IA
E
S
In
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
In
t
e
l
l
i
g
e
n
c
e
(
IJ
-
AI
)
Vo
l
.
8
, N
o
. 1
, M
a
r
c
h
201
9
, p
p
.
6
3
~
7
6
IS
S
N
:
2
2
5
2
-
8938
,
DOI
:
1
0
.
1
1
5
9
1
/
i
j
a
i
.
v
8
.i
1
.p
p
6
3
-
7
6
r
63
Jou
r
n
al
h
om
e
page
:
ht
t
p:
/
/
i
ae
s
co
r
e
.c
o
m
/o
n
lin
e
/in
d
e
x
.p
h
p
/I
J
AI
An
i
m
p
r
o
v
e
d
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
wo
r
k
s
i
n
n
e
t
wo
r
k
s
we
i
g
h
t
s
a
d
j
u
s
t
m
e
n
t
f
o
r
t
r
a
i
n
i
n
g
r
e
a
l
-
wo
r
l
d
n
o
n
l
i
n
e
a
r
d
a
t
a
s
e
t
s
Li
m
En
g
A
i
k
1
, T
a
n
W
e
i H
o
n
g
2
, A
h
m
a
d
K
a
d
r
i J
u
n
o
h
3
1
,3
In
s
t
i
t
u
t
M
a
te
m
a
tik
K
e
ju
r
u
te
r
a
a
n
, U
n
iv
e
r
s
ity
Ma
l
a
y
s
i
a
P
e
r
l
i
s
,
0
2
6
0
0
A
r
a
u
,
P
e
r
l
i
s
,
Ma
l
a
y
s
i
a
.
2
Sc
h
o
o
l
o
f
M
e
c
ha
t
r
oni
c
E
ngi
ne
e
r
i
ng,
U
ni
ve
r
s
i
t
y
Ma
l
a
y
s
i
a
P
e
r
l
i
s
,
0
2
6
0
0
A
r
a
u
,
P
e
r
l
i
s
,
Ma
l
a
y
s
i
a
.
Ar
t
i
c
l
e
I
n
f
o
AB
S
T
RACT
Ar
t
i
c
l
e
h
i
s
t
o
r
y
:
Re
c
e
i
v
e
d
De
c
2
7,
201
8
Re
v
i
s
e
d
Fe
b
17,
201
9
Ac
c
e
p
t
e
d
F
e
b
2
8
,
2
0
1
9
In
n
e
u
ra
l
n
e
t
w
o
rk
s
,
t
h
e
a
c
c
u
ra
c
i
e
s
o
f
i
t
s
n
e
t
w
o
rk
s
a
re
m
a
i
n
l
y
re
l
y
i
n
g
o
n
t
w
o
im
p
o
r
ta
n
t
f
a
c
to
r
s
w
h
ic
h
a
r
e
th
e
c
e
n
te
r
s
a
n
d
th
e
n
e
tw
o
r
k
s
w
e
ig
h
t.
T
h
e
gr
a
di
e
nt
de
s
c
e
nt
a
l
gor
i
t
hm
i
s
a
w
i
de
l
y
us
e
d
w
e
i
ght
a
dj
us
t
m
e
nt
a
l
gor
i
t
hm
i
n
mo
s
t
o
f
n
e
u
r
a
l
n
e
t
wo
r
k
s
t
r
a
i
n
i
n
g
a
l
g
o
r
i
t
h
m.
Ho
we
v
e
r
,
t
h
e
me
t
h
o
d
i
s
k
n
o
wn
fo
r
i
t
s
w
e
a
k
n
e
s
s
fo
r
e
a
s
i
l
y
t
ra
p
i
n
l
o
c
a
l
m
i
n
i
m
a
.
It
s
u
ffe
rs
fro
m
a
ra
n
d
o
m
we
i
g
h
t
g
e
n
e
r
a
t
e
d
f
o
r
t
h
e
n
e
t
wo
r
k
s
d
u
r
i
n
g
i
n
i
t
i
a
l
s
t
a
g
e
o
f
t
r
a
i
n
i
n
g
a
t
i
n
p
u
t
la
y
e
r
to
h
id
d
e
n
la
y
e
r
n
e
tw
o
r
k
s
.
Th
e
p
e
r
f
o
r
m
a
n
c
e
o
f
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
ne
t
w
or
ks
(
R
B
F
N
)
ha
s
be
e
n
i
m
pr
ove
d
f
r
om
di
f
f
e
r
e
nt
pe
r
s
pe
c
t
i
ve
s
,
i
nc
l
udi
ng
cen
t
r
o
i
d
i
n
i
t
i
al
i
zat
i
o
n
p
r
o
b
l
em
t
o
w
ei
g
h
t
co
r
r
ect
i
o
n
st
ag
e
o
v
er
t
h
e
y
ear
s.
Un
f
o
r
t
u
n
a
t
e
l
y
,
t
h
e
s
o
l
u
t
i
o
n
d
o
e
s
n
o
t
p
r
o
v
i
d
e
a
g
o
o
d
t
r
a
d
e
-
of
f
be
t
w
e
en
q
u
al
i
t
y
an
d
ef
f
i
ci
en
cy
o
f
t
h
e
w
ei
g
h
t
p
r
o
d
u
ces
b
y
t
h
e
al
g
o
r
i
t
h
m
.
T
o
so
l
v
e
t
h
i
s
pr
obl
e
m
,
a
n
i
m
pr
ove
d
gr
a
di
e
nt
de
s
c
e
nt
a
l
gor
i
t
hm
f
or
f
i
ndi
ng
i
ni
t
i
a
l
w
e
i
ght
an
d
i
m
p
r
o
v
e
t
h
e
o
v
er
al
l
n
et
w
o
r
k
s
w
ei
g
h
t
i
s
p
r
o
p
o
sed
.
T
h
i
s
i
m
p
r
o
v
ed
v
er
si
o
n
al
g
o
r
i
t
h
m
i
s
i
n
co
r
p
o
r
at
ed
in
to
R
B
F
N
tr
a
in
in
g
a
lg
o
r
ith
m
f
o
r
u
p
d
a
tin
g
w
e
ig
h
t.
