I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
8
,
No
.
3
,
J
u
n
e
201
8
,
p
p
.
1
7
2
0
~
1
7
3
0
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v8
i
3
.
p
p
1
7
2
0
-
1730
1720
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
M
usic E
m
o
tion C
la
ss
ificatio
n
b
a
sed
o
n Ly
rics
-
Au
di
o
using
Co
rpus
b
a
sed E
mo
tion
F
ik
a
H
a
s
t
a
rit
a
Ra
ch
m
a
n
1
,
R
iy
a
na
rt
o
Sa
rno
2
,
Cha
s
t
ine F
a
t
icha
h
3
1,
2
,
3
De
p
a
rtm
e
n
t
o
f
In
f
o
rm
a
ti
c
s
,
In
stit
u
t
T
e
k
n
o
lo
g
i
S
e
p
u
l
u
h
No
p
e
m
b
e
r
,
In
d
o
n
e
sia
1
De
p
a
rtme
n
t
o
f
In
f
o
rm
a
ti
c
s,
Un
iv
e
rsit
y
o
f
T
ru
n
o
jo
y
o
M
a
d
u
ra
,
I
n
d
o
n
e
sia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
1
0
,
2
0
1
8
R
ev
i
s
ed
A
p
r
2
,
2
0
1
8
A
cc
ep
ted
A
p
r
11
,
2
0
1
8
M
u
sic
h
a
s
ly
rics
a
n
d
a
u
d
io
.
T
h
a
t‟s
c
o
m
p
o
n
e
n
ts
c
a
n
b
e
a
f
e
a
tu
re
f
o
r
m
u
sic
e
m
o
ti
o
n
c
las
sif
ica
ti
o
n
.
Ly
ric
f
e
a
t
u
re
s we
re
e
x
trac
t
e
d
f
ro
m
tex
t
d
a
t
a
a
n
d
a
u
d
io
f
e
a
tu
re
s
w
e
re
e
x
trac
ted
f
ro
m
a
u
d
io
sig
n
a
l
d
a
ta.In
t
h
e
c
las
si
f
ica
ti
o
n
o
f
e
m
o
ti
o
n
s,
e
m
o
ti
o
n
c
o
rp
u
s
is
re
q
u
ired
f
o
r
ly
rica
l
f
e
a
tu
re
e
x
trac
ti
o
n
.
C
o
rp
u
s
Ba
se
d
E
m
o
ti
o
n
(CBE)
su
c
c
e
e
d
to
in
c
re
a
se
th
e
v
a
lu
e
o
f
F
-
M
e
a
su
re
f
o
r
e
m
o
ti
o
n
c
las
sif
ica
ti
o
n
o
n
tex
t
d
o
c
u
m
e
n
ts.
T
h
e
m
u
sic
d
o
c
u
m
e
n
t
h
a
s
a
n
u
n
stru
c
t
u
re
d
f
o
rm
a
t
c
o
m
p
a
re
d
w
it
h
th
e
a
rti
c
le
tex
t
d
o
c
u
m
e
n
t.
S
o
i
t
re
q
u
ires
g
o
o
d
p
re
p
ro
c
e
ss
in
g
a
n
d
c
o
n
v
e
rsio
n
p
ro
c
e
ss
b
e
f
o
re
c
las
si
f
ica
ti
o
n
p
ro
c
e
ss
.
W
e
u
se
d
M
IRE
X
Da
tas
e
t
f
o
r
th
is
re
se
a
rc
h
.
P
sy
c
h
o
li
n
g
u
isti
c
a
n
d
sty
li
s
ti
c
f
e
a
tu
re
s
w
e
r
e
u
se
d
a
s
l
y
rics
f
e
a
tu
re
s.
P
s
y
c
h
o
li
n
g
u
isti
c
f
e
a
tu
re
wa
s
a
fe
a
tu
re
th
a
t
re
late
d
to
t
h
e
c
a
teg
o
ry
o
f
e
m
o
ti
o
n
.
I
n
t
h
is
re
se
a
rc
h
,
CBE
u
se
d
t
o
su
p
p
o
rt
t
h
e
e
x
trac
ti
o
n
p
ro
c
e
ss
o
f
p
sy
c
h
o
li
n
g
u
isti
c
f
e
a
tu
re
.
S
ty
li
stic
fe
a
tu
re
s
re
late
d
w
it
h
u
sa
g
e
o
f
u
n
iq
u
e
w
o
rd
s
in
th
e
l
y
ri
c
s,
e
.
g
.
„o
o
h
‟,
„a
h
‟
,
„y
e
a
h
‟,
e
t
c
.
En
e
rg
y
,
tem
p
o
ra
l
a
n
d
sp
e
c
tru
m
f
e
a
tu
re
s
we
r
e
e
x
tra
c
t
e
d
f
o
r
a
u
d
io
f
e
a
tu
re
s.
T
h
e
b
e
st
tes
t
re
su
lt
f
o
r
m
u
sic
e
m
o
ti
o
n
c
las
sif
ic
a
ti
o
n
w
a
s
th
e
a
p
p
li
c
a
ti
o
n
o
f
R
a
n
d
o
m
F
o
re
st
m
e
th
o
d
s
f
o
r
l
y
rics
a
n
d
a
u
d
i
o
f
e
a
tu
re
s.
T
h
e
v
a
lu
e
o
f
F
-
m
e
a
su
re
w
a
s
5
6
.
8
%
.
K
ey
w
o
r
d
:
Au
d
io
f
ea
t
u
r
es
C
B
E
C
o
r
p
u
s
b
ased
e
m
o
tio
n
E
m
o
tio
n
L
y
r
ic
f
ea
t
u
r
es
Mu
s
ic
C
las
s
i
f
icatio
n
Co
p
y
rig
h
t
©
2
0
1
8
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Fik
a
Ha
s
tar
ita
R
ac
h
m
an
,
Dep
ar
t
m
en
t o
f
I
n
f
o
r
m
atic
s
,
I
n
s
tit
u
t T
ek
n
o
lo
g
i Sep
u
l
u
h
No
p
em
b
er
,
Jl
.
T
ek
n
ik
Ki
m
ia,
Ged
u
n
g
T
ek
n
ik
I
n
f
o
r
m
ati
k
a
,
Ka
m
p
u
s
I
T
S Su
k
o
lilo
Su
r
ab
a
y
a
6
0
1
1
1
,
I
n
d
o
n
esia.
E
m
ail:
f
i
k
a1
4
@
m
h
s
.
if
.
it
s
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
Mu
s
ic
e
m
o
t
io
n
r
ep
r
ese
n
ted
i
n
2
m
o
d
els
:
ca
teg
o
r
ical
a
n
d
d
i
m
e
n
s
io
n
al
m
o
d
els.
C
ate
g
o
r
ical
an
d
d
i
m
en
s
io
n
al
m
o
d
els
ar
e
ac
t
u
all
y
i
n
ter
r
elate
d
.
E
ac
h
m
o
d
el
h
as
ad
v
a
n
tag
e
s
a
n
d
d
is
ad
v
a
n
tag
e
s
.
C
ateg
o
r
ical
m
o
d
el
u
s
e
h
u
m
an
lan
g
u
ag
e
f
o
r
ca
teg
o
r
y
lab
el,
s
o
it
's
ea
s
y
f
o
r
u
s
er
to
u
n
d
er
s
ta
n
d
it.
T
h
e
d
im
en
s
io
n
a
l
m
o
d
el
i
s
an
e
m
o
t
io
n
al
m
o
d
el
t
h
at
d
es
cr
ib
es
e
m
o
tio
n
s
i
n
a
d
i
m
en
s
i
o
n
al
v
ec
to
r
s
p
ac
e
.
T
h
e
C
o
r
p
u
s
t
h
at
u
s
ed
f
o
r
t
h
e
C
at
eg
o
r
ical
m
o
d
el
r
e
f
er
en
ce
i
s
W
o
r
d
Net
Af
f
ec
t
o
f
E
m
o
tio
n
(
W
NA
)
[
1
]
.
W
N
A
i
s
a
d
e
v
el
o
p
m
e
n
t
o
f
W
o
r
d
Ne
t
th
at
h
av
e
a
n
e
m
o
tio
n
al
l
ab
el
s
b
ased
o
n
E
k
m
a
n
E
m
o
tio
n
[
2
]
.
T
h
er
e
ar
e
6
ca
teg
o
r
ies
o
f
e
m
o
tio
n
s
:
s
ad
n
e
s
s
,
an
g
er
,
j
o
y
,
d
is
g
u
s
t,
f
ea
r
,
an
d
s
u
r
p
r
is
e.
W
h
ile
Af
f
ec
ti
v
e
No
r
m
s
f
o
r
E
n
g
li
s
h
W
o
r
d
s
(
A
NE
W
)
is
th
e
d
ataset
th
at
is
o
f
te
n
u
s
ed
f
o
r
r
esear
ch
o
n
d
i
m
en
s
io
n
al
m
o
d
el
[
3
]
.
E
ac
h
ter
m
i
n
A
NE
W
co
n
tai
n
ed
3
d
im
e
n
s
io
n
v
al
u
e
s
.
T
h
er
e
ar
e
Vale
n
ce
,
A
r
o
u
s
al
,
an
d
Do
m
i
n
a
n
ce
.
Vale
n
ce
an
d
A
r
o
u
s
al
ar
e
s
ee
n
f
r
o
m
ea
ch
p
er
s
o
n
al.
Vale
n
c
e
s
a
m
e
a
s
le
v
el
o
f
p
leas
u
r
e.
I
n
v
ec
to
r
s
p
ac
e,
th
e
r
an
g
e
o
f
v
al
en
ce
v
al
u
e
f
r
o
m
n
e
g
ati
v
e
in
to
p
o
s
itif
a
n
d
ar
o
u
s
a
l
h
as
v
alu
e
lo
w
u
n
til
h
ig
h
.
Di
f
f
er
en
t
w
i
th
Vale
n
ce
a
n
d
A
r
o
u
s
al,
Do
m
in
a
n
ce
is
r
elat
io
n
b
et
w
ee
n
p
eo
p
le
an
d
th
eir
e
n
v
ir
o
n
m
e
n
t.
A
lt
h
o
u
g
h
an
y
r
esear
ch
f
o
c
u
s
to
co
m
b
i
n
atio
n
2
m
o
d
el
[
3
]
,
b
u
t
m
o
r
e
r
esear
ch
u
s
i
n
g
o
n
l
y
o
n
e
e
m
o
t
io
n
al
m
o
d
el
(
i.e
.
[
4
]
-
[
7
]
)
.
Mu
s
ic
h
a
s
l
y
r
ics
an
d
a
u
d
io
f
ea
tu
r
es
th
at
ca
n
b
e
u
s
ed
f
o
r
a
r
ef
er
en
ce
in
m
u
s
ic
cla
s
s
i
f
ic
atio
n
.
T
h
e
l
y
r
ics
o
f
m
u
s
ic
ar
e
m
o
r
e
d
o
m
in
an
t
to
Vale
n
ce
th
a
n
A
r
o
u
s
al
.
A
n
d
a
u
d
io
o
f
m
u
s
ic
m
o
r
e
r
e
p
r
esen
t
to
A
r
o
u
s
a
l
d
i
m
en
s
io
n
.
T
h
e
p
r
ev
io
u
s
r
ese
ar
ch
th
at
d
is
c
u
s
s
es
t
h
e
ex
tr
ac
tio
n
o
f
e
m
o
t
io
n
s
f
r
o
m
a
u
d
io
[
8
]
-
[
1
0
]
h
av
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
Mu
s
ic
E
mo
tio
n
C
la
s
s
ifica
tio
n
b
a
s
ed
o
n
Lyr
ics
-
A
u
d
io
u
s
in
g
C
B
E
(
F
ika
Ha
s
ta
r
ita
R
a
ch
ma
n
)
1721
co
n
clu
s
io
n
t
h
at
a
u
d
io
ca
n
b
e
a
f
ea
t
u
r
e
in
m
u
s
ic
e
m
o
tio
n
cl
ass
i
f
icatio
n
.
Si
m
ilar
l
y
w
it
h
t
h
e
p
r
ev
io
u
s
r
esear
c
h
u
s
i
n
g
l
y
r
ic
s
ex
tr
ac
tio
n
[
7
]
-
[
1
0
]
,
it
p
r
o
v
es
th
at
l
y
r
ic
s
ar
e
i
m
p
o
r
tan
t
to
o
f
o
r
f
ea
tu
r
e
in
m
u
s
ic
e
m
o
tio
n
class
i
f
icatio
n
.
