
o
n
t
h
e
sm
o
o
t
h
i
n
g
t
e
c
h
n
i
q
u
e
f
o
r
o
b
t
a
i
n
i
n
g
e
m
i
ssi
o
n
p
r
o
b
a
b
i
l
i
t
i
e
s
i
n
h
i
d
d
e
n
M
a
r
k
o
v
m
o
d
e
l
s,”
O
ri
e
n
t
a
l
J
o
u
r
n
a
l
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
1
,
n
o
.
1
,
p
p
.
1
5
–
2
4
,
2
0
0
8
.
[
1
7
]
D
.
V
i
l
a
r
,
H
.
N
e
y
,
A
.
Ju
a
n
,
a
n
d
E.
V
i
d
a
l
,
“
Ef
f
e
c
t
o
f
f
e
a
t
u
r
e
sm
o
o
t
h
i
n
g
me
t
h
o
d
s
i
n
t
e
x
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
t
a
s
k
s
,
”
i
n
P
ro
c
e
e
d
i
n
g
s
o
f
t
h
e
4
t
h
I
n
t
e
rn
a
t
i
o
n
a
l
W
o
rks
h
o
p
o
n
Pa
t
t
e
r
n
R
e
c
o
g
n
i
t
i
o
n
i
n
I
n
f
o
rm
a
t
i
o
n
S
y
st
e
m
s
,
p
p
.
1
0
8
–
1
1
7
,
2
0
0
4
.
[
1
8
]
F
.
H
o
s
s
e
y
n
d
o
o
s
t
a
n
d
M
.
T
e
s
h
n
e
h
l
a
b
,
“
I
m
p
r
o
v
i
n
g
h
i
d
d
e
n
M
a
r
k
o
v
m
o
d
e
l
p
e
r
f
o
r
m
a
n
c
e
i
n
p
h
o
n
e
m
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
b
y
f
u
z
z
y
s
m
o
o
t
h
i
n
g
,
”
i
n
E
S
P
O
C
O
'
0
5
:
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
4
t
h
W
S
E
A
S
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
E
l
e
c
t
r
o
n
i
c
,
S
i
g
n
a
l
P
r
o
c
e
s
s
i
n
g
a
n
d
C
o
n
t
r
o
l
,
2
0
0
7
.
[
1
9
]
R
.
F
e
c
h
n
e
r
,
J
.
D
ö
r
p
i
n
g
h
a
u
s,
R
.
R
o
c
k
e
n
f
e
l
l
e
r
,
J
.
F
a
b
e
r
,
“
A
g
e
n
e
r
i
c
f
r
a
me
w
o
r
k
f
o
r
h
i
d
d
e
n
M
a
r
k
o
v
m
o
d
e
l
s
o
n
b
i
o
me
d
i
c
a
l
d
a
t
a
,”
a
rXi
v
:
2
3
0
7
.
1
3
2
8
8
,
2
0
2
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
6
,
Dec
em
b
er
20
25
:
5
1
8
3
-
5
1
9
2
5192
[
2
0
]
N
.
M
.
N
.
M
a
t
h
i
v
a
n
a
n
,
R
.
M
.
J
a
n
o
r
,
S
.
A
.
R
a
z
a
k
,
a
n
d
N
.
A
.
M
.
G
h
a
n
i
,
“
F
e
a
t
u
r
e
su
b
s
t
i
t
u
t
i
o
n
u
si
n
g
l
a
t
e
n
t
D
i
r
i
c
h
l
e
t
a
l
l
o
c
a
t
i
o
n
f
o
r
t
e
x
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
6
,
n
o
.
1
,
p
p
.
1
0
8
7
–
1
0
9
8
,
2
0
2
5
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
2
5
.
0
1
6
0
1
1
0
5
.
[
2
1
]
D
.
Ju
r
a
f
sk
y
a
n
d
J
.
H
.
M
a
r
t
i
n
,
S
p
e
e
c
h
a
n
d
l
a
n
g
u
a
g
e
p
r
o
c
e
ssi
n
g
:
a
n
i
n
t
r
o
d
u
c
t
i
o
n
t
o
n
a
t
u
r
a
l
l
a
n
g
u
a
g
e
p
ro
c
e
ss
i
n
g
,
c
o
m
p
u
t
a
t
i
o
n
a
l
l
i
n
g
u
i
s
t
i
c
s,
a
n
d
s
p
e
e
c
h
re
c
o
g
n
i
t
i
o
n
,
H
o
b
o
k
e
n
,
U
n
i
t
e
d
S
t
a
t
e
s:
P
r
e
n
t
i
c
e
-
H
a
l
l
,
I
n
c
.
,
2
0
0
0
.
[
2
2
]
A
.
S
v
e
t
e
,
N
.
B
o
r
e
n
s
t
e
i
n
,
M
.
Z
h
o
u
,
I
.
A
u
g
e
n
st
e
i
n
,
a
n
d
R
.
C
o
t
t
e
r
e
l
l
,
“
C
a
n
t
r
a
n
sf
o
r
mers
l
e
a
r
n
n
-
g
r
a
m
l
a
n
g
u
a
g
e
mo
d
e
l
s
?