He
n
c
e
,
t
h
i
s
p
a
p
e
r
p
r
e
s
e
n
t
e
d
a
n
i
mp
r
o
v
e
d
R
B
F
N
i
n
t
e
r
m
o
f
a
l
g
o
r
i
t
h
m
f
o
r
im
p
r
o
v
in
g
th
e
w
e
ig
h
t
a
d
ju
s
tm
e
n
t
in
R
B
F
N
d
u
r
in
g
tr
a
in
in
g
p
r
o
c
e
s
s
.
T
h
e
pr
opos
e
d
t
r
a
i
ni
ng
a
l
gor
i
t
hm
,
w
hi
c
h
us
e
s
i
m
pr
ove
d
gr
a
di
e
nt
de
s
c
e
nt
al
g
o
r
i
t
h
m
fo
r
w
e
i
g
h
t
a
d
j
u
s
t
m
e
n
t
fo
r
t
ra
i
n
i
n
g
R
B
F
N
,
o
b
t
a
i
n
e
d
s
i
g
n
i
fi
c
a
n
t
i
m
p
ro
v
e
m
e
n
t
in
p
r
e
d
ic
tio
n
s
c
o
m
p
a
r
e
d
to
th
e
s
ta
n
d
a
r
d
R
B
F
N
.
T
h
e
p
r
o
p
o
s
e
d
tr
a
in
in
g
al
g
o
r
i
t
h
m
w
as
i
m
p
l
em
en
t
ed
i
n
M
A
T
L
A
B
en
v
i
r
o
n
m
en
t
.
T
h
e
p
r
o
p
o
sed
im
p
r
o
v
e
d
n
e
tw
o
r
k
c
a
lle
d
I
R
B
F
N
w
a
s
te
s
te
d
a
ga
i
ns
t
t
he
s
t
a
nda
r
d
R
B
F
N
i
n
pr
e
di
c
t
i
ons
.
T
he
e
xpe
r
i
m
e
nt
a
l
m
ode
l
s
w
e
r
e
t
e
s
t
e
d
on
f
our
l
i
t
e
r
a
t
ur
e
s
nonl
i
ne
a
r
f
unc
t
i
on
a
nd
f
our
r
e
a
l
-
wo
r
l
d
a
p
p
l
i
c
a
t
i
o
n
p
r
o
b
l
e
ms
,
p
a
r
t
i
c
u
l
a
r
l
y
i
n
Ai
r
p
o
l
l
u
t
a
n
t
p
r
o
b
l
e
m,
B
i
o
c
h
e
mi
c
a
l
Ox
y
g
e
n
De
ma
n
d
(
B
OD)
p
r
o
b
l
e
m,
Ph
y
t
o
p
l
a
n
k
t
o
n
p
r
o
b
le
m
,
a
n
d
f
o
r
e
x
p
a
ir
E
U
R
U
S
D
.
T
h
e
r
e
s
u
lts
a
r
e
c
o
m
p
a
r
e
d
to
I
R
B
F
N
f
o
r
r
o
o
t
m
e
a
n
s
q
u
a
r
e
e
r
r
o
r
(
R
M
S
E
)
v
a
lu
e
s
w
ith
s
ta
n
d
a
r
d
R
B
F
N
.
Th
e
I
R
B
FN
y
i
e
l
d
e
d
a
p
r
o
m
i
s
i
n
g
r
e
s
u
l
t
w
i
t
h
a
n
a
v
e
r
a
g
e
i
m
p
r
o
v
e
m
e
n
t
pe
r
c
e
nt
a
ge
m
or
e
t
ha
n
40
pe
r
c
e
nt
i
n
R
M
S
E
.
Ke
y
wo
r
d
s
:
Gr
a
d
i
e
n
t
De
s
c
e
n
t
Im
p
ro
v
e
d
R
B
F
N
Ne
u
r
a
l
n
e
t
wo
r
k
Ra
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
We
i
g
h
t
A
d
j
u
s
t
m
e
n
t
Co
p
y
r
i
g
h
t
©
2
0
1
9
In
s
t
i
t
u
t
e
o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
r
i
n
g
a
n
d
S
c
i
e
n
c
e
.
Al
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
Co
r
r
e
s
p
o
n
d
i
n
g
Au
t
h
o
r
:
Li
m
En
g
A
i
k
,
In
s
t
i
t
u
t
M
a
t
e
m
a
t
i
k
K
e
j
u
ru
t
e
ra
a
n
,
Un
i
v
e
r
s
i
t
y
Ma
l
a
y
s
i
a
P
e
r
l
i
s
,
02600
A
r
a
u,
P
e
r
l
i
s
,
M
a
l
a
ys
i
a
.
Em
a
i
l
:
e
.
a
.
l
i
m
8
0
@
g
m
a
i
l
1.
IN
T
R
O
D
U
C
T
IO
N
Th
e
t
e
r
m
R
a
d
i
a
l
B
a
s
i
s
F
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
(
R
B
F
N
)
i
s
a
s
s
o
c
i
a
t
e
d
w
i
t
h
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
(
R
B
F
)
in
s
in
g
le
-
la
y
e
r
e
d
n
e
tw
o
r
k
s
w
ith
s
tr
u
c
tu
r
e
a
s
s
h
o
w
n
in
F
ig
u
r
e
1
.
R
a
d
ia
l
B
a
s
is
F
u
n
c
tio
n
(
R
B
F
)
N
e
tw
o
r
k
s
de
r
i
ve
d
f
r
om
t
he
t
he
or
y
of
f
unc
t
i
on
a
ppr
oxi
m
a
t
i
on.
I
t
w
a
s
or
i
gi
na
l
l
y
us
e
d
i
n
e
xa
c
t
i
nt
e
r
pol
a
t
i
on
i
n
mu
l
t
i
d
i
me
n
s
i
o
n
a
l
s
p
a
c
e
b
y
M
o
o
d
y
a
n
d
D
a
r
k
e
n
[1
]
.