T
w
o
m
ai
n
f
ea
t
u
r
es
o
f
m
u
s
ic,
au
d
io
an
d
l
y
r
ics
,
w
as
u
s
ed
in
r
esear
ch
[
1
1
]
-
[
1
3
]
.
B
u
t
in
t
h
e
p
r
o
ce
s
s
,
th
e
y
d
id
n
‟
t
u
s
e
a
co
m
b
in
at
io
n
o
f
d
i
m
e
n
s
io
n
al
a
n
d
ca
teg
o
r
ica
l
m
o
d
els.
T
h
e
p
u
r
p
o
s
e
o
f
t
h
i
s
r
esear
ch
i
s
to
p
r
o
v
e
th
e
lin
k
a
g
e
o
f
ca
teg
o
r
ical
m
o
d
els
an
d
d
i
m
en
s
io
n
al
m
o
d
els
f
o
r
class
if
ica
tio
n
o
f
e
m
o
tio
n
s
i
n
m
u
s
ic.
W
e
u
s
e
th
e
co
m
b
i
n
ed
e
m
o
tio
n
al
co
r
p
u
s
b
et
w
ee
n
th
e
d
i
m
e
n
s
io
n
al
m
o
d
e
l
an
d
th
e
c
ate
g
o
r
y
m
o
d
el
f
o
r
ex
tr
ac
tio
n
o
f
l
y
r
ic
f
ea
t
u
r
e
.
T
h
is
is
o
u
r
co
n
tr
ib
u
ted
f
o
r
th
is
r
esear
c
h
.
L
y
r
ic
s
b
ec
o
m
e
a
n
i
m
p
o
r
ta
n
t
p
ar
t
in
th
e
d
etec
tio
n
o
f
e
m
o
tio
n
s
.
On
e
o
f
t
h
e
l
y
r
ic
s
f
ea
t
u
r
es
is
p
s
y
ch
o
li
n
g
u
is
tic
f
ea
t
u
r
e.
th
ese
f
ea
tu
r
es
ca
n
b
e
p
r
esen
ted
d
if
f
er
en
tl
y
d
ep
en
d
i
n
g
o
n
m
o
d
el
o
f
e
m
o
tio
n
an
d
t
y
p
e
o
f
co
r
p
u
s
u
s
ed
.
Vip
in
Ku
m
ar
e
x
tr
ac
t
p
s
y
co
li
n
g
u
i
s
tic
f
ea
t
u
r
es
l
y
r
ic
s
f
r
o
m
Se
n
ti
w
o
r
d
n
et
[
1
4
]
.
Sen
t
i
w
o
r
d
n
et
i
s
a
co
r
p
u
s
th
at
h
a
s
a
p
o
s
it
iv
e
-
n
e
g
ati
v
e
s
co
r
e
[
1
5
]
.
Usi
n
g
s
en
ti
w
o
r
d
n
et,
t
h
e
l
y
r
ics
an
al
y
ze
d
p
o
s
itiv
e
-
n
e
g
ati
v
e
o
f
s
en
ti
m
e
n
t
v
al
u
e
n
o
t
e
m
o
tio
n
al
v
al
u
e
.
D
i
m
e
n
s
io
n
al
m
o
d
el
s
ca
n
also
af
f
ec
t
th
e
v
a
lu
e
o
f
p
s
y
ch
o
li
n
g
u
i
s
tic
f
ea
t
u
r
es.
W
ith
an
A
NE
W
e
m
o
tio
n
co
r
p
u
s
,
th
is
f
ea
t
u
r
e
ca
n
b
e
w
o
r
t
h
th
e
v
alu
e
o
f
th
e
v
alan
ce
an
d
ar
o
u
s
al
d
i
m
en
s
io
n
s
[
1
6
]
.
Ou
r
r
esear
ch
u
s
e
C
B
E
f
o
r
ex
tr
ac
tin
g
p
s
y
c
o
lin
g
u
i
s
tic
f
ea
t
u
r
e
o
f
e
m
o
tio
n
f
r
o
m
l
y
r
ic.
C
o
r
p
u
s
B
ased
E
m
o
tio
n
(
C
B
E
)
ap
p
lies
th
e
co
m
b
in
ed
co
n
ce
p
t
o
f
ca
teg
o
r
ical
an
d
d
i
m
en
s
io
n
al
d
atasets
.
No
t
o
n
l
y
co
m
b
i
n
i
n
g
,
b
u
t
also
e
x
p
an
d
c
o
r
p
u
s
u
s
i
n
g
s
i
m
ilar
it
y
w
o
r
d
a
n
d
eu
c
lid
ea
n
d
i
s
tan
ce
co
n
ce
p
t
s
.
Gen
er
al
I
n
q
u
ir
er
(
GI
)
an
d
W
o
r
d
n
et
d
atasets
ar
e
u
s
ed
to
s
u
p
p
o
r
t th
e
s
u
cc
es
s
o
f
th
is
r
esear
c
h
to
o
.
A
u
d
io
s
i
g
n
al
ca
n
b
e
f
r
o
m
s
p
e
ec
h
o
r
n
o
t.
T
h
e
s
p
ee
ch
f
ea
tu
r
e
is
tak
e
n
f
r
o
m
t
h
e
h
u
m
a
n
v
o
i
ce
w
it
h
o
u
t
th
e
i
n
s
tr
u
m
en
t.
Sp
ee
ch
f
ea
t
u
r
e
ca
n
al
s
o
b
e
class
if
ied
i
n
to
e
m
o
tio
n
[
1
7
]
.
B
u
t,
f
o
r
t
h
is
m
u
s
ic
d
o
cu
m
e
n
t,
t
h
e
au
d
io
f
ea
t
u
r
es
to
b
e
u
s
ed
ar
e
n
o
n
-
s
p
ee
ch
-
s
h
ap
ed
s
ig
n
als
o
r
w
a
v
s
i
g
n
als.
Au
d
io
ca
n
b
e
ex
t
r
ac
ted
in
to
Stan
d
ar
d
Au
d
io
an
d
Me
lo
d
ic
A
u
d
io
f
ea
tu
r
es
.
U
s
i
n
g
th
e
ap
p
licatio
n
o
f
to
o
lb
o
x
,
ex
tr
ac
ted
au
d
io
f
ea
tu
r
es
ca
n
r
ea
ch
m
o
r
e
th
an
a
h
u
n
d
r
ed
.
T
h
e
ap
p
licatio
n
o
f
t
h
e
r
elief
F
al
g
o
r
ith
m
a
n
d
P
C
A
(
P
r
in
cip
le
C
o
m
p
o
n
e
n
t
An
al
y
s
i
s
)
is
u
s
ed
f
o
r
r
ed
u
ctin
g
d
i
m
e
n
s
io
n
an
d
s
elec
tio
n
o
f
au
d
io
f
ea
t
u
r
es,
s
o
it
is
k
n
o
w
n
w
h
ic
h
f
ea
tu
r
es
ar
e
m
o
r
e
im
p
o
r
tan
t
to
u
s
e
[
4
]
.
Van
L
o
i
Ng
u
y
e
n
[
1
8
]
d
iv
id
ed
au
d
io
f
ea
tu
r
es
i
n
to
t
w
o
s
u
b
s
ets
o
f
d
i
m
e
n
s
io
n
:
A
r
o
u
s
al
an
d
Vale
n
ce
.
T
h
e
s
u
b
s
et
s
is
u
s
ed
f
o
r
e
m
o
tio
n
al
class
i
f
ic
atio
n
w
it
h
th
e
d
i
m
e
n
s
io
n
al
m
o
d
el
an
d
co
n
v
er
t
it
i
n
t
o
ca
teg
o
r
ical
u
s
in
g
T
h
ay
er
‟s
m
o
d
el
.
9
k
i
n
d
s
o
f
s
p
ec
tr
al
s
h
ap
e
au
d
io
ex
tr
ac
tio
n
r
esu
l
ts
ca
n
also
b
e
u
s
ed
as
a
f
ea
tu
r
e
o
f
m
u
s
ic
e
m
o
tio
n
cla
s
s
i
f
icat
io
n
[
1
9
]
.
R
o
u
g
h
n
es
s
f
ea
tu
r
e
i
n
a
u
d
io
s
p
ec
tr
u
m
is
ca
n
b
e
m
ea
n
t
a
s
s
p
ec
tr
al
f
l
u
x
.
A
n
d
co
n
tin
u
ed
i
m
p
le
m
e
n
tatio
n
o
f
e
m
o
tio
n
al
c
lasi
f
icat
io
n
ca
n
le
ad
to
ap
p
licatio
n
in
t
h
e
m
u
s
i
c
r
ec
o
m
m
e
n
d
atio
n
s
y
s
te
m
[
2
0
]
.
Ou
r
r
esear
ch
w
ill
b
e
test
ed
u
s
i
n
g
s
ta
n
d
ar
d
au
d
io
f
ea
tu
r
e
s
o
b
tain
ed
f
r
o
m
th
e
T
o
o
lb
o
x
:
MI
R
T
o
o
lb
o
x
an
d
P
s
y
s
o
u
n
d
.
I
n
ca
teg
o
r
ical
m
u
s
ic
o
f
e
m
o
ti
o
n
,
th
er
e
h
a
v
e
b
ee
n
p
r
ev
io
u
s
r
esear
ch
u
s
in
g
au
d
io
an
d
l
y
r
i
c
f
ea
tu
r
e
s
[
1
2
]
,
[
1
3
]
.
B
u
t th
e
t
w
o
o
f
r
ese
ar
ch
d
id
n
o
t
u
s
e
t
h
e
e
m
o
tio
n
al
co
r
p
u
s
i
n
it
s
f
ea
t
u
r
e
e
x
tr
ac
tio
n
p
r
o
ce
s
s
.
O
n
e
u
s
e
s
th
e
J
l
y
r
ic
s
f
r
a
m
e
w
o
r
k
to
o
b
tain
s
ta
tis
t
ical
f
ea
tu
r
e
s
[
1
2
]
.
An
d
o
th
er
s
s
ee
th
e
w
o
r
d
s
p
ar
cit
y
th
at
ap
p
ea
r
s
i
n
t
h
e
l
y
r
ics
[
1
3
]
.
T
h
e
d
if
f
er
e
n
ce
o
f
t
h
is
r
esear
c
h
w
i
th
p
r
ev
io
u
s
r
esear
c
h
i
s
o
n
t
h
e
l
y
r
ics
an
d
a
u
d
io
f
ea
t
u
r
es.
L
y
r
ic
s
f
ea
t
u
r
es
t
h
at
ex
tr
ac
ted
is
a
c
o
m
b
in
at
io
n
o
f
p
s
y
co
li
n
g
u
i
s
tic
a
n
d
s
t
y
lis
tic
f
ea
tu
r
es.
W
h
ile
t
h
e
au
d
io
f
ea
t
u
r
e
u
s
ed
w
a
s
ta
k
e
n
f
r
o
m
MI
R
T
o
o
lb
o
x
an
d
P
s
y
s
o
u
n
d
.
P
r
ev
io
u
s
r
esear
ch
u
s
i
n
g
a
u
d
io
an
d
l
y
r
ics
f
ea
tu
r
e
s
w
i
th
ca
teg
o
r
ical
m
o
d
el
ap
p
r
o
ac
h
.
T
h
is
r
esear
ch
w
ill
b
e
co
m
b
i
n
ed
ly
r
ic
s
an
d
a
u
d
io
f
ea
t
u
r
es
w
it
h
t
h
e
ap
p
r
o
ac
h
o
f
th
e
t
w
o
m
o
d
els o
f
e
m
o
tio
n
,
ca
teg
o
r
ical
an
d
Di
m
en
s
io
n
al.
2.
RE
S
E
ARCH
M
E
T
H
O
D
E
m
o
tio
n
d
etec
tio
n
p
r
o
ce
s
s
to
b
e
p
er
f
o
r
m
ed
in
t
h
i
s
r
esear
ch
in
cl
u
d
e
m
u
lti
m
o
d
al
f
ea
t
u
r
es.
T
h
e
m
u
s
ic
f
ea
t
u
r
es
e
x
tr
ac
ted
f
r
o
m
l
y
r
ic
s
a
n
d
au
d
io
co
m
p
o
n
en
t
s
.