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
2
4
C
o
n
f
e
ren
c
e
o
n
Em
p
i
r
i
c
a
l
M
e
t
h
o
d
s
i
n
N
a
t
u
ra
l
L
a
n
g
u
a
g
e
Pr
o
c
e
ssi
n
g
,
p
p
.
9
8
5
1
–
9
8
6
7
,
d
o
i
:
1
0
.
1
8
6
5
3
/
v
1
/
2
0
2
4
.
e
m
n
l
p
-
m
a
i
n
.
5
5
0
.
[
2
3
]
S
.
Q
i
n
,
G
.
Z
h
a
n
g
,
Y
.
W
u
,
a
n
d
Z.
Z
h
u
,
“
B
a
y
e
s
i
a
n
g
r
o
u
p
i
n
g
-
G
i
b
b
s sa
m
p
l
i
n
g
e
st
i
mat
i
o
n
o
f
h
i
g
h
-
d
i
m
e
n
s
i
o
n
a
l
l
i
n
e
a
r
mo
d
e
l
w
i
t
h
n
o
n
-
sp
a
r
si
t
y
,
”
C
o
m
p
u
t
a
t
i
o
n
a
l
S
t
a
t
i
st
i
c
s
&
D
a
t
a
A
n
a
l
y
si
s
,
v
o
l
.
2
0
3
,
2
0
2
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
sd
a
.
2
0
2
4
.
1
0
8
0
7
2
.
[
2
4
]
A
.
E.
G
e
l
f
a
n
d
,
“
G
i
b
b
s
S
a
m
p
l
i
n
g
,
”
J
o
u
rn
a
l
o
f
t
h
e
Am
e
ri
c
a
n
S
t
a
t
i
st
i
c
a
l
Asso
c
i
a
t
i
o
n
,
v
o
l
.
9
5
,
n
o
.
4
5
2
,
p
p
.
1
3
0
0
–
1
3
0
4
,
2
0
0
0
,
d
o
i
:
1
0
.
2
3
0
7
/
2
6
6
9
7
7
5
.
[
2
5
]
E.
S
e
n
g
e
r
,
Y
.
C
a
mp
b
e
l
l
,
R
.
v
a
n
d
e
r
G
o
o
t
,
a
n
d
B
.
P
l
a
n
k
,
“
K
A
R
R
I
ER
EW
EG
E:
A
l
a
r
g
e
sc
a
l
e
c
a
r
e
e
r
p
a
t
h
p
r
e
d
i
c
t
i
o
n
d
a
t
a
s
e
t
,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
3
1
st
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
a
t
i
o
n
a
l
L
i
n
g
u
i
st
i
c
s:
I
n
d
u
s
t
ry
T
ra
c
k
,
p
p
.
5
3
3
–
5
4
5
,
2
0
2
5
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
No
r
sy
e
la
Mu
h
a
m
m
a
d
No
o
r
M
a
th
iv
a
n
a
n
is
a
P
h
.
D
.
stu
d
e
n
t
a
t
th
e
S
c
h
o
o
l
o
f
M
a
th
e
m
a
ti
c
a
l
S
c
ien
c
e
s,
F
a
c
u
lt
y
o
f
Co
m
p
u
ter
a
n
d
M
a
th
e
m
a
ti
c
a
l
S
c
ien
c
e
s,
Un
iv
e
rsiti
Tek
n
o
lo
g
i
M
ARA
(UiTM
),
M
a
l
a
y
sia
,
u
n
d
e
r
t
h
e
su
p
e
rv
isi
o
n
o
f
No
r
Az
u
ra
M
d
.
G
h
a
n
i
a
n
d
Ro
z
iah
M
o
h
d
Ja
n
o
r.
S
h
e
is
a
lso
a
lec
tu
re
r
a
t
th
e
Un
iv
e
rsity
o
f
Wo
ll
o
n
g
o
n
g
M
a
lay
sia
.
He
r
re
se
a
rc
h
in
tere
sts
a
re
re
late
d
to
n
a
tu
ra
l
lan
g
u
a
g
e
p
r
o
c
e
ss
in
g
,
t
e
x
t
a
n
a
ly
ti
c
s,
a
n
d
m
a
c
h
i
n
e
lea
rn
in
g
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
n
o
rsy
e
la.m
@u
o
w.ed
u
.
m
y
.
Ro
z
ia
h
Mo
h
d
J
a
n
o
r
is
a
re
ti
re
d
P
r
o
fe
ss
o
r
o
f
S
tatisti
c
s
w
h
o
f
o
rm
e
rly
se
rv
e
d
a
s
th
e
Vic
e
-
Ch
a
n
c
e
ll
o
r
o
f
U
n
iv
e
r
siti
Tek
n
o
lo
g
i
M
ARA
(UiT
M
)
M
a
lay
sia
.