R
B
F
N
d
is
p
la
y
e
d
its
a
d
v
a
n
ta
g
e
s
o
v
e
r
o
th
e
r
ty
p
e
s
o
f
n
e
u
r
a
l
ne
t
w
or
ks
w
i
t
h
be
t
t
e
r
a
ppr
oxi
m
a
t
i
on
a
bi
l
i
t
i
e
s
,
s
i
m
pl
e
ne
t
w
or
k
de
s
i
gn
a
nd
f
a
s
t
e
r
l
e
a
r
ni
ng
a
l
gor
i
t
hm
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
r
IS
S
N
:
2252
-
8938
IJ
-
AI
Vo
l
.
8
, N
o
.
1,
M
a
r
c
h
201
9
:
6
3
–
7
6
64
Co
n
s
e
q
u
e
n
t
l
y
,
RBF
N
b
e
c
o
m
e
s
a
p
o
p
u
l
a
r
t
o
o
l
s
a
m
o
n
g
r
e
s
e
a
r
c
h
e
r
s
f
o
r
m
a
n
y
a
p
p
l
i
c
a
t
i
o
n
s
s
u
c
h
a
s
cl
as
s
i
f
i
cat
i
o
n
,
p
at
t
er
n
r
eco
g
n
i
t
i
o
n
an
d
ap
p
r
o
x
i
m
at
i
o
n
m
o
d
el
.
Nu
m
e
r
o
u
s
r
e
s
e
a
r
c
h
e
r
s
wh
o
m
h
a
v
e
b
e
e
n
wo
r
k
i
n
g
t
o
i
m
p
r
o
v
e
s
R
B
F
N
t
r
a
i
n
i
n
g
a
l
g
o
r
i
t
h
m
s
,
s
e
t
al
o
n
g
s
i
d
e
t
h
e
s
t
a
nda
r
d
t
e
c
hni
que
s
[2
–
8]
.
R
B
F
N
a
r
e
u
s
e
f
u
l
in
a
p
p
r
o
x
im
a
tio
n
p
r
o
b
le
m
s
,
b
u
t
it
is
tim
e
-
co
n
s
u
m
i
n
g
t
o
t
r
ai
n
t
h
e
n
et
w
o
r
k
s
as
i
t
co
n
n
ect
t
o
a
l
ar
g
e
n
u
m
b
er
s
o
f
t
r
ai
n
i
n
g
d
at
a,
n
o
n
et
h
el
es
s
g
en
er
at
e
h
i
g
h
er
r
o
r
d
u
e
t
o
p
o
s
s
i
b
l
e
i
n
v
al
i
d
da
t
a
or
de
s
i
gna
t
i
on
of
w
e
i
ght
s
i
n
hi
dde
n
l
a
ye
r
.
A
l
t
hough
a
s
t
a
nda
r
d
R
B
F
N
ha
s
be
e
n
pr
ove
d
by
S
a
r
i
m
ve
i
s
[2
]
to
b
e
f
a
s
te
r
in
tr
a
in
in
g
w
ith
r
e
a
s
o
n
a
b
le
a
c
c
u
r
a
c
y
,
it
s
till
p
r
o
d
u
c
e
s
s
u
b
s
ta
n
tia
l
er
r
o
r
.
S
u
ch
o
u
t
co
m
es
w
er
e
cau
s
ed
b
y
t
h
e
s
t
an
d
ar
d
R
B
F
N
t
ha
t
l
a
c
k
t
he
a
bi
l
i
t
y
t
o
c
om
put
e
s
a
c
c
ur
a
t
e
w
e
i
ght
s
fo
r
h
i
d
d
e
n
l
a
y
e
r
t
h
a
t
a
b
l
e
t
o
re
p
re
s
e
n
t
t
h
e
l
e
v
e
l
o
f
i
m
p
o
rt
a
n
c
e
fo
r
e
a
c
h
h
i
d
d
e
n
n
o
d
e
s
.
T
h
ro
u
g
h
i
m
p
ro
v
i
n
g
t
h
e
we
i
g
h
t
s
c
a
l
c
u
l
a
t
i
o
n
e
q
u
a
t
i
o
n
i
n
s
t
a
n
d
a
r
d
R
B
F
N
d
u
r
i
n
g
u
p
d
a
t
i
n
g
s
t
a
g
e
,
we
c
a
n
f
i
x
t
h
e
p
r
o
b
l
em
s
t
at
ed
ab
o
v
e.
No
t
e
d
t
h
a
t
t
h
e
m
o
r
e
a
c
c
u
r
a
t
e
t
h
e
we
i
g
h
t
s
a
r
e
a
s
s
i
g
n
s
t
o
e
a
c
h
n
o
d
e
,
m
o
r
e
a
c
c
u
r
a
t
e
t
h
e
i
n
f
o
r
m
a
t
i
o
n
th
a
t
f
e
e
d
s
to
th
e
o
u
tp
u
t
la
y
e
r
o
f
n
e
tw
o
r
k
,
r
e
s
u
ltin
g
in
a
c
c
u
r
a
te
r
e
s
u
lts
.
I
n
th
is
p
a
p
e
r
,
a
n
im
p
r
o
v
e
d
R
B
F
N
th
a
t
out
pe
r
f
or
m
s
t
he
s
t
a
nda
r
d
R
B
F
N
in
te
r
m
o
f
a
c
c
u
r
a
c
ie
s
is
p
r
e
s
e
n
te
d
.