F
o
r
m
al
s
en
te
n
ce
s
tr
u
c
tu
r
e
is
n
o
t
o
w
n
ed
b
y
l
y
r
ics
.
L
y
r
ic
s
h
as
a
s
m
al
l
o
f
w
o
r
d
s
w
it
h
li
m
ited
v
o
ca
b
u
lar
y
.
I
n
t
h
e
l
y
r
ic
s
t
h
er
e
is
a
p
h
r
a
s
e
o
r
i
d
eo
m
th
at
m
ak
e
s
it
d
if
f
ic
u
lt
to
k
n
o
w
th
e
tr
u
e
m
e
an
in
g
.
I
t
is
a
c
h
alle
n
g
e
to
b
e
ab
le
to
ex
p
r
es
s
t
h
e
e
m
o
tio
n
o
f
m
u
s
ic
b
ased
o
n
l
y
r
ics.
T
h
er
e
ar
e
s
ev
er
al
f
ea
tu
r
es
th
at
ca
n
b
e
e
x
tr
ac
ted
:
p
s
y
c
h
o
lin
g
u
is
tic
a
n
d
s
t
y
li
s
ti
c
f
ea
t
u
r
es
o
f
te
x
t
.
P
s
y
c
h
o
li
n
g
u
i
s
tic
f
ea
t
u
r
es
ar
e
p
s
y
ch
o
lo
g
ical
o
f
la
n
g
u
a
g
e
f
ea
tu
r
es
i
n
t
h
e
l
y
r
ic
s
.
T
h
is
f
ea
tu
r
e
ca
n
b
e
f
o
u
n
d
w
it
h
th
e
h
elp
o
f
e
m
o
tio
n
al
co
r
p
u
s
:
GI
an
d
C
B
E
.
St
y
li
s
tic
f
ea
t
u
r
es
o
f
l
y
r
ic
s
ar
e
i
n
ter
j
ec
tio
n
w
o
r
d
s
(
e.
g
.
,
"
o
o
h
,
"
"
ah
")
an
d
s
p
ec
ial
p
u
n
ct
u
atio
n
s
(
e.
g
.
,
"
!
,
"
"?
"
)
.
A
u
d
io
f
e
atu
r
es
e
x
tr
ac
ted
u
s
i
n
g
to
o
lb
o
x
P
s
y
s
o
u
n
d
3
an
d
MI
R
T
o
o
lb
o
x
.
Fig
u
r
e
1
is
a
p
r
o
p
o
s
ed
m
o
d
el
in
t
h
i
s
r
esea
r
ch
.
T
h
e
Featu
r
e
s
i
n
clu
d
e
f
ea
tu
r
e
o
f
e
n
er
g
y
a
n
d
f
ea
t
u
r
es
o
f
s
p
ec
tr
u
m
.
W
e
u
s
e
d
2
m
ai
n
f
ea
tu
r
e,
b
ec
au
s
e
en
e
r
g
y
a
n
d
s
p
ec
tr
u
m
o
f
a
u
d
io
al
w
a
y
s
s
u
cc
e
s
s
f
u
l
f
o
r
d
etec
tio
n
e
m
o
tio
n
[
1
6
]
,
[
1
3
]
.
D
y
n
a
m
ic
lo
u
d
n
ess
r
ep
r
esen
t
f
r
o
m
e
n
er
g
y
f
ea
t
u
r
e.
R
o
u
g
h
n
es
s
an
d
i
n
h
ar
m
o
n
icit
y
r
ep
r
esen
t sp
ec
tr
u
m
f
ea
t
u
r
e.
T
h
e
r
esu
lt o
f
f
ea
t
u
r
e
ex
tr
ac
tio
n
a
r
e
u
s
ed
in
m
u
s
ic
e
m
o
t
io
n
clas
s
if
ica
tio
n
.
T
h
e
d
ataset
u
s
ed
f
o
r
th
e
ex
tr
a
ctio
n
o
f
p
s
y
c
h
o
li
n
g
u
is
tic
f
ea
t
u
r
es
i
n
t
h
e
l
y
r
ics
ar
e
ANE
W
,
W
NA
a
n
d
GI
.
A
NE
W
d
ataset
is
o
f
ten
u
s
e
d
as
a
r
ef
er
en
ce
in
e
m
o
tio
n
al
d
etec
tio
n
f
o
r
d
i
m
en
s
io
n
a
l
m
o
d
els.
W
h
ile
W
N
A
is
a
r
ef
er
en
ce
d
ataset
u
s
ed
in
e
m
o
tio
n
al
d
etec
tio
n
f
o
r
ca
teg
o
r
ical
m
o
d
el.
I
n
A
NE
W
th
er
e
is
a
b
ig
o
f
w
o
r
d
w
it
h
e
m
o
tio
n
lab
el.
Si
x
b
asic e
m
o
ti
o
n
s
o
f
E
k
m
an
ar
e
u
s
ed
as i
ts
e
m
o
tio
n
al
lab
el
s
.
Data
d
is
tr
ib
u
t
io
n
af
ter
t
h
e
m
er
g
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
3
,
J
u
n
e
2
0
1
8
:
1
7
2
0
–
1730
1722
p
r
o
ce
s
s
b
et
w
ee
n
A
NE
W
an
d
W
NA
is
f
r
o
m
1
0
3
0
ter
m
s
i
n
A
NE
W
an
d
1
1
9
7
o
f
W
NA
.
T
h
er
e
ar
e
1
0
5
ter
m
s
th
at
h
a
v
e
Vala
n
ce
-
A
r
o
u
s
a
l
-
Do
m
i
n
an
ce
v
al
u
e
an
d
e
m
o
tio
n
al
lab
el
(
Fig
u
r
e
2
)
.
C
o
r
p
u
s
B
ase
d
E
m
o
tio
n
al
(
C
B
E
)
r
ep
r
esen
ts
2
m
o
d
el,
ca
teg
o
r
ica
l a
n
d
d
i
m
en
s
io
n
a
l
m
o
d
els.
Fig
u
r
e
1
.
P
r
o
p
o
s
ed
m
o
d
el
Fig
u
r
e
2
.
Data
d
is
tr
ib
u
tio
n
i
n
A
NE
W
an
d
W
N
A
d
ataset
s
I
n
th
i
s
r
esear
ch
,
d
ata
r
es
u
lt
o
r
p
r
ed
ictio
n
u
s
ed
ca
teg
o
r
ical
m
o
d
el
o
n
l
y
.
B
u
t
to
o
b
tain
p
s
y
c
h
o
lin
g
u
i
s
tic
f
ea
t
u
r
es
s
till
r
e
f
e
r
s
to
th
e
C
B
E
w
h
ich
r
ep
r
esen
t
s
ca
teg
o
r
ic
al
an
d
d
im
e
n
s
io
n
al
m
o
d
el
s
.
C
B
E
is
a
co
m
b
i
n
ed
co
r
p
u
s
b
et
w
ee
n
ANE
W
an
d
W
NA
w
h
ic
h
h
as
p
r
o
ce
d
u
r
es
o
f
au
to
m
atic
i
n
co
m
p
le
te
d
ata
an
d
C
B
E
e
x
p
an
d
[
2
1
]
.
A
u
to
m
ated
d
ata
p
r
o
ce
d
u
r
es
f
o
r
i
n
co
m
p
le
te
d
ata
ar
e
u
s
ed
f
o
r
d
ata
t
h
at
h
a
s
n
o
lab
el
o
r
d
i
m
e
n
s
io
n
v
alu
e
s
.
T
h
e
co
n
ce
p
t
o
f
Me
r
g
i
n
g
W
NA
an
d
A
NE
W
,
ca
u
s
i
n
g
i
n
c
o
m
p
lete
d
ata.
I
n
[
2
1
]
,
I
SEAR
Data
s
et
u
s
ed
f
o
r
ex
p
an
d
C
B
E
.
Fig
u
r
e
3
s
h
o
w
t
h
e
C
B
E
s
c
h
e
m
e
[
2
1
]
.
Fig
u
r
e
3
.
C
B
E
s
ch
e
m
e
[
2
1
]
W
N
A
A
N
E
W
M
e
r
g
i
n
g
o
f
d
a
t
a
A
u
t
o
t
a
g
g
i
n
g
d
a
t
a
i
n
c
o
mp
l
e
t
e
Ex
p
a
n
d
CBE
1
092
1
05
925
WN
A
(
6 l
ab
e
l
e
mo
si
)
A
N
E
W
(
n
i
l
a
i
d
i
me
n
si
V
A
D
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
Mu
s
ic
E
mo
tio
n
C
la
s
s
ifica
tio
n
b
a
s
ed
o
n
Lyr
ics
-
A
u
d
io
u
s
in
g
C
B
E
(
F
ika
Ha
s
ta
r
ita
R
a
ch
ma
n
)
1723
I
t
ca
n
b
e
s
ee
n
i
n
Fi
g
u
r
e
2
,
th
e
r
e
ar
e
2
0
1
7
d
ata
is
in
co
m
p
lete
.
E
v
en
f
o
r
n
e
w
d
ata,
it
is
ce
r
ta
in
t
h
at
th
e
ter
m
h
as
n
o
lab
el
o
r
d
i
m
e
n
s
io
n
v
al
u
e.
A
l
g
o
r
ith
m
f
o
r
Au
to
t
ag
g
i
n
g
o
f
i
n
co
m
p
lete
d
ata
w
a
s
cr
ea
ted
to
h
an
d
le
th
is
ca
s
e.
T
h
is
p
r
o
ce
d
u
r
e
is
m
ad
e
w
it
h
t
h
e
co
n
ce
p
t
o
f
s
y
n
o
n
y
m
o
f
w
o
r
d
(
s
y
n
s
et)
,
r
elate
d
n
ess
m
ea
s
u
r
e
a
n
d
E
u
clid
ea
n
Dis
ta
n
ce
.
T
h
e
m
e
th
o
d
o
f
r
elate
d
n
ess
m
ea
s
u
r
e
u
s
e
d
A
d
ap
ted
L
E
S
K
[
2
2
]
.
W
e
h
av
e
tes
t
th
e
P
ea
r
s
o
n
C
o
r
r
elatio
n
v
al
u
e
o
f
A
d
ap
ted
L
E
SK
a
n
d
E
u
clid
ea
n
Dis
t
an
ce
,
th
e
v
al
u
e
is
-
0
.
9
2
.
I
t
m
ea
n
s
t
h
er
e
is
a
r
elatio
n
s
h
ip
b
et
w
ee
n
A
d
ap
ted
L
E
SK
an
d
E
u
clid
ea
n
D
is
ta
n
c
e.
T
h
e
T
-
test
v
al
u
e
b
et
w
ee
n
A
d
ap
ted
L
E
SK
w
it
h
E
u
clid
ea
n
Di
s
ta
n
ce
in
ter
m
s
lab
eled
'
j
o
y
'
is
6
,
6
0
4
3
,
h
en
ce
th
e
r
e
ca
n
b
e
co
r
r
elatio
n
b
et
w
ee
n
th
at
v
ar
iab
le.
C
B
E
ex
p
an
d
al
g
o
r
ith
m
i
s
also
ad
d
ed
to
ex
p
an
d
th
e
co
r
p
u
s
.
T
h
e
d
ata
test
i
n
g
is
d
ata
t
h
at
n
o
t
co
n
tai
n
ed
in
th
e
p
r
ev
io
u
s
C
B
E
.
T
h
u
s
,
C
B
E
is
co
r
p
u
s
o
f
e
m
o
tio
n
al
ter
m
t
h
at
h
as
a
V
A
D
d
i
m
e
n
s
io
n
v
al
u
e
an
d
an
e
m
o
t
io
n
lab
el.
Fig
u
r
e
4.
Fil
ter
in
g
ter
m
u
s
i
n
g
s
y
n
s
et
i
n
W
o
r
d
N
et
B
ef
o
r
e
th
e
a
u
ta
g
g
i
n
g
o
f
i
n
c
o
m
p
lete
d
ata
p
r
o
ce
d
u
r
e
is
e
x
ec
u
ted
,
i
t
is
n
ec
e
s
s
ar
y
to
ch
ec
k
th
e
s
y
n
o
n
y
m
s
o
f
ter
m
i
n
W
o
r
d
Net
(
Fig
u
r
e
4
)
.