He
r
sta
ti
stica
l
a
d
v
a
n
c
e
m
e
n
ts
re
v
o
l
v
e
a
ro
u
n
d
m
o
d
e
li
n
g
i
n
v
o
lv
in
g
m
u
lt
i
-
g
r
o
u
p
late
n
t
v
a
riab
le
m
o
d
e
ls
,
stru
c
tu
ra
l
e
q
u
a
ti
o
n
m
o
d
e
li
n
g
,
a
n
d
d
a
ta
e
n
v
e
lo
p
m
e
n
t
a
n
a
l
y
sis
(DEA
).
S
h
e
h
a
s
ta
u
g
h
t
a
d
v
a
n
c
e
d
sta
ti
stica
l
m
o
d
e
li
n
g
a
t
t
h
e
p
o
st
g
ra
d
u
a
te
le
v
e
l,
e
x
p
l
o
rin
g
m
a
c
h
in
e
le
a
rn
in
g
u
sin
g
R.
He
r
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
fo
c
u
s
o
n
e
n
v
i
ro
n
m
e
n
tal,
so
c
ial,
a
n
d
g
o
v
e
r
n
a
n
c
e
(ES
G
)
c
rit
e
ria
a
n
d
th
e
su
sta
in
a
b
le
d
e
v
e
l
o
p
m
e
n
t
g
o
a
ls
(S
DG
).
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
ro
z
iah
.
jan
o
r@g
m
a
il
.
c
o
m
.
S
h
u
k
o
r
Abd
R
a
z
a
k
is
a
P
r
o
fe
ss
o
r
a
n
d
De
p
u
t
y
Vic
e
Ch
a
n
c
e
ll
o
r
(Re
se
a
rc
h
&
In
n
o
v
a
ti
o
n
)
o
f
Un
iv
e
rsit
i
S
u
lt
a
n
Zai
n
a
l
Ab
i
d
in
(UN
IS
ZA),
T
e
re
n
g
g
a
n
u
,
M
a
lay
sia
.
His
re
se
a
rc
h
in
tere
sts
a
re
o
n
t
h
e
se
c
u
rit
y
issu
e
s
fo
r
M
o
b
il
e
Ad
H
o
c
Ne
two
rk
s,
M
o
b
il
e
IP
v
6
,
Ve
h
icu
lar
Ad
Ho
c
Ne
two
rk
s,
a
n
d
n
e
two
rk
se
c
u
rit
y
.
He
a
lso
a
c
ti
v
e
l
y
c
o
n
d
u
c
ts
se
v
e
ra
l
re
se
a
rc
h
in
d
i
g
it
a
l
fo
re
n
sic
in
v
e
stig
a
ti
o
n
,
wire
les
s
s
e
n
so
r
n
e
two
rk
s,
a
n
d
c
lo
u
d
c
o
m
p
u
ti
n
g
.
He
is
th
e
a
u
th
o
r
a
n
d
c
o
-
a
u
th
o
r
fo
r
m
a
n
y
jo
u
rn
a
ls
a
n
d
c
o
n
fe
re
n
c
e
p
r
o
c
e
e
d
in
g
s
a
t
n
a
ti
o
n
a
l
a
n
d
in
tern
a
ti
o
n
a
l
le
v
e
ls.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
sh
u
k
o
rra
z
a
k
@
u
n
isz
a
.
e
d
u
.
m
y
.
No
r
Az
u
r
a
Md
.
G
h
a
n
i
is
a
P
ro
fe
ss
o
r
a
t
th
e
S
c
h
o
o
l
o
f
M
a
th
e
m
a
ti
c
a
l
S
c
ien
c
e
s
,
F
a
c
u
lt
y
o
f
C
o
m
p
u
ter
a
n
d
M
a
t
h
e
m
a
ti
c
a
l
S
c
ien
c
e
s,
Un
iv
e
rsiti
Te
k
n
o
lo
g
i
M
ARA
,
M
a
lay
sia
,
a
n
d
F
o
rm
e
r
Ch
a
ir
IEE
E
Co
m
p
u
ter
S
o
c
iety
M
a
lay
sia
Ch
a
p
ter.
C
u
rre
n
t
ly
,
s
h
e
se
rv
e
s
a
s
Dire
c
to
r
a
t
Re
se
a
rc
h
M
a
n
a
g
e
m
e
n
t
Ce
n
ter,
Un
iv
e
rsiti
Te
k
n
o
lo
g
i
M
ARA
,
M
a
lay
sia
.
He
r
e
x
p
e
rti
se
is
in
b
ig
d
a
ta,
ima
g
e
p
r
o
c
e
ss
in
g
,
a
rti
ficia
l
n
e
u
ra
l
n
e
tw
o
rk
s,
sta
ti
stica
l
p
a
tt
e
r
n
re
c
o
g
n
it
i
o
n
,
a
n
d
f
o
re
n
sic
sta
ti
stics
.
S
h
e
is
th
e
a
u
th
o
r
o
r
c
o
-
a
u
th
o
r
o
f
m
a
n
y
j
o
u
r
n
a
ls
a
n
d
c
o
n
fe
re
n
c
e
p
ro
c
e
e
d
in
g
s
a
t
n
a
ti
o
n
a
l
a
n
d
in
ter
n
a
ti
o
n
a
l
lev
e
ls.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
a
z
u
ra
@tm
sk
.
u
it
m
.
e
d
u
.
m
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.