T
h
e
p
r
o
p
o
s
e
d
n
e
tw
o
r
k
s
e
f
f
ic
ie
n
c
y
is
de
m
ons
t
r
a
t
e
d
t
hr
ough
t
he
a
ppl
i
c
a
t
i
on
of
e
i
ght
e
xpe
r
i
m
e
nt
a
l
m
ode
l
s
,
w
i
t
h
f
our
nonl
i
ne
a
r
m
ode
l
f
r
om
lite
r
a
tu
r
e
s
,
th
r
e
e
r
e
a
l
-
wo
r
l
d
p
r
o
b
l
e
m
s
d
a
t
a
o
b
t
a
i
n
e
d
f
r
o
m
L
i
m
[9
]
an
d
o
n
e
r
eal
-
wo
r
l
d
t
i
m
e
-
se
r
i
e
s
d
a
t
a
f
r
o
m
XM
m
e
t
a
t
r
a
d
e
r
4
tr
a
d
in
g
p
la
tf
o
r
m
[1
0
]
.
T
h
e
a
d
v
a
n
ta
g
e
s
o
f
th
e
p
r
e
s
e
n
te
d
p
r
o
p
o
s
ed
n
et
w
o
r
k
s
ar
e
i
d
en
t
i
f
i
ed
an
d
t
h
e
r
es
u
l
t
s
ar
e
co
m
p
ar
ed
w
i
t
h
s
t
an
d
ar
d
R
B
F
N
an
d
d
i
s
cu
s
s
ed
.
Fi
g
u
r
e
1
.
Th
e
a
r
c
h
i
t
e
c
t
u
r
e
o
f
R
B
F
N
2.
RE
L
AT
E
D
W
O
RK
S
Ra
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
(
RBF
N
)
a
r
e
w
e
l
l
-
know
n
f
or
i
t
s
a
bi
l
i
t
y
t
o
ge
ne
r
a
l
i
z
e
a
nd
a
ppr
oxi
m
a
t
e
a
s
am
p
l
e
d
at
a
w
i
t
h
o
u
t
t
h
e
r
eq
u
i
r
em
en
t
f
o
r
t
h
e
eq
u
at
i
o
n
an
d
co
ef
f
i
ci
en
t
s
,
p
ar
t
i
cu
l
ar
l
y
w
h
en
an
u
n
k
n
o
w
n
m
o
d
el
de
s
c
r
i
bi
ng
a
n
unknow
n
c
om
pl
e
x
r
e
l
a
t
i
on
w
i
t
h
a
bund
an
t
t
r
ai
n
i
n
g
d
at
a.
D
u
e
t
o
t
h
ei
r
ab
i
l
i
t
y
t
o
g
en
er
al
i
ze
su
b
st
a
n
t
i
a
l
l
y
,
R
B
F
N
a
r
e
u
su
a
l
l
y
se
l
e
c
t
e
d
f
o
r
t
h
i
s
p
u
r
p
o
se
[1
1
-
23]
.
F
u
r
th
e
r
m
o
r
e
,
in
th
is
b
ig
d
a
ta
e
r
a
,
m
a
n
y
dom
a
i
ns
s
uc
h
a
s
i
m
a
ge
pr
oc
e
s
s
i
ng,
t
e
xt
c
a
t
e
gor
i
z
a
t
i
on,
bi
om
e
t
r
i
c
,
m
ic
r
o
a
r
r
a
y
,
e
tc
.
h
a
d
th
e
s
iz
e
o
f
d
a
ta
s
e
ts
s
o
la
r
g
e
, th
a
t r
e
a
l
-
tim
e
s
y
s
te
m
r
e
q
u
ir
e
s
lo
n
g
tim
e
a
n
d
m
e
m
o
r
y
s
to
r
a
g
e
to
p
r
o
c
e
s
s
th
e
m
.
Un
d
e
r
t
h
e
s
a
m
e
g
r
o
u
p
,
t
h
e
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
n
e
u
r
a
l
n
e
t
wo
r
k
s
a
l
s
o
c
a
n
g
e
n
e
r
a
l
i
z
e
a
n
d
a
p
p
r
o
x
i
m
a
t
e
co
m
p
l
ex
d
at
as
et
.
H
o
w
ev
er
,
d
u
e
t
o
t
he
a
r
c
hi
t
e
c
t
ur
e
of
ba
c
kpr
opa
ga
t
i
on
ne
ur
a
l
ne
t
w
or
ks
t
ha
t
w
or
ks
i
n
f
or
w
a
r
d
an
d
b
ack
w
ar
d
d
i
r
ect
i
o
n
,
d
el
ay
s
i
n
t
r
ai
n
i
n
g
t
i
m
e
i
s
u
n
av
o
i
d
ab
l
e.
F
u
r
t
h
er
m
o
r
e,
i
n
t
er
m
o
f
ap
p
r
o
x
i
m
at
i
o
n
,
mu
l
t
i
p
l
e
l
i
t
e
r
a
t
u
r
e
s
s
h
o
w
e
d
t
h
a
t
t
h
e
R
B
F
N
i
s
mo
r
e
s
u
p
e
r
i
o
r
t
h
a
n
b
a
c
k
p
r
o
p
a
g
a
t
i
on
ne
ur
a
l
ne
t
w
or
ks
,
i
n
t
e
r
m
of
tr
a
in
in
g
s
p
e
e
d
a
n
d
a
c
c
u
r
a
c
y
[2
4
-
28]
.
Tw
o
m
a
i
n
c
r
i
t
e
r
i
a
t
h
a
t
d
e
t
e
r
m
i
n
e
t
h
e
a
c
c
u
r
a
c
y
o
f
a
n
R
B
F
N
a
r
e
t
h
e
in
itia
l
c
e
n
tr
e
a
n
d
th
e
n
e
tw
o
r
k
s
w
e
ig
h
ts
.
I
n
th
is
p
a
p
e
r
,
w
e
f
o
c
u
s
m
a
in
ly
o
n
n
e
tw
o
r
k
s
w
e
ig
h
t
a
d
ju
s
tm
e
n
t.
Th
r
o
u
g
h
o
u
t
m
a
n
y
l
i
t
e
r
a
t
u
r
e
s
,
w
e
f
o
u
n
d
t
h
a
t
w
e
i
g
h
t
a
d
j
us
t
m
e
nt
i
n
R
B
F
N
pl
a
y
vi
t
a
l
r
ol
e
i
n
de
t
e
r
m
i
ne
x
1
x
2
x
3
x
N
.