I
f
a
s
y
n
o
n
y
m
o
f
t
er
m
is
f
o
u
n
d
in
t
h
e
p
r
ev
io
u
s
C
B
E
,
au
to
m
atica
ll
y
th
e
ter
m
h
a
s
a
s
a
m
e
lab
el
v
al
u
e.
A
u
to
m
atic
tag
g
i
n
g
p
r
o
ce
d
u
r
e
h
a
n
d
li
n
g
f
o
r
ca
s
es
th
a
t
d
o
n
o
t
h
a
v
e
lab
el
s
b
u
t
h
av
e
V
AD
v
al
u
es a
n
d
v
a
n
t i
n
s
tead
.
I
n
p
r
ev
io
u
s
C
B
E
[
2
0
]
,
P
OS
T
ag
g
i
n
g
an
d
w
o
r
d
f
ilter
i
n
g
in
t
h
e
E
m
o
tio
n
ca
teg
o
r
y
o
n
GI
h
a
s
n
o
t
b
ee
n
u
s
ed
.
I
t
ca
u
s
e
t
h
e
a
u
to
m
at
ic
ta
g
g
i
n
g
al
g
o
r
ith
m
h
a
s
n
o
t
p
r
o
d
u
ce
d
th
e
o
p
ti
m
al
o
u
tp
u
t.
T
h
e
d
eter
m
i
n
a
tio
n
o
f
th
e
e
m
o
tio
n
al
cl
u
s
ter
ce
n
ter
b
e
f
o
r
e
th
e
Au
to
m
atic
T
ag
g
i
n
g
al
g
o
r
ith
m
is
s
till
b
a
s
ed
o
n
th
e
r
esear
c
h
er
'
s
ass
u
m
p
tio
n
s
,
s
o
t
h
er
e
ar
e
s
till
n
o
t
m
a
x
i
m
al
r
es
u
lt
s
.
I
n
t
h
is
r
esear
ch
,
w
e
i
m
p
r
o
v
e
m
e
n
t
C
B
E
w
it
h
u
s
ed
P
OS
T
ag
g
in
g
,
GI
f
il
ter
in
g
an
d
cl
u
s
t
er
ce
n
ter
d
eter
m
in
at
io
n
.
A
s
well
as th
e
co
n
ce
p
t o
f
K
-
Nea
r
es
t n
ei
g
h
b
o
r
s
(
KNN)
,
th
is
r
e
s
ea
r
ch
u
s
e
s
t
h
e
clo
s
e
n
o
d
es
to
th
e
m
o
d
el.
T
h
e
d
if
f
er
en
ce
is
K
-
n
ea
r
est
is
u
s
ed
f
o
r
class
i
f
icatio
n
[
2
3
]
,
w
h
ile
th
i
s
r
esear
c
h
is
u
s
ed
t
o
lo
o
k
f
o
r
t
h
e
s
co
r
e
o
f
Vala
n
ce
-
A
r
o
u
s
al
-
Do
m
i
n
an
ce
(
V
A
D)
.
C
lu
s
ter
ce
n
ter
d
eter
m
in
at
io
n
is
b
a
s
ed
o
n
V
A
D
a
v
er
ag
e
v
al
u
e
i
n
ev
er
y
e
m
o
tio
n
lab
el.
W
it
h
th
a
t
i
m
p
r
o
v
e
m
en
t,
C
B
E
is
ex
p
ec
ted
to
d
o
b
etter
an
d
p
r
o
d
u
ce
m
o
r
e
ac
cu
r
ate
o
u
tp
u
t r
es
u
l
t.
Step
o
f
au
to
ta
g
g
i
n
g
p
r
o
ce
d
u
r
es o
f
in
co
m
p
lete
d
ata
ar
e:
1)
Def
i
n
e
t
h
e
ce
n
ter
o
f
cl
u
s
ter
f
o
r
ea
ch
lab
el
o
f
e
m
o
tio
n
'
J
o
y
'
,
'Sad
'
,
'An
g
er
'
,
'
D
is
g
u
s
t
'
,
'
Fear
'
,
'
S
u
r
p
r
is
e
'
.
T
h
e
ce
n
ter
o
f
t
h
e
cl
u
s
ter
is
ta
k
e
n
f
r
o
m
th
e
clo
s
est
ter
m
to
t
h
e
a
v
er
ag
e
ter
m
d
ata
o
f
ea
ch
clu
s
te
r
.
T
h
e
ce
n
ter
o
f
clu
s
ter
ce
r
tai
n
l
y
h
as V
AD
v
al
u
e
an
d
e
m
o
tio
n
lab
el
.
N
o
t
a
v
a
i
l
a
b
l
e
T
e
r
m x
i
n
c
o
m
p
l
e
t
e
(
d
o
e
sn
‟
t
h
a
v
e
l
a
b
e
l
o
r
d
i
me
n
s
i
o
n
v
a
l
u
e
)
S
e
a
r
c
h
sy
n
o
n
y
m o
f
t
e
r
m x
i
n
w
o
r
d
n
e
t
C
h
e
c
k
e
mo
t
i
o
n
a
l
l
a
b
e
l
o
r
d
i
me
n
s
i
o
n
a
l
v
a
l
u
e
o
f
sy
n
o
n
y
m x
i
n
W
N
A
-
A
N
E
W
A
u
t
o
mat
i
c
t
a
g
g
i
n
g
p
r
o
c
e
d
u
r
e
U
se
d
a
t
r
i
b
u
t
o
f
sy
n
o
n
y
m x
f
o
r
t
e
r
m
x
a
v
a
i
l
a
b
l
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
3
,
J
u
n
e
2
0
1
8
:
1
7
2
0
–
1730
1724
Fig
u
r
e
5
.
I
llu
s
tr
atio
n
p
o
s
it
io
n
o
f
ter
m
s
ac
co
r
d
in
g
to
V
A
D
v
a
lu
e
2)
Fo
r
ter
m
a
t
h
at
h
a
s
n
o
e
m
o
tio
n
lab
el:
a.
I
f
„
ter
m
a‟
d
o
es
n
o
t
h
a
v
e
V
AD,
ch
ec
k
t
h
e
A
d
ap
ted
L
E
S
K
v
alu
e
b
et
w
ee
n
ter
m
x
w
i
th
ea
ch
ce
n
ter
o
f
clu
s
ter
.
A
d
ap
ted
L
E
SK
v
alu
e
is
th
e
v
alu
e
o
f
th
e
p
r
o
x
i
m
it
y
b
et
w
ee
n
th
e
t
w
o
ter
m
s
.
T
h
e
h
ig
h
e
s
t
v
alu
e
d
escr
ib
e
th
e
clo
s
e
s
t
ter
m
.
T
h
e
e
m
o
tio
n
lab
el
b
et
w
ee
n
t
w
o
te
r
m
s
i
s
co
n
s
id
e
r
ed
s
a
m
e.
T
h
e
l
ab
el
o
f
ter
m
a
eq
u
als
w
it
h
th
e
e
m
o
tio
n
lab
e
l o
n
its
ter
m
cl
u
s
ter
.
b.
I
f
„
ter
m
a
‟
h
as
V
A
D
v
a
lu
e,
eu
clid
ea
n
d
is
ta
n
ce
v
al
u
e
b
ec
o
m
e
s
r
e
f
er
en
ce
to
f
i
n
d
t
h
e
n
ea
r
est
ter
m
.
eu
clid
ea
n
d
is
ta
n
ce
v
al
u
e
b
ec
o
m
es
r
ef
er
e
n
ce
to
f
in
d
th
e
n
ea
r
est
ter
m
.
Seen
i
n
Fi
g
u
r
e
5
,
'
E
u
cDis
a
-
s
ad
'
is
th
e
E
u
clid
ea
n
d
is
ta
n
ce
b
et
w
ee
n
ter
m
a
w
it
h
t
h
e
ce
n
ter
o
f
t
h
e
s
ad
cl
u
s
ter
.
An
d
'
E
u
cDi
s
a
-
j
o
y
'
i
s
t
h
e
E
u
clid
ea
n
d
is
ta
n
ce
b
et
w
ee
n
te
r
m
a
w
it
h
th
e
ce
n
ter
o
f
th
e
j
o
y
cl
u
s
ter
.
F
o
r
m
u
la
(
1
)
is
a
f
o
r
m
u
la
to
f
in
d
th
e
v
al
u
e
o
f
eu
clid
ea
n
d
is
t
an
c
e
b
et
w
ee
n
„
ter
m
a‟
a
n
d
ter
m
ce
n
ter
o
f
clu
s
ter
(
„
ter
m
p
c‟
)
.
T
h
e
s
m
a
lles
t
eu
cled
ian
v
alu
e
r
ep
r
esen
t
s
th
e
p
r
o
x
im
i
t
y
b
et
w
ee
n
ter
m
a
a
n
d
th
e
ter
m
clu
s
ter
.
So
th
e
ter
m
ce
n
ter
o
f
clu
s
ter
lab
eled
is
co
n
s
id
er
ed
eq
u
al
to
ter
m
a.
(
1
)
W
e
u
s
ed
Mu
s
ic
I
n
f
o
r
m
at
io
n
R
etr
iev
al
E
v
al
u
atio
n
eXc
h
an
g
e
(
MI
R
E
X
)
d
ataset
f
o
r
Mu
s
ic
d
o
cu
m
e
n
t.
MI
R
E
X
[
2
4
]
is
a
m
u
s
ic
d
atas
et
f
o
r
Mo
o
d
C
lass
i
f
icatio
n
T
ask
i
n
I
n
ter
n
atio
n
al
So
ciet
y
f
o
r
Mu
s
ic
I
n
f
o
r
m
atio
n
R
etr
iev
al
(
I
SMI
R
)
co
n
f
er
e
n
ce
s
.
T
h
is
m
o
d
el
class
i
f
ie
s
e
m
o
ti
o
n
s
in
to
f
i
v
e
d
is
ti
n
ct
g
r
o
u
p
s
o
r
c
lu
s
ter
s
(
T
ab
le
1
)
,
ea
ch
cl
u
s
ter
co
m
p
r
is
i
n
g
f
i
v
e
to
s
ev
e
n
r
elate
d
e
m
o
t
io
n
s
(
a
d
j
ec
tiv
es).
T
h
er
e
ar
e
9
0
3
d
ata
in
3
0
-
s
ec
o
n
d
s
o
f
au
d
io
.
I
ts
d
iv
id
ed
in
to
5
m
o
o
d
clu
s
ter
s
.
E
ac
h
cl
u
s
ter
h
as
b
alan
ce
s
n
u
m
b
er
o
f
d
ata
(
1
7
0
,
1
6
4
,
2
1
5
,
1
9
1
,
1
6
3
ex
ce
r
p
ts
)
.
Of
9
0
3
au
d
io
d
ata,
7
6
4
h
as
au
d
io
a
n
d
l
y
r
ics.
B
u
t
b
ec
au
s
e
t
h
er
e
i
s
a
p
r
o
ce
s
s
o
f
c
o
n
v
er
ti
n
g
d
ata
i
n
to
T
h
ay
er
m
o
d
el,
t
h
en
t
h
e
test
d
a
ta
u
s
ed
is
4
5
6
d
ata.
T
ab
le
1
.
Fiv
e
clu
s
ter
i
n
MI
R
E
X
Data
s
et
C
l
u
st
e
r
s
M
o
o
d
a
d
j
e
c
t
i
v
e
s
C
l
u
st
e
r
1
P
a
ssi
o
n
a
t
e
,
r
o
u
s
i
n
g
,
c
o
n
f
i
d
e
n
t
,
b
o
i
st
e
r
o
u
s,
r
o
w
d
y
C
l
u
st
e
r
2
R
o
l
l
i
c
k
i
n
g
,
c
h
e
e
r
f
u
l
,
f
u
n
,
sw
e
e
t
,
a
mi
a
b
l
e
/
g
o
o
d
n
a
t
u
r
e
d
C
l
u
st
e
r
3
L
i
t
e
r
a
t
e
,
p
o
i
g
n
a
n
t
,
w
i
st
f
u
l
,
b
i
t
t
e
r
sw
e
e
t
,
a
u
n
t
u
mn
a
l
,
b
r
o
o
d
i
n
g
C
l
u
st
e
r
4
H
u
mo
r
o
u
s,
s
i
l
l
y
,
c
a
m
p
y
,
q
u
i
r
k
y
,
w
h
i
msic
a
l
,
w
i
t
t
y
,
w
r
y
C
l
u
st
e
r
5
A
g
g
r
e
ssi
v
e
,
f
i
e
r
y
,
t
e
n
se
/
a
n
x
x
i
o
u
s
,
i
n
t
e
n
se
,
v
o
l
a
t
i
l
e
,
v
i
sce
r
a
l
E
m
o
tio
n
al
lab
el
o
f
MI
R
E
X
d
er
iv
ed
f
r
o
m
R
u
s
s
el
Mo
d
el,
s
o
it
u
s
es
5
e
m
o
tio
n
cl
u
s
ter
la
b
els
.
C
B
E
h
av
e
6
e
m
o
t
io
n
al
lab
el
d
er
iv
e
d
f
r
o
m
s
i
x
E
k
m
a
n
b
asic
e
m
o
t
i
o
n
.
I
n
o
r
d
er
to
co
o
p
er
ate
to
s
u
p
p
o
r
t
th
is
r
esear
ch
,
it
tak
e
s
co
n
v
er
s
io
n
p
r
o
ce
s
s
to
g
et
t
h
e
u
n
i
f
o
r
m
d
ata
f
o
r
e
m
o
t
i
o
n
al
lab
el
.
T
h
is
r
e
s
ea
r
ch
u
s
e
s
T
h
ay
er
s
m
o
d
el
f
o
r
th
e
u
n
i
f
o
r
m
i
t
y
o
f
lab
el
d
ata
,
b
ec
au
s
e
it
h
as
4
clas
s
e
m
o
tio
n
s
w
it
h
clea
r
li
m
it
s
o
n
d
i
m
e
n
s
io
n
s
p
ac
es
[
2
5
]
.
Fro
m
th
e
d
ata
co
n
v
er
s
io
n
p
r
o
ce
s
s
g
e
n
er
ated
4
class
e
m
o
tio
n
s
,
n
a
m
el
y
: c
las
s
1
.
class
2
,
class
3
,
an
d
class
4
.
Fig
u
r
e
6
i
s
t
h
e
m
ap
p
in
g
o
f
MI
R
E
X
lab
el
to
T
h
a
y
er
m
o
d
el.
Of
th
e
i
m
ag
e
is
clea
r
l
y
v
is
ib
le
t
h
at
„
C
l
u
s
ter
5
‟
o
n
th
e
MI
R
E
X
w
il
l
b
e
co
n
v
er
ted
to
„
C
lass
2
‟
o
n
T
h
ay
er
,
„
C
l
u
s
ter
2
‟
w
ill
b
e
co
n
v
er
ted
to
„
C
lass
1
‟
o
n
T
h
a
y
er
,
a
n
d
„
C
l
u
s
ter
3
‟
w
il
l
b
e
co
n
v
er
ted
to
„
C
las
s
3
‟
T
h
a
y
er
.
I
n
d
i
m
e
n
s
io
n
s
p
ac
es,
„
C
l
u
s
ter
1
‟
a
n
d
„
C
l
u
s
ter
4
‟
o
f
MI
R
E
X
ar
e
l
o
ca
ted
in
t
h
e
s
lice
ar
ea
b
et
w
ee
n
2
e
m
o
tio
n
al
cla
s
s
e
s
o
f
T
h
a
y
er
.
„
C
l
u
s
ter
1
‟
MI
R
E
X
is
lo
ca
ted
in
th
e
ar
ea
o
f
„
C
las
s
1
‟
an
d
„
C
las
s
2
‟
.
W
h
er
ea
s
„
C
l
u
s
ter
4
‟
MI
R
E
X
is
lo
ca
ted
in
th
e
ar
ea
o
f
„
C
la
s
s
1
‟
an
d
„
C
las
s
4
‟
.
I
n
t
h
i
s
r
esear
ch
th
er
e
i
s
n
o
h
a
n
d
li
n
g
o
f
it,
s
o
o
n
l
y
m
u
s
ic
d
ata
o
n
MI
R
E
X
w
it
h
„
C
lu
s
ter
2
‟
,
„
C
l
u
s
ter
3
‟
,
a
n
d
„
C
l
u
s
ter
5
‟
t
h
at
u
s
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
Mu
s
ic
E
mo
tio
n
C
la
s
s
ifica
tio
n
b
a
s
ed
o
n
Lyr
ics
-
A
u
d
io
u
s
in
g
C
B
E
(
F
ika
Ha
s
ta
r
ita
R
a
ch
ma
n
)
1725
Fig
u
r
e
6
.
C
o
n
v
er
s
io
n
m
ap
p
in
g
lab
el
o
f
e
m
o
tio
n
MI
R
E
X
to
T
h
a
y
er
m
o
d
el
T
h
e
co
n
v
er
s
io
n
m
ap
p
in
g
o
f
th
e
e
m
o
tio
n
al
lab
el
f
r
o
m
C
B
E
lab
els
to
T
h
a
y
e
r
lab
el
s
i
s
s
h
o
w
n
i
n
Fig
u
r
e
7
.
I
n
th
e
Fig
u
r
e
,
it a
p
p
ea
r
s
th
at
t
h
e
'
Hap
p
y
'
a
n
d
'
Su
r
p
r
is
e
'
lab
els ar
e
co
n
v
er
ted
to
th
e
'
C
las
s
1
'
i
n
T
h
a
y
er
class
.
'
Di
s
g
u
s
t
'
,
„
Fear
‟
,
a
n
d
„
A
n
g
er
‟
lab
els
co
n
v
er
ted
o
n
„
C
l
ass
2
'
T
h
a
y
er
.
An
d
„
Sad
n
es
s
‟
lab
el
ar
e
co
n
v
er
ted
ti „
C
las
s
3
‟
in
T
h
a
y
er
.
Fig
u
r
e
7
.
C
o
n
v
er
s
io
n
m
ap
p
in
g
lab
el
em
o
tio
n
o
f
C
B
E
to
T
h
ay
er
m
o
d
el
T
ha
y
e
r
M
o
de
l
La
be
l
e
m
o
t
i
o
n
o
f
CB
E
L
a
b
e
l
e
mo
t
i
o
n
o
f
M
I
R
EX
T
h
a
y
e
r
M
o
d
e
l
Emo
t
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
3
,
J
u
n
e
2
0
1
8
:
1
7
2
0
–
1730
1726
Fig
u
r
e
9
is
th
e
d
ata
f
lo
w
p
r
o
ce
s
s
f
r
o
m
t
h
e
co
n
v
er
t
e
m
o
t
io
n
al
lab
el
p
r
o
ce
s
s
,
p
r
ep
r
o
ce
s
s
in
g
u
n
til
l
y
r
ic
s
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
B
ef
o
r
e
th
e
ly
r
ic
s
f
ea
t
u
r
e
ex
tr
ac
tio
n
p
r
o
ce
s
s
,
th
er
e
is
a
p
r
ep
r
o
ce
s
s
in
g
d
ata.
I
n
th
e
ly
r
ic
s
th
er
e
ar
e
m
a
n
y
in
f
o
r
m
al
w
o
r
d
s
,
s
u
c
h
as
n
‟
w
h
ich
m
ea
n
s
'
an
d
'
,
'
l
l
w
h
ic
h
m
ea
n
s
'w
i
ll
'
,
'
e
m
w
h
ic
h
m
ea
n
s
'
t
h
e
m
',
an
d
o
th
er
s
.
I
t
is
n
ec
ess
ar
y
f
o
r
r
ep
air
d
ata
to
im
p
r
o
v
e
th
e
wo
r
d
in
to
a
f
o
r
m
al
w
o
r
d
s
tr
u
ct
u
r
e
f
o
r
m
.
B
ef
o
r
e
th
e
r
ep
air
d
ata
p
r
o
ce
s
s
,
th
er
e
is
a
ch
ec
k
in
g
p
o
s
it
io
n
o
f
w
o
r
d
s
u
s
i
n
g
P
OS
T
ag
g
i
n
g
Sta
n
d
f
o
r
d
.
T
h
is
is
i
m
p
o
r
tan
t
b
ec
au
s
e
th
er
e
ar
e
p
ar
ts
o
f
s
e
n
t
en
ce
s
t
h
at
w
il
l
n
o
t
b
e
p
r
o
ce
s
s
ed
(
p
r
e
p
o
s
itio
n
,
ar
ticle,
p
o
s
s
esiv
e
p
r
o
n
o
u
n
,
etc.
)
.
P
ar
ts
o
f
w
o
r
d
s
th
at
ar
e
n
o
t i
n
cl
u
d
ed
in
th
e
r
ep
air
p
r
o
ce
s
s
w
i
ll
b
e
f
ilter
ed
f
r
o
m
th
e
d
ata.
I
n
ac
co
r
d
an
ce
w
it
h
[
2
6
]
,
an
au
d
io
f
ea
t
u
r
e
th
at
a
f
f
ec
ts
t
h
e
m
u
s
ic
e
m
o
tio
n
r
ec
o
g
n
it
io
n
a
s
s
h
o
w
n
i
n
T
ab
le
2
.
I
n
th
is
s
t
u
d
y
,
a
u
d
io
f
ea
tu
r
es
w
er
e
ex
tr
ac
ted
u
s
i
n
g
t
h
e
P
s
y
s
o
u
n
d
3
an
d
MI
R
T
o
o
lb
o
x
.
Featu
r
es
i
n
cl
u
d
e
f
ea
t
u
r
e
o
f
en
er
g
y
(
d
y
n
a
m
ic
lo
u
d
n
es
s
)
,
f
ea
t
u
r
e
o
f
te
m
p
o
r
al
(
te
m
p
o
)
an
d
f
ea
tu
r
e
s
o
f
s
p
ec
tr
u
m
(
r
o
u
g
h
n
es
s
an
d
in
h
ar
m
o
n
icit
y
)
.
Fig
u
r
e
8
.
Flo
w
p
r
o
ce
s
s
f
o
r
ex
tr
ac
tio
n
f
ea
t
u
r
e
r
o
u
g
h
n
es
s
T
h
e
p
r
o
ce
s
s
o
f
f
ea
t
u
r
e
ex
tr
ac
t
io
n
is
s
h
o
w
n
i
n
Fi
g
u
r
e
9
.
T
h
at
p
r
o
ce
s
s
u
s
ed
m
ir
to
o
lb
o
x
f
u
n
ctio
n
.
Fo
r
r
o
u
g
h
n
e
s
s
f
ea
t
u
r
e,
t
h
e
f
u
n
ctio
n
s
m
ir
a
u
d
io
(
)
an
d
m
ir
r
o
u
g
h
n
es
s
(
)
ar
e
u
s
ed
.
T
h
e
ex
tr
ac
ti
o
n
f
lo
w
f
ea
tu
r
e
f
o
r
o
th
er
f
ea
t
u
r
es u
s
es a
s
a
m
e
p
at
h
lik
e
Fig
u
r
e
9
w
ith
t
h
e
u
s
e
o
f
d
if
f
er
e
n
t f
u
n
ctio
n
s
o
n
m
ir
to
o
lb
o
x
o
r
p
s
y
s
o
u
n
d
3.
Fig
u
r
e
9
.
P
r
o
c
es
s
o
f
ly
r
ic
s
I
n
th
e
cla
s
s
i
f
ic
atio
n
p
r
o
ce
s
s
,
l
y
r
ics
a
n
d
au
d
io
f
ea
t
u
r
es
ar
e
te
s
ted
u
s
i
n
g
t
h
r
ee
class
i
f
icat
io
n
m
et
h
o
d
s
:
th
e
S
u
p
p
o
r
t V
ec
to
r
Ma
ch
in
e
(
SVM
)
m
et
h
o
d
,
th
e
R
a
n
d
o
m
F
o
r
est
m
eth
o
d
an
d
t
h
e
Naiv
e
B
a
y
e
s
m
et
h
o
d
.