.
.
.
.
.
In
p
u
t l
a
y
e
r
Hi
d
d
e
n
L
a
y
e
r
o
f
RB
F
Ou
t
p
u
t
L
a
y
e
r
ψ
ψ
ψ
ψ
w
N
.
.
.
w
2
w
3
w
1
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
IS
S
N
:
2252
-
8938
r
An
i
m
p
r
o
v
e
d
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
in
n
e
tw
o
r
k
s
w
e
ig
h
ts
a
d
ju
s
tm
e
n
t fo
r
…
(
Li
m
E
n
g
A
i
k
)
65
its
a
c
c
u
r
a
c
y
in
a
p
p
r
o
x
im
a
tio
n
p
r
o
b
le
m
.
W
a
n
g
e
t
a
l.
[2
9
]
st
u
d
y
t
h
e
e
f
f
e
c
t
i
v
e
n
e
ss
o
f
e
x
t
r
e
m
e
l
e
a
r
n
i
n
g
m
a
c
h
i
n
e
by
a
dj
us
t
i
ng
i
t
s
ne
t
w
or
ks
w
e
i
ght
s
,
i
t
i
s
f
ound
t
ha
t
ne
t
w
or
ks
w
e
i
ght
s
i
gni
f
i
c
a
nt
ly
in
f
lic
ts
in
o
u
tp
u
t
a
c
c
u
r
a
c
y
.
A
y
e
a
r
a
f
t
e
r
W
a
n
g
’
s
t
e
a
m
f
i
n
d
i
ng,
H
a
di
e
t
a
l
.
[3
0
]
pr
oof
t
he
w
e
a
k
r
obus
t
ne
s
s
i
n
m
ul
t
i
la
y
e
r
p
e
r
c
e
p
tr
o
n
a
n
d
RBF
N
i
n
e
s
t
i
m
a
t
i
n
g
l
a
r
g
e
d
a
t
a
s
e
t
w
a
s
d
u
e
t
o
t
h
e
l
a
c
k
i
n
w
e
i
g
h
t
c
o
r
r
e
c
t
i
o
n
a
l
g
o
r
i
t
h
m
i
n
b
o
t
h
t
h
e
n
e
t
w
o
r
k
s
.
Th
e
w
o
r
k
r
e
l
a
t
e
d
t
o
w
e
i
g
h
t
c
o
r
r
e
c
t
i
o
n
i
s
t
h
e
n
s
u
p
p
o
r
t
e
d
b
y
G
o
d
o
y
e
t
a
l
.
[3
1
]
an
d
S
o
d
h
i
et
al
.
[3
2
]
th
a
t
p
r
o
o
f
wh
e
n
a
b
e
t
t
e
r
we
i
g
h
t
i
s
as
s
i
g
n
ed
t
o
t
h
e
n
et
w
o
r
k
s
,
t
h
e
o
u
t
p
u
t
y
i
el
d
s
i
g
n
i
f
i
can
t
l
y
b
et
t
er
i
n
accu
r
acy
.
R
ecen
t
l
y
,
Ca
o
e
t
a
l
.
[3
3
]
in
th
e
ir
r
e
v
ie
w
w
o
r
k
,
s
ta
te
d
th
a
t
r
a
n
d
o
m
w
e
ig
h
t
a
p
p
r
o
a
c
h
e
s
in
R
B
F
N
c
a
u
s
e
s
th
e
r
e
d
u
c
e
th
e
n
e
tw
o
r
k
a
c
c
u
r
a
c
y
in
m
o
s
t a
p
p
r
o
x
im
a
tio
n
c
a
s
e
s
tu
d
ie
s
.
In
fi
n
d
i
n
g
s
o
l
u
t
io
n
to
o
b
ta
in
e
d
b
e
tte
r
w
e
ig
h
ts
a
d
ju
s
tm
e
n
t
a
n
d
g
a
in
in
g
a
c
c
u
r
a
te
a
p
p
r
o
x
im
a
tio
n
,
Na
v
a
r
r
o
e
t
a
l
.
[3
4
,
35]
de
m
ons
t
r
a
t
e
d
t
ha
t
t
he
hybr
i
d
of
e
vol
ut
i
ona
r
y
a
l
gor
i
t
hm
s
uc
h
a
s
pa
r
t
i
c
l
e
s
w
a
r
m
opt
i
m
i
z
a
t
i
on
(
P
S
O
)
w
i
t
h
R
B
F
N
s
how
s
a
good
pe
r
f
or
m
a
nc
e
i
n
c
l
a
s
s
i
f
i
c
a
t
i
on
pr
obl
e
m
s
a
nd
a
ppr
oxi
m
a
tio
n
pr
obl
e
m
s
.
T
hi
s
w
or
k
i
s
e
xt
e
nde
d
by
L
e
ung
e
t
a
l
.
[3
6
,
37]
by
i
nt
r
oduc
i
ng
a
nove
l
P
S
O
m
e
t
hod
or
a
dj
us
t
i
ng
t
he
RBF
N
n
e
t
w
o
r
k
s
w
e
i
g
h
t
.
M
o
r
e
o
v
e
r
,
t
h
e
f
o
l
l
o
w
i
n
g
y
e
a
r
s
,
e
v
o
l
u
t
i
o
n
a
r
y
a
l
g
or
i
t
hm
t
ur
n
popul
a
r
a
nd
be
c
om
i
ng
th
e
f
o
c
u
s
to
o
ls
f
o
r
w
e
ig
h
t
a
d
ju
s
tm
e
n
t
in
R
B
F
N
,
w
h
e
r
e
th
e
c
o
m
b
in
a
tio
n
o
f
m
u
ltip
le
e
v
o
lu
tio
n
a
r
y
a
lg
o
r
ith
m
su
c
h
a
s
g
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
a
n
d
P
S
O
w
i
t
h
R
B
F
N
[3
8
,
39]
.