Ly
r
i
c
s
C
o
nv
e
r
t
e
m
o
t
i
o
na
l
l
ab
e
l
i
nt
o
T
ha
y
e
r
M
o
de
l
P
r
e
pr
o
c
e
ss
i
ng
D
at
a
R
e
pa
i
r
d
at
a:
n’ ; ‘l
l
;
n’t
;
t
’
;
‘
e
m
;
‘
r
e
;
‘
v
e
;
‘
n
; ‘s ;
‘
c
au
se
;
…
F
i
l
t
e
r
i
ng
wo
r
d
us
i
ng
P
O
S
T
ag
g
i
ng
St
an
df
o
r
d
E
xt
r
ac
t
i
ng
f
e
at
ur
e
o
f
l
y
r
i
c
s:
ps
y
c
ho
l
i
ng
ui
st
i
c
a
nd
st
y
l
i
s
t
i
c
f
e
at
ur
e
s
f
i
t
ur
M
us
i
c
.w
av
M
IR
T
o
o
l
bo
x
F
un
c
t
i
o
n:
m
i
r
au
di
o
(
)
F
un
c
t
i
o
n:
M
i
r
r
o
ug
hn
e
ss
(
)
si
g
na
l
R
o
ug
hn
e
ss
r
e
su
l
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
Mu
s
ic
E
mo
tio
n
C
la
s
s
ifica
tio
n
b
a
s
ed
o
n
Lyr
ics
-
A
u
d
io
u
s
in
g
C
B
E
(
F
ika
Ha
s
ta
r
ita
R
a
ch
ma
n
)
1727
T
ab
le
2
.
A
u
d
io
Featu
r
e
f
o
r
Mu
s
ic
E
m
o
tio
n
R
ec
o
g
n
itio
n
[
2
6
]
F
e
a
t
u
r
e
se
t
Ex
t
r
a
c
t
o
r
F
e
a
t
u
r
e
s
En
e
r
g
y
P
sy
s
o
u
n
d
D
y
n
a
mi
c
l
o
a
d
n
e
ss
S
D
T
A
u
d
i
o
p
o
w
e
r
,
t
o
t
a
l
l
o
u
d
n
e
ss,
a
n
d
s
p
e
c
i
f
i
c
l
o
u
d
n
e
ss se
n
sa
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
s
R
h
y
t
h
m
M
a
r
sy
a
s
B
e
a
t
h
i
s
t
o
g
r
a
m
M
A
t
o
o
l
b
o
x
,
R
P
e
x
t
r
a
c
t
o
r
R
h
y
t
h
m
p
a
t
t
e
r
n
,
r
h
y
t
h
m
h
i
st
o
g
r
a
m,
a
n
d
t
e
m
p
o
M
I
R
t
o
o
l
b
o
x
R
h
y
t
h
m
st
r
e
n
g
t
h
,
r
h
y
t
h
m re
g
u
l
a
r
i
t
y
,
r
h
y
t
h
m c
l
a
r
i
t
y
,
a
v
e
r
a
g
e
o
n
se
t
f
r
e
q
u
e
n
c
y
,
a
n
d
a
v
e
r
a
g
e
t
e
mp
o
T
e
mp
o
r
a
l
S
D
T
Z
e
r
o
-
c
r
o
ssi
n
g
s,
t
e
mp
o
r
a
l
c
e
n
t
r
o
i
d
,
a
n
d
l
o
g
a
t
t
a
c
k
t
i
me
S
p
e
c
t
r
u
m
M
a
r
sy
a
s,
S
D
T
S
p
e
c
t
r
a
l
c
e
n
t
r
o
i
d
,
s
p
e
c
t
r
a
l
r
o
l
l
o
f
f
,
sp
e
c
t
r
a
l
f
l
u
x
,
s
p
e
c
t
r
a
l
f
l
a
t
n
e
ss
me
a
su
r
e
s,
a
n
d
sp
e
c
t
r
a
l
c
r
e
st
f
a
c
t
o
r
s
M
A
t
o
o
l
b
o
x
,
M
a
r
sy
a
s,
S
D
T
M
e
l
-
f
r
e
q
u
e
n
c
y
c
e
p
st
r
a
l
c
o
e
f
f
i
c
i
e
n
t
s
M
A
TL
A
B
S
p
e
c
t
r
a
l
c
o
n
t
r
a
s
t
,
D
a
u
b
e
c
h
i
e
s w
a
v
e
l
e
t
s c
o
e
f
f
i
c
i
e
n
t
h
i
st
o
g
r
a
m,
t
r
i
st
i
m
u
l
u
s,
e
v
e
n
-
h
a
r
m
,
a
n
d
o
d
d
-
h
a
r
m
.
M
I
R
t
o
o
l
b
o
x
R
o
u
g
h
n
e
ss,
i
r
r
e
g
u
l
a
r
i
t
y
,
a
n
d
i
n
h
a
r
mo
n
i
c
i
t
y
H
a
r
mo
n
y
M
I
R
t
o
o
l
b
o
x
S
a
l
i
e
n
t
p
i
t
c
h
,
c
h
r
o
mag
r
a
m c
e
n
t
r
o
i
d
,
k
e
y
c
l
a
r
i
t
y
,
mu
si
c
a
l
mo
d
e
,
a
n
d
h
a
r
mo
n
i
c
c
h
a
n
g
e
M
a
r
sy
a
s
P
i
t
c
h
h
i
s
t
o
g
r
a
m
P
sy
S
o
u
n
d
S
a
w
t
o
o
t
h
w
a
v
e
f
o
r
m i
n
sp
i
r
e
d
p
i
t
c
h
e
st
i
ma
t
e
2.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
T
h
e
d
ataset
u
s
ed
as
te
s
t
d
ata
i
s
th
e
MI
R
E
X
-
li
k
e
m
o
o
d
d
ataset
[
2
4
]
.
I
n
MI
R
E
X,
7
6
4
d
ata
h
as
l
y
r
ic
s
an
d
au
d
io
.
B
u
t
b
ec
au
s
e
o
f
co
n
v
er
s
io
n
lab
el
e
m
o
tio
n
p
r
o
ce
s
s
to
T
h
a
y
er
m
o
d
el,
w
e
u
s
ed
4
5
6
d
ata.
T
h
ats
d
ata
h
as
a
'
cl
u
s
ter
2
'
,
'
cl
u
s
ter
3
'
,
an
d
'
cl
u
s
ter
5
'
e
m
o
tio
n
lab
el
s
.
T
h
e
d
ata
w
ill
b
e
u
s
ed
in
th
e
m
u
s
ic
e
m
o
tio
n
class
i
f
icatio
n
.
T
h
er
e
ar
e
2
t
esti
n
g
m
o
d
els
:
C
B
E
ac
cu
r
ac
y
te
s
ti
n
g
f
o
r
e
m
o
tio
n
al
clas
s
if
ica
tio
n
b
ased
o
n
p
s
y
ch
o
li
n
g
u
is
tic
f
ea
t
u
r
es,
an
d
e
m
o
tio
n
cla
s
s
i
f
icatio
n
t
esti
n
g
o
f
m
u
s
ic
w
i
th
v
ar
io
u
s
f
ea
t
u
r
es
.
T
h
e
f
ir
s
t
test
w
as
co
n
d
u
cted
w
it
h
th
e
ai
m
o
f
a
n
a
l
y
zi
n
g
t
h
e
b
est
C
B
E
ca
s
e
to
b
e
u
s
ed
f
o
r
p
s
y
c
h
o
li
n
g
u
i
s
tic
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
T
h
e
s
ec
o
n
d
test
is
d
o
n
e
w
i
th
t
h
e
ai
m
o
f
f
i
n
d
in
g
th
e
b
es
t
f
ea
t
u
r
e
t
h
at
w
ill
b
e
u
s
ed
f
o
r
th
e
c
la
s
s
i
f
icatio
n
o
f
e
m
o
tio
n
m
u
s
ic.
I
n
t
h
e
f
ir
s
t
te
s
t,
t
h
er
e
ar
e
3
c
ases
o
f
C
B
E
,
n
a
m
el
y
:
C
B
E
1
,
C
B
E
2
,
an
d
C
B
E
3
.
C
B
E
1
is
a
m
er
g
in
g
A
NE
W
an
d
W
N
A
d
ataset
w
i
t
h
n
o
ex
p
a
n
s
io
n
p
r
o
ce
s
s
o
r
au
to
m
a
tic
tag
g
i
n
g
p
r
o
ce
d
u
r
e.
C
B
E
2
is
C
B
E
1
w
h
ich
h
as
u
n
d
er
g
o
n
e
a
u
to
m
atic
tag
g
in
g
p
r
o
ce
s
s
u
s
in
g
W
o
r
d
n
et
s
y
n
o
n
y
m
co
n
ce
p
t.
A
n
d
C
B
E
3
is
th
e
d
ev
elo
p
m
e
n
t
o
f
C
B
E
2
w
h
ich
h
as
u
n
d
er
g
o
n
e
a
u
to
m
at
ic
tag
g
i
n
g
p
r
o
ce
s
s
u
s
i
n
g
E
u
clid
ea
n
Dis
tan
ce
co
n
ce
p
t.
Fo
r
th
ese
test
s
,
th
e
th
r
ee
C
B
E
s
ar
e
u
s
ed
in
ter
ch
a
n
g
ea
b
l
y
f
o
r
th
e
ex
tr
ac
tio
n
o
f
p
s
y
ch
o
li
n
g
u
i
s
tic
f
ea
tu
r
e
s
in
th
e
l
y
r
ics.
T
h
e
r
esu
lt
o
f
its
f
ea
t
u
r
e
ex
tr
ac
tio
n
is
u
s
ed
f
o
r
th
e
class
i
f
icatio
n
o
f
e
m
o
tio
n
m
u
s
ic.
Fig
u
r
e
1
0
s
h
o
w
s
th
e
d
ep
lo
y
m
en
t
o
f
C
B
E
1
d
ata
in
d
i
m
e
n
s
io
n
s
p
ac
e
,
w
h
er
e
C
1
is
Vale
n
ce
,
C
2
is
A
r
o
u
s
al
an
d
C
3
is
Do
m
i
n
an
c
e
.
Fro
m
Fi
g
u
r
e
1
0
,
it
ap
p
ea
r
s
th
at
t
h
e
d
ata
is
s
ca
tter
ed
w
el
l
o
n
th
e
d
i
m
e
n
s
io
n
a
l
s
p
ac
e.
I
t
m
a
k
i
n
g
ea
s
ier
f
o
r
t
h
e
co
n
v
er
s
io
n
p
r
o
ce
s
s
i
n
to
t
h
e
t
h
a
y
er
m
o
d
el.
T
h
e
o
b
s
tacle
is
t
h
e
d
i
f
f
er
en
ce
ex
is
te
n
ce
o
f
d
i
m
en
s
io
n
al.
C
B
E
h
av
e
3
d
i
m
e
n
s
io
n
w
h
ile
T
h
a
y
er
m
o
d
el
o
n
l
y
h
a
v
e
2
d
i
m
e
n
s
io
n
.
Fo
r
th
e
ti
m
e
be
in
g
,
we
ad
j
u
s
ted
th
e
d
ata
w
i
th
T
h
a
y
er
m
o
d
el
u
s
i
n
g
2
d
i
m
e
n
s
io
n
(
Vale
n
ce
-
A
r
o
u
s
al)
.
C
B
E
2
is
f
o
r
m
ed
w
i
th
t
h
e
h
el
p
o
f
S
y
n
s
et
o
f
W
o
r
d
n
et.
Fi
g
u
r
e
1
1
s
h
o
w
s
t
h
e
ce
n
tr
al
p
o
s
iti
o
n
o
f
th
e
clu
s
ter
in
d
i
m
e
n
s
io
n
s
p
ac
e.
T
h
e
ce
n
ter
o
f
c
lu
s
ter
is
th
e
in
it
ial
s
tep
r
esu
lt
o
f
au
to
ta
g
g
i
n
g
p
r
o
ce
d
u
r
e
o
f
in
co
m
p
lete
d
ata.
T
h
is
ce
n
ter
o
f
clu
s
ter
w
ill b
e
th
e
ce
n
ter
ter
m
to
f
o
r
m
at
io
n
o
f
C
B
E
3
.
Fig
u
r
e
1
0
.
Sp
r
ea
d
o
f
C
B
E
1
d
a
ta
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
3
,
J
u
n
e
2
0
1
8
:
1
7
2
0
–
1730
1728
Fig
u
r
e
11
.
C
lu
s
ter
ce
n
ter
p
o
s
itio
n
in
d
i
m
en
s
io
n
s
p
ac
e
T
h
e
f
o
r
m
u
la
o
f
ac
cu
r
ac
y
test
i
n
g
d
ata
ter
lih
a
t
p
ad
a
f
o
r
m
u
la
(
1
)
.