S
o
m
e
r
e
s
e
a
r
c
h
e
r
u
s
e
s
a
m
o
d
if
ie
d
v
e
r
s
io
n
o
f
g
e
n
e
tic
al
g
o
r
i
t
h
m
as
n
et
w
o
r
k
s
w
ei
g
h
t
s
co
n
t
r
o
l
l
er
f
o
r
g
ai
n
i
n
g
b
et
t
er
r
es
u
l
t
s
[4
0
-
42]
an
d
m
o
s
t
r
ecen
t
o
n
u
s
i
n
g
m
o
d
i
f
i
ed
PSO
i
n
R
B
FN
f
o
r
w
e
i
g
h
t
a
d
j
u
s
t
m
e
n
t
b
y
M
i
r
j
a
l
i
l
i
[1
1
]
.
T
h
e
u
s
e
d
o
f
e
v
o
lu
tio
n
a
r
y
a
lg
o
r
ith
m
a
s
a
to
o
l
f
o
r
ne
t
w
or
k
w
e
i
ght
s
a
dj
us
t
m
e
nt
i
n
R
B
F
N
t
r
a
i
ni
ng
w
a
s
i
nde
e
d
a
good
m
e
t
hod
i
f
t
he
ne
t
w
or
ks
t
r
a
i
ni
ng
s
pe
e
d
a
nd
co
m
p
u
t
at
i
o
n
co
s
t
ar
e
n
o
t
m
ai
n
co
n
c
er
n
s
.
Al
o
n
g
s
i
d
e
wi
t
h
u
s
e
d
o
f
e
v
o
l
u
t
i
o
n
a
r
y
a
l
g
o
r
i
t
h
m
,
t
h
e
r
e
a
r
e
a
l
s
o
s
o
m
e
r
e
s
e
a
r
c
h
o
n
u
s
i
n
g
f
u
z
z
y
in
f
e
r
e
n
c
e
r
u
le
a
s
w
e
ig
h
t
s
e
ttin
g
f
o
r
R
B
F
N
w
h
ic
h
y
ie
ld
g
o
o
d
r
e
s
u
lts
[4
3
,
44]
.
H
o
w
e
v
e
r
,
th
e
a
p
p
lic
a
tio
n
o
f
fu
z
z
y
i
n
fe
re
n
c
e
i
n
R
BF
N
w
a
s
o
n
l
y
t
e
s
t
e
d
i
n
c
l
a
s
s
i
f
i
c
a
t
i
o
n
p
r
o
b
l
e
m
,
a
n
d
t
h
e
s
e
t
t
i
n
g
o
f
t
h
e
f
u
z
z
y
r
u
l
e
i
s
a
ch
al
l
en
g
i
n
g
t
as
k
f
o
r
r
eg
u
l
ar
r
es
ear
ch
er
s
t
h
at
d
o
n
o
t
h
av
e
f
u
zzy
t
h
eo
r
y
b
ack
g
r
o
u
n
d
.
I
n
co
n
t
r
ar
y
,
t
h
e
u
s
ed
o
f
st
a
t
i
st
i
c
a
l
m
e
t
h
o
d
i
n
o
b
t
a
i
n
i
n
g
su
i
t
a
b
l
e
w
e
i
g
h
t
f
o
r
R
B
F
N
w
e
r
e
re
p
o
rt
e
d
i
n
n
u
m
e
ro
u
s
l
i
t
e
ra
t
u
re
s
[4
5
-
48]
,
th
a
t
us
e
s
s
t
oc
ha
s
t
i
c
m
e
t
hod,
G
r
ove
r
s
e
a
r
c
hi
ng
a
l
gor
i
t
hm
,
Hi
e
r
a
r
c
h
y
M
a
r
k
o
v
i
a
n
m
a
t
r
i
x
a
n
d
a
t
t
r
i
b
u
t
e
d
c
l
a
s
s
co
r
r
el
at
i
o
n
m
et
h
o
d
f
o
r
t
h
at
p
u
r
p
o
s
es
.
F
u
r
t
h
er
m
o
r
e,
i
t
i
s
al
s
o
r
ep
o
r
t
ed
t
h
at
t
h
e
l
ear
n
i
n
g
al
g
o
r
i
t
h
m
f
o
r
n
et
w
o
r
k
s
tr
a
in
in
g
m
a
y
p
e
r
f
o
r
m
w
o
r
s
e
w
ith
th
e
in
c
r
e
a
s
e
s
o
f
d
a
ta
s
e
t
[4
9
]
.
In
c
o
rp
o
ra
t
i
n
g
s
t
a
t
i
s
t
i
c
a
l
m
e
t
h
o
d
i
n
t
o
R
B
F
N
i
s
a
co
m
p
l
i
cat
ed
t
as
k
,
r
eq
u
i
r
es
d
eep
unde
r
s
t
a
ndi
ng
i
n
f
l
ow
of
a
l
gor
i
t
hm
a
nd
s
t
a
t
i
s
t
i
c
a
l
t
he
or
y
t
o
i
m
pl
e
m
e
nt
s
uc
h
ap
p
r
o
ach
es
.
F
u
r
t
h
er
m
o
r
e,
t
h
e
r
ep
o
r
t
ed
i
m
p
r
o
v
em
en
t
o
n
r
es
u
l
t
s
i
n
m
en
t
i
o
n
ed
l
i
t
er
at
u
r
es
ar
e
n
o
t
s
i
g
n
i
f
i
can
t
en
o
u
g
h
f
o
r
t
h
e
co
m
p
l
i
cat
ed
i
m
p
l
em
en
t
at
i
o
n
p
r
o
ces
s
i
n
v
o
l
v
ed
.
M
o
r
eo
v
er
,
t
h
er
e
ar
e
o
t
h
er
s
s
t
u
d
i
es
u
s
i
n
g
ex
t
r
em
e
l
ear
n
i
n
g
m
ach
i
n
e
as
t
o
o
l
f
o
r
w
ei
g
h
t
ad
j
u
s
t
m
en
t
s
h
o
w
n
b
y
T
an
g
an
d
H
u
an
g
[5
0
]
an
d
,
Da
s
h
an
d
D
as
h
[5
1
]
.