T
h
e
f
o
r
m
u
la
is
r
atio
b
et
w
e
en
th
e
s
u
m
o
f
th
e
tr
u
e
p
r
ed
icted
w
it
h
to
tal
d
o
cu
m
e
n
t.
An
d
Ta
b
le
3
s
h
o
ws
t
h
e
ac
cu
r
ac
y
v
al
u
es
o
b
tain
ed
f
o
r
d
if
f
er
en
t
C
B
E
u
s
e
s
i
n
t
h
e
e
x
tr
ac
tio
n
p
r
o
ce
s
s
o
f
p
s
y
co
li
n
g
u
i
s
tic
f
ea
t
u
r
es.
I
t
is
s
ee
n
th
a
t
t
h
e
u
s
e
o
f
C
B
E
3
h
as
a
b
etter
p
er
ce
n
tag
e
ac
c
u
r
ac
y
v
al
u
e
o
f
0
.
3
7
.
So
th
e
p
s
y
co
li
n
g
u
i
s
tic
f
ea
t
u
r
e
u
s
ed
f
o
r
th
e
cla
s
s
i
f
i
ca
tio
n
p
r
o
ce
s
s
i
s
a
p
s
y
co
lin
g
u
is
t
ic
f
ea
t
u
r
e
u
s
i
n
g
C
B
E
3
.
(
2
)
T
ab
le
3
.
T
h
e
A
cc
u
r
ac
y
V
alu
e
s
f
o
r
d
if
f
er
e
n
t
C
a
s
e
o
f
C
B
E
C
a
se
T
r
u
e
p
r
e
d
i
c
t
i
o
n
A
c
c
u
r
a
c
y
C
B
E
1
1
1
8
0
,
2
5
8
C
B
E
2
1
3
7
0
,
3
0
0
4
C
B
E
3
1
6
9
0
,
3
7
T
h
e
class
i
f
icatio
n
p
r
o
ce
s
s
w
a
s
test
ed
u
s
in
g
S
VM
,
R
a
n
d
o
m
Fo
r
est
an
d
Nai
v
e
B
a
y
es
m
o
d
el
s
.
T
h
e
class
i
f
icatio
n
p
r
o
ce
s
s
u
s
i
n
g
th
e
h
elp
o
f
W
ek
a
to
o
ls
w
i
th
p
er
ce
n
tag
e
s
p
lit
6
6
%.
T
h
e
F
-
Me
a
s
u
r
e
v
al
u
e
o
f
ea
c
h
m
et
h
o
d
is
s
h
o
w
n
i
n
t
h
e
T
ab
le
4
.
T
h
er
e
ar
e
4
test
ca
s
es,
ea
ch
ca
s
e
u
s
i
n
g
d
if
f
er
en
t
f
ea
t
u
r
es.
T
h
e
f
ir
s
t
ca
s
e
u
s
in
g
th
e
au
d
io
f
ea
t
u
r
e
o
n
l
y
.
T
h
e
au
d
io
f
ea
tu
r
e
ar
e
d
y
n
a
m
ic
lo
u
d
n
es
s
,
te
m
p
o
,
r
o
u
g
h
n
e
s
s
a
n
d
in
h
ar
m
o
n
icit
y
.
I
t
ap
p
ea
r
s
th
at
th
e
b
e
s
t
r
es
u
lt
s
o
b
tain
ed
b
y
u
s
in
g
Nai
v
e
B
a
y
es
w
it
h
v
al
u
e
o
f
0
,
4
6
0
.
T
h
e
s
ec
o
n
d
ca
s
e
u
s
e
s
t
y
li
s
tic
f
ea
t
u
r
e.
T
h
e
f
ea
tu
r
e
is
o
n
l
y
ca
p
ab
le
o
f
b
r
in
g
in
g
th
e
r
es
u
lt
o
f
0
,
4
5
6
w
it
h
Nai
v
e
B
ay
e
s
m
et
h
o
d
.
Un
iq
u
e
to
th
e
th
ir
d
ca
s
e,
th
e
p
s
y
c
h
o
lin
g
u
is
ti
c
f
ea
tu
r
es
ar
e
n
o
t
af
f
ec
ted
b
y
t
h
r
ee
class
i
f
icatio
n
m
et
h
o
d
s
.
T
h
e
ac
cu
r
ac
y
r
e
s
u
lt
s
ar
e
eq
u
al
f
o
r
all.
T
h
e
v
alu
e
o
f
ac
cu
r
ac
y
is
0
,
3
5
4
.
T
ab
le
4
. F
-
Me
asu
r
e
Val
u
e
o
f
C
las
s
i
f
icatio
n
M
et
h
o
d
e
F
e
a
t
u
r
e
S
M
O
R
a
n
d
o
m F
o
r
e
st
N
a
i
v
e
B
a
y
e
s
A
u
d
i
o
0
,
2
8
1
0
,
4
2
8
0
,
4
6
0
S
t
y
l
i
st
i
c
0
,
3
5
8
0
,
4
3
3
0
,
4
5
6
P
sy
c
h
o
l
i
n
g
u
i
st
i
c
0
,
3
5
4
0
,
3
5
4
0
,
3
5
4
A
u
d
i
o
,
S
t
y
l
i
st
i
c
,
p
sy
c
h
o
l
i
n
g
u
i
st
i
c
0
,
4
3
7
0
,
5
6
8
0
,
4
5
6
T
h
e
last
ca
s
e
is
u
s
ed
all
o
f
f
ea
tu
r
es
in
ca
s
e
o
n
e,
ca
s
e
t
w
o
an
d
ca
s
e
tr
ee
.
I
t
is
s
ee
n
th
at
th
e
R
an
d
o
m
Fo
r
est
m
et
h
o
d
w
it
h
t
h
e
u
s
e
o
f
A
u
d
io
,
St
y
lis
tic
an
d
P
s
y
co
li
n
g
u
i
s
tic
f
ea
tu
r
es
h
as
t
h
e
b
est
F
-
m
ea
s
u
r
e
v
al
u
e.
T
h
er
e
is
0
.
5
6
8
.
3.
CO
NCLU
SI
O
N
T
h
is
r
esear
ch
s
h
o
w
t
h
at
t
h
e
u
s
e
o
f
C
B
E
is
ab
le
to
s
u
p
p
o
r
t
th
e
p
r
o
ce
s
s
o
f
clas
s
if
icatio
n
e
m
o
tio
n
o
f
m
u
s
ic.
W
it
h
t
h
e
b
es
t F
-
m
ea
s
u
r
e
f
o
r
R
a
n
d
o
m
Fo
r
est
m
et
h
o
d
o
f
5
6
.
8
%.
Fo
r
f
u
r
t
h
er
r
esear
ch
,
ad
d
itio
n
al
p
r
o
ce
s
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
Mu
s
ic
E
mo
tio
n
C
la
s
s
ifica
tio
n
b
a
s
ed
o
n
Lyr
ics
-
A
u
d
io
u
s
in
g
C
B
E
(
F
ika
Ha
s
ta
r
ita
R
a
ch
ma
n
)
1729
w
il
l
b
e
d
ev
elo
p
ed
to
i
m
p
r
o
v
e
th
e
ex
tr
ac
tio
n
p
er
f
o
r
m
a
n
ce
o
f
l
y
r
ics
a
n
d
au
d
io
f
ea
t
u
r
es.
So
t
h
at
ca
n
b
e
o
b
tain
e
d
b
etter
ac
cu
r
ac
y
v
al
u
e.
On
ce
an
al
y
ze
d
,
th
e
li
k
eli
h
o
o
d
o
f
er
r
o
r
s
o
cc
u
r
r
in
g
in
l
y
r
ical
f
ea
t
u
r
e
ex
tr
ac
tio
n
is
t
h
e
ab
s
en
ce
o
f
t
h
e
co
n
ce
p
t
o
f
W
o
r
d
Sen
s
e
Dis
a
m
b
i
g
u
a
tio
n
(
W
S
D)
[
2
7
]
,
A
d
v
er
b
-
A
d
j
ec
tiv
e
C
o
m
p
o
n
en
t
(
AAC)
o
r
Neg
atio
n
w
o
r
d
.
An
d
f
o
r
th
e
a
u
d
io
f
ea
t
u
r
e
n
ee
d
s
to
d
o
m
o
r
e
co
m
b
in
at
io
n
o
f
f
ea
t
u
r
e
ex
tr
ac
tio
n
,
s
o
i
t
ca
n
b
e
d
o
n
e
test
in
g
t
h
e
b
est au
d
io
f
ea
tu
r
e
f
o
r
th
e
e
m
o
tio
n
al
cla
s
s
i
f
ic
atio
n
o
f
t
h
is
m
u
s
ic.
RE
F
E
R
E
NC
E
S
[1
]
C.
S
trap
p
a
ra
v
a
a
n
d
A
.
V
a
li
tu
tt
i
,
“
W
o
rd
Ne
t
-
A
ffe
c
t :
a
n
Affe
c
ti
v
e
Ex
ten
sio
n
o
f
W
o
rd
Ne
t”
,
L
RE
C
,
p
p
.
1
0
8
3
-
1
0
8
6
,
2
0
0
4
.
[2
]
P
.
Ek
m
a
n
,
“
A
n
-
A
r
g
u
m
e
n
t
-
F
o
r
-
Ba
sic
-
Em
o
ti
o
n
s.p
d
f
”
,
Co
g
n
.
Emo
t
.
,
v
o
l.
6
,
n
o
.
3
,
p
p
.
1
6
9
-
2
0
0
,
1
9
9
2
.
[3
]
H.
Co
ro
n
a
a
n
d
M
.
P
.
O.
M
a
h
o
n
y
,
“
A
n
Ex
p
lo
ra
ti
o
n
o
f
M
o
o
d
Clas
sif
ica
ti
o
n
in
t
h
e
M
il
li
o
n
S
o
n
g
s Da
tas
e
t
”
,
1
2
th
S
o
u
n
d
a
n
d
M
u
sic
Co
mp
u
ti
n
g
Co
n
fer
e
n
c
e
,
2
0
1
5
.
[4
]
B.
Ro
c
h
a
,
R.
P
a
n
d
a
,
a
n
d
R.
P
.
P
a
iv
a
,
“
M
u
sic
Em
o
ti
o
n
Re
c
o
g
n
i
ti
o
n
:
T
h
e
Im
p
o
rtan
c
e
o
f
M
e
lo
d
i
c
F
e
a
tu
re
s
”
,
in
In
ter
n
a
t
io
n
a
l
W
o
rk
sh
o
p
o
n
M
a
c
h
i
n
e
L
e
a
rn
i
n
g
a
n
d
M
u
sic
(
M
M
L
)
,
n
o
.
2
0
0
8
,
2
0
1
3
.
[5
]
Y.
Hu
,
X
.
Ch
e
n
,
a
n
d
D.
Ya
n
g
,
“
L
y
ri
c
-
B
a
se
d
S
o
n
g
Em
o
ti
o
n
De
tec
ti
o
n
w
it
h
A
ff
e
c
ti
v
e
Lex
ico
n
a
n
d
F
u
z
z
y
Clu
ste
rin
g
M
e
th
o
d
”
,
IS
M
IR
2
0
0
9
,
p
p
.
1
2
3
-
1
2
8
,
2
0
0
9
.
[6
]
J.
S
.
Do
w
n
ie,
“
W
h
e
n
Ly
rics
Ou
terf
o
r
m
A
u
d
io
f
o
r
M
u
sic
M
o
o
d
Cl
a
s
sif
i
c
a
ti
o
n
:
A
F
e
a
tu
re
A
n
a
l
y
sis
”
,
IS
M
IR
2
0
1
0
,
p
p
.
6
1
9
-
6
2
4
,
2
0
1
0
.
[7
]
M
.
Kim
a
n
d
H.
Kw
o
n
,
“
Ly
ric
s
-
b
a
se
d
Em
o
ti
o
n
Clas
sif
ica
ti
o
n
u
sin
g
F
e
a
tu
re
S
e
lec
ti
o
n
b
y
P
a
rti
a
l
S
y
n
tac
ti
c
A
n
a
l
y
si
s
”
,
2
0
1
1
.
[8
]
J.
A
.
Rid
o
e
a
n
,
R.
S
a
rn
o
,
D.
S
u
n
a
ry
o
,
a
n
d
D.R.