B
o
th
s
tu
d
ie
s
u
s
e
d
e
x
tr
e
m
e
le
a
r
n
in
g
m
a
c
h
in
e
f
o
r
c
o
n
tr
o
llin
g
th
e
n
e
tw
o
r
k
s
w
e
ig
h
ts
d
u
r
in
g
tr
a
in
in
g
p
r
o
c
e
s
s
.
T
h
e
r
e
a
r
e
n
u
m
e
r
o
u
s
o
th
e
r
s
le
s
s
k
n
o
w
n
m
e
th
o
d
u
s
e
d
f
o
r
R
B
F
N
w
e
ig
h
t
a
d
ju
s
tm
e
n
t
a
r
e
pr
oj
e
c
t
i
on
ba
s
e
d
l
e
a
r
ni
ng
m
e
t
hod
[5
2
]
,
lin
e
a
r
in
te
r
v
a
l
r
e
g
r
e
s
s
io
n
w
e
ig
h
ts
m
e
th
o
d
[5
3
]
,
s
e
lf
-
or
ga
ni
z
e
d
R
B
F
N
ba
s
e
d
on
m
ut
ua
l
i
nf
or
m
a
t
i
on
a
nd
ne
ur
ons
a
c
t
i
vi
t
y
[5
4
]
,
n
o
v
e
l
tw
o
-
st
e
p
s
a
l
g
o
r
i
t
h
m
s
[5
5
]
,
s
e
lf
-
co
n
s
t
r
u
ct
i
n
g
le
a
s
t
-
Wi
l
c
o
x
o
n
m
e
t
h
o
d
[5
6
]
, a
n
d
se
m
i
-
an
al
y
t
i
c
co
m
p
u
t
at
i
o
n
o
f
La
u
r
e
n
t
s
e
r
i
e
s
[5
7
]
.
Fi
n
a
l
l
y
,
r
e
s
e
a
r
c
h
e
r
a
l
s
o
i
n
c
o
r
p
o
r
a
t
i
n
g
p
a
r
t
o
f
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
n
e
u
r
a
l
n
e
tw
o
r
k
s
a
lg
o
r
ith
m
th
a
t
in
v
o
lv
e
d
th
e
w
e
ig
h
t
u
p
d
a
tin
g
s
ta
g
e
in
to
R
B
F
N
to
e
n
h
a
n
c
e
th
e
p
e
r
f
o
r
m
a
n
c
e
o
f
w
e
ig
h
t
c
o
r
r
e
c
tio
n
a
b
ility
in
th
e
al
g
o
r
i
t
h
m
.
I
n
2
0
1
1
,
P
h
i
l
i
p
et
al
.
[5
8
]
us
e
d
f
e
e
d
fo
rw
a
rd
n
e
u
ra
l
n
e
t
w
o
rk
t
o
c
o
rre
c
t
t
h
e
w
e
i
g
h
t
fo
r
ba
c
kpr
opa
ga
t
i
on
ne
ur
a
l
ne
t
w
or
ks
dur
i
ng
l
e
a
r
ni
ng
s
t
ag
e.
I
n
t
h
e
s
am
e
y
ear
,
M
al
v
i
y
a
et
al
.
[5
9
]
us
e
d
ba
c
kpr
opa
ga
t
i
on
ne
ur
a
l
ne
t
w
or
k
a
nd
P
S
O
a
s
w
e
i
ght
t
uni
ng
t
ool
f
or
R
B
F
N
.
S
i
nc
e
t
he
n,
m
a
ny
r
e
s
e
a
r
c
h
f
oc
us
e
s
on
gr
a
di
e
nt
de
s
c
e
nt
(
G
D
)
a
l
gor
i
t
hm
w
e
r
e
pr
oc
e
e
ds
w
i
t
h
good
r
e
s
ul
t
s
,
a
s
G
D
a
l
gor
i
t
hm
w
a
s
t
he
w
e
i
ght
upda
t
i
ng
a
l
gor
i
t
hm
i
n
ba
c
kpr
opa
ga
t
i
on
ne
ur
a
l
ne
t
w
or
ks
.
In
2
0
1
2
,
M
o
h
s
e
n
i
e
t
a
l
.
[6
0
]
in
tr
o
d
u
c
e
d
a
m
ix
e
d
tr
a
in
in
g
a
lg
o
r
ith
m
u
s
in
g
b
a
c
k
p
r
o
p
a
g
a
tio
n
n
e
u
r
a
l
ne
t
w
or
ks
a
nd
va
r
i
a
bl
e
s
t
r
uc
t
ur
e
s
ys
t
e
m
t
o
opt
i
m
i
z
e
w
e
i
ght
upda
t
i
ng
i
n
R
B
F
N
.
A
s
s
a
f
e
t
a
l
.
[6
1
]
th
e
n
p
r
o
p
o
s
e
d
a
n
ew
t
r
ai
n
i
n
g
m
o
d
e
u
s
i
n
g
u
n
s
u
p
er
v
i
s
ed
cl
as
s
i
f
i
cat
i
o
n
al
g
o
r
i
t
h
m
w
h
er
e
t
h
e
co
n
n
ect
i
o
n
s
o
f
w
ei
g
h
t
s
ar
e
le
a
r
n
e
d
u
s
in
g
G
D
a
lg
o
r
ith
m
.
T
h
e
n
X
ie
e
t
a
l.
[6
2
]
fu
rt
h
e
r
e
x
t
e
n
d
e
d
t
h
e
u
s
e
s
o
f
G
D
a
l
g
o
ri
t
h
m
t
o
s
e
c
o
n
d
o
rd
e
r
GD
a
l
g
o
r
i
t
h
m
i
n
t
r
a
i
n
i
n
g
R
B
F
N
wh
e
r
e
t
h
e
we
i
g
h
t
i
s
a
d
j
u
s
t
e
d
d
u
r
i
n
g
t
r
a
i
n
i
n
g
.