W
ij
a
y
a
,
“
M
u
sic
m
o
o
d
c
las
sif
ica
ti
o
n
u
sin
g
a
u
d
i
o
p
o
w
e
r
a
n
d
a
u
d
io
h
a
rm
o
n
icity
b
a
se
d
o
n
M
P
EG
-
7
a
u
d
i
o
f
e
a
tu
re
s
a
n
d
S
u
p
p
o
r
t
V
e
c
to
r
M
a
c
h
in
e
”
,
2
0
1
7
3
r
d
In
t.
C
o
n
f.
S
c
i.
In
f.
T
e
c
h
n
o
l.
,
p
p
.
7
2
-
7
6
,
2
0
1
7
.
[9
]
L
.
L
u
,
D.
L
iu
,
a
n
d
H.
Zh
a
n
g
,
“
A
u
to
m
a
ti
c
M
o
o
d
De
tec
ti
o
n
a
n
d
T
ra
c
k
in
g
o
f
M
u
sic
A
u
d
io
S
ig
n
a
ls
”
,
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Au
d
i
o
,
S
p
e
e
c
h
,
a
n
d
L
a
n
g
u
a
g
e
Pro
c
e
ss
in
g
,
v
o
l.
1
4
,
n
o
.
1
,
p
p
.
5
-
1
8
,
Ja
n
2
0
0
6
.
[1
0
]
A
.
S
c
h
in
d
ler
a
n
d
A
.
Ra
u
b
e
r,
“
C
a
p
tu
ri
n
g
th
e
Te
m
p
o
ra
l
Do
m
a
in
i
n
Ech
o
n
e
st
F
e
a
tu
re
s
f
o
r
I
m
p
ro
v
e
d
Clas
sif
ic
a
ti
o
n
Eff
e
c
ti
v
e
n
e
ss
”
,
In
ter
n
a
ti
o
n
a
l
W
o
rk
sh
o
p
o
n
Ad
a
p
ti
v
e
M
u
l
ti
me
d
ia
Re
triev
a
l,
p
p
.
1
-
1
5
,
2
0
1
5
.
[1
1
]
R.
M
a
lh
e
ir
o
,
R.
P
a
n
d
a
,
P
.
G
o
m
e
s,
a
n
d
R.
P
.
P
a
iv
a
,
“
M
u
sic
Em
o
ti
o
n
Re
c
o
g
n
it
i
o
n
f
ro
m
Ly
rics
:
A
Co
m
p
a
ra
ti
v
e
S
tu
d
y
”
,
in
I
n
ter
n
a
t
io
n
a
l
W
o
rk
sh
o
o
n
M
a
c
h
in
e
L
e
a
rn
i
n
g
a
n
d
M
u
sic
(
M
M
L
)
,
p
p
.
9
-
1
2
,
2
0
1
3
.
[1
2
]
R.
P
a
n
d
a
,
R.
M
a
lh
e
iro
,
B.
R
o
c
h
a
,
A
.
Oliv
e
ira,
a
n
d
R.
P
.
P
a
iv
a
,
“
M
u
lt
i
-
M
o
d
a
l
M
u
sic
Em
o
ti
o
n
Re
c
o
g
n
it
io
n
:
A
Ne
w
Da
tas
e
t,
M
e
th
o
d
o
l
o
g
y
a
n
d
Co
m
p
a
ra
ti
v
e
A
n
a
l
y
sis”
,
1
0
’th
In
ter
n
a
ti
o
n
a
l
S
y
mp
o
si
u
m
o
n
C
o
mp
u
ter
M
u
sic
M
u
lt
id
isc
ip
l
in
a
ry
Res
e
a
rc
h
,
p
p
.
1
-
1
3
,
2
0
1
3
.
[1
3
]
F
.
X
u
e
,
Ha
o
;
X
u
e
,
L
ik
e
;
S
u
,
“
M
u
lt
im
o
d
a
l
M
u
sic
M
o
o
d
Clas
sif
ica
ti
o
n
b
y
F
u
sio
n
o
f
A
u
d
io
a
n
d
Ly
ric
s
”
,
in
2
1
st
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
,
M
u
lt
iM
e
d
ia
M
o
d
e
li
n
g
,
p
p
.
2
6
-
3
7
,
2
0
1
5
.
[1
4
]
V
.
K
u
m
a
r,
“
M
o
o
d
Clas
sif
iac
ti
o
n
o
f
Ly
rics
u
sin
g
S
e
n
ti
W
o
rd
N
e
t”
,
In
t.
C
o
n
f.
C
o
mp
u
t.
C
o
mm
u
n
.
In
f
o
rm
a
ti
c
s
(
ICCCI
-
2013)
,
p
p
.
1
-
5
,
2
0
1
3
.
[1
5
]
A
.
Esu
li
,
F
.
S
e
b
a
stian
i
,
a
n
d
V.
G
.
M
o
ru
z
z
i,
“
S
ENT
IW
OR
DN
E
T
:
A
P
u
b
li
c
ly
A
v
a
il
a
b
le
L
e
x
ic
a
l
Re
so
u
rc
e
f
o
r
Op
in
i
o
n
M
in
in
g
”
,
Pro
c
.
L
r.
2
0
0
6
,
p
p
.
4
1
7
-
4
2
2
,
2
0
0
6
.
[1
6
]
A
.
Ja
m
d
a
r,
J.
A
b
ra
h
a
m
,
K.
Kh
a
n
n
a
,
a
n
d
R.
Du
b
e
y
,
“
E
m
o
ti
o
n
A
n
a
l
y
si
so
f
S
o
n
g
s
Ba
se
d
o
n
L
y
r
ica
l
a
n
d
A
u
d
i
o
F
e
a
tu
re
s”
,
In
t.
J
.
Arti
f.
I
n
tell.
Ap
p
l.
,
v
o
l.
6
,
n
o
.
3
,
p
p
.
3
5
-
5
0
,
2
0
1
5
.
[1
7
]
H.K.
P
a
lo
a
n
d
M
.
N.
M
o
h
a
n
ty
,
“
Clas
si
f
ica
ti
o
n
o
f
E
m
o
ti
o
n
a
l
S
p
e
e
c
h
o
f
Ch
il
d
re
n
Us
in
g
P
r
o
b
a
b
il
isti
c
Ne
u
ra
l
Ne
tw
o
rk
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
),
v
o
l
.
5
,
n
o
.
2
,
p
p
.
3
1
1
-
3
1
7
,
2
0
1
5
.
[1
8
]
V
.
L
.
Ng
u
y
e
n
,
D.
Ki
m
,
V
.
P
.
Ho
,
a
n
d
Y.
L
i
m
,
“
A
Ne
w
R
e
c
o
g
n
it
io
n
M
e
th
o
d
f
o
r
V
isu
a
li
z
i
n
g
M
u
sic
Em
o
ti
o
n
”
,
In
t.
J
.
El
e
c
tr.
Co
mp
u
t.
E
n
g
.
,
v
o
l
.
7
,
n
o
.
3
,
p
p
.
1
2
4
6
-
1
2
5
4
,
2
0
1
7
.
[1
9
]
M
.
S
u
d
a
rm
a
a
n
d
I.
G
.
Ha
rs
e
m
a
d
i,
“
De
sig
n
a
n
d
A
n
a
l
y
sis
S
y
s
tem
o
f
KN
N
a
n
d
ID3
A
lg
o
rit
h
m
f
o
r
M
u
sic
Clas
sif
ic
a
ti
o
n
b
a
se
d
o
n
M
o
o
d
F
e
a
tu
r
e
Ex
trac
ti
o
n
”
,
In
t
.
J
.
El
e
c
tr.
Co
mp
u
t
.
En
g
(
IJ
ECE
)
,
v
o
l.
7
,
n
o
.
1
,
p
p
.
4
8
6
-
4
9
5
,
2
0
1
7
.
[2
0
]
C.
S
c
ien
c
e
,
A
.
Ha
rjo
k
o
,
B.
Jim
b
a
ra
n
,
a
n
d
S
.
Uta
ra
,
“
M
u
sic
Re
c
o
m
m
e
n
d
a
ti
o
n
S
y
ste
m
Ba
se
d
o
n
Co
n
tex
t
Us
in
g
Ca
se
-
Ba
s
e
d
Re
a
so
n
in
g
a
n
d
S
e
lf
Org
a
n
izin
g
M
a
p
”
,
In
d
o
n
e
s.
J
.
El
e
c
tr.
En
g
.
C
o
mp
u
t.
S
c
i.
,
v
o
l.
4
,
n
o
.
2
,
p
p
.
4
5
9
-
4
6
4
,
2
0
1
6
.
[2
1
]
F
.
H.
Ra
c
h
m
a
n
,
R.
S
a
rn
o
,
a
n
d
C.
F
a
ti
c
h
a
h
,
“
CBE :
C
o
rp
u
s
-
Ba
se
d
o
f
Em
o
ti
o
n
f
o
r
Em
o
ti
o
n
De
tec
ti
o
n
i
n
T
e
x
t
Do
c
u
m
e
n
t
”
,
in
ICIT
ACE
E
,
2
0
1
6
,
p
p
.
3
3
1
-
33
5
.
[2
2
]
S
.
Ba
n
e
rjee
a
n
d
T
.
P
e
d
e
rse
n
,
“
An
A
d
a
p
ted
L
e
s
k
A
lg
o
rit
h
m
f
o
r
W
o
rd
S
e
n
se
Disa
m
b
ig
u
a
ti
o
n
Us
in
g
W
o
rd
Ne
t
”
,
in
T
h
ird
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
L
i
n
g
u
isti
c
s a
n
d
In
tell
ig
e
n
t
T
e
x
t
Pro
c
e
ss
in
g
,
p
p
.
1
3
6
-
1
4
5
,
2
0
0
2
.
[2
3
]
B.
Y.
P
ra
tam
a
a
n
d
R.
S
a
rn
o
,
“
P
e
rso
n
a
li
ty
Clas
si
f
ic
a
ti
o
n
Ba
se
d
o
n
T
w
it
ter
T
e
x
t
Us
in
g
Na
iv
e
B
a
y
e
s,
KN
N
a
n
d
S
V
M
”
,
in
2
0
1
5
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Da
t
a
a
n
d
S
o
ft
w
a
re
En
g
in
e
e
rin
g
(
ICo
DS
E)
,
p
p
.
1
7
0
1
7
4
,
2
0
1
5
.
[2
4
]
X
.
Hu
,
J.S
.
Do
w
n
ie,
C.
Lau
rier,
M
.
Ba
y
,
a
n
d
A
.
F
.
Eh
m
a
n
n
,
“
T
h
e
2
0
0
7
M
IRE
X
A
u
d
io
M
o
o
d
Clas
s
if
ica
ti
o
n
T
a
sk
:
L
e
ss
o
n
L
e
a
rn
e
d
Un
iv
e
rsit
y
o
f
Ill
in
o
is
a
t
Urb
a
n
a
-
Ch
a
m
p
a
ig
n
M
u
si
c
T
e
c
h
n
o
lo
g
y
G
ro
u
p
,
U
n
iv
e
rsitat
P
o
m
p
e
u
F
a
b
ra
c
lau
rier@iu
a
.
u
p
f
.
e
d
u
”
,
in
Pro
c
e
e
d
in
g
s
o
f
t
h
e
I
n
ter
n
a
t
io
n
a
l
S
y
mp
o
si
u
m
o
n
M
u
sic
I
n
fo
rm
a
ti
o
n
Retrie
v
a
l
,
p
p
.
4
6
2
-
4
6
7
,
2
0
0
8
.
[2
5
]
R.
E.
T
h
a
y
e
r,
T
h
e
Bi
o
p
sy
c
h
o
l
o
g
y
o
f
M
o
o
d
a
n
d
Aro
u
sa
l
,
Ne
w
Yo
rk
:
Ox
f
o
rd
Un
iv
e
rsit
y
P
re
ss
,
1
9
8
9
.
[2
6
]
Y.
Ya
n
g
a
n
d
H.H.
C
h
e
n
,
M
u
sic
E
mo
ti
o
n
Rec
o
g
n
it
i
o
n
,
CRC
P
re
ss
,
2
0
1
1
.
[2
7
]
B.
S
.
Ri
n
ty
a
rn
a
a
n
d
R.
S
a
rn
o
,
“
A
d
a
p
ted
W
e
ig
h
ted
G
ra
p
h
f
o
r
W
o
rd
S
e
n
se
Disa
m
b
ig
u
a
ti
o
n
”
,
Ic
o
ICT
,
2
0
1
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.