T
h
o
u
g
h
t
h
e
GD
a
l
g
o
r
i
t
h
m
i
s
a
gr
e
a
t
a
ppr
oa
c
he
s
f
or
f
i
ndi
ng
w
e
i
gh
t,
h
o
w
e
v
e
r
,
th
e
G
D
a
lg
o
r
ith
m
a
ls
o
h
a
s
w
e
a
k
n
e
s
s
s
u
c
h
a
s
e
a
s
ily
tr
a
p
in
lo
c
a
l m
in
im
a
[6
3
,
64]
.
To
o
v
e
r
c
o
m
e
t
h
e
w
e
a
k
n
e
s
s
o
f
G
D
a
l
g
o
r
i
t
h
m
,
a
f
e
w
s
t
u
d
i
e
s
w
e
r
e
d
o
n
e
.
G
a
n
a
p
a
t
h
y
e
t
a
l
.
[6
5
]
de
m
ons
t
r
a
t
e
d
t
he
e
f
f
i
c
i
e
nc
y
of
a
n
i
m
pr
ove
d
s
t
e
e
pe
s
t
de
s
c
e
nt
a
l
gor
i
t
hm
f
or
de
t
e
r
m
i
ni
ng
w
e
i
ght
of
hi
dde
n
la
y
e
r
to
o
u
tp
u
t
la
y
e
r
n
e
tw
o
r
k
.
C
h
a
n
g
e
t
a
l.
[2
1
]
pr
opos
e
d
t
he
us
e
d
of
e
r
r
or
f
e
e
dba
c
k
s
c
he
m
e
i
n
R
B
F
N
f
or
we
i
g
h
t
c
o
r
r
e
c
t
i
o
n
wh
i
c
h
c
a
n
a
v
o
i
d
t
h
e
l
o
c
a
l
m
i
n
i
m
a
p
r
o
b
l
e
m
.
R
e
c
e
n
t
l
y
,
M
o
h
a
m
m
a
d
i
e
t
a
l
.
[1
6
]
ap
p
l
i
ed
a
ps
e
udo
-
in
v
e
r
s
e
a
lg
o
r
ith
m
f
o
r
c
a
lc
u
la
tin
g
w
e
ig
h
t
u
p
d
a
ti
ng
m
a
t
r
i
x
e
ve
n
w
i
t
h
l
a
r
ge
da
t
a
s
e
t
t
ha
t
c
a
n
f
i
x
th
e
lo
c
a
l
mi
n
i
ma
p
r
o
b
l
e
m.
In
t
h
e
m
e
n
t
i
o
n
e
d
l
i
t
e
ra
t
u
re
s
,
c
l
e
a
rl
y
m
o
s
t
a
p
p
ro
a
c
h
e
s
i
n
v
o
l
v
e
d
c
o
m
p
l
i
c
a
t
e
d
p
ro
c
e
s
s
to
p
e
r
f
o
r
m
th
e
c
o
m
p
u
ta
tio
n
a
n
d
s
o
m
e
a
p
p
r
o
a
c
h
e
s
y
ie
ld
n
o
t
s
ig
n
if
ic
a
n
t
im
p
r
o
v
e
m
e
n
t
in
a
c
c
u
r
a
c
ie
s
.
Fu
r
t
h
e
r
m
o
r
e
,
m
a
n
y
a
p
p
r
o
a
c
h
e
s
f
o
c
u
s
o
n
a
p
p
l
y
i
n
g
s
e
p
a
r
a
t
e
p
a
r
t
o
f
m
e
t
h
o
d
s
o
r
a
l
g
o
r
i
t
h
m
f
o
r
w
e
i
g
h
t
ad
j
u
s
t
m
en
t
i
n
s
t
ead
o
f
f
o
cu
s
i
n
g
o
n
i
m
p
r
o
v
i
n
g
R
B
F
N
i
n
t
er
n
al
al
g
o
r
i
th
m
.
I
n
th
is
p
a
p
e
r
,
w
e
f
o
c
u
s
o
n
im
p
r
o
v
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
r
IS
S
N
:
2252
-
8938
IJ
-
AI
Vo
l
.
8
, N
o
.
1,
M
a
r
c
h
201
9
:
6
3
–
7
6
66
th
e
in
te
r
n
a
l
a
lg
o
r
ith
m
f
o
r
w
e
ig
h
t
c
o
r
r
e
c
tio
n
s
ta
g
e
a
n
d
r
e
d
u
c
e
c
o
m
p
lic
a
te
d
a
lg
o
r
ith
m
w
h
ic
h
in
tu
r
n
m
a
k
in
g
RBF
N
a
l
g
o
r
i
t
h
m
p
r
e
c
e
d
e
f
a
s
t
e
r
c
o
m
p
u
t
a
t
i
o
n
w
i
t
h
l
o
w
c
o
m
p
u
t
a
t
i
o
n
c
o
s
t
.
T
h
i
s
p
a
p
e
r
i
s
o
r
g
a
n
i
z
e
d
w
i
t
h
t
h
e
fo
l
l
o
w
in
g
s
e
c
tio
n
d
e
s
c
r
ib
in
g
th
e
s
ta
n
d
a
r
d
R
B
F
N
,
f
o
llo
w
s
b
y
s
te
p
s
in
im
p
r
o
v
in
g
th
e
s
ta
n
d
a
r
d
R
B
F
N
a
n
d
u
s
e
d
fo
r
s
i
m
u
l
a
t
i
n
g
a
n
d
p
re
d
i
c
t
i
n
g
t
h
e
e
i
g
h
t
e
x
p
e
ri
m
e
n
t
a
l
m
o
d
e
l
s
.
T
h
e
n
,
i
n
s
e
c
t
i
o
n
4
d
i
s
c
u
s
s
e
d
t
h
e
re
s
u
l
t
s
o
f
e
a
c
h
mo
d
e
l
s
a
n
d
c
o
mp
a
r
e
d
w
i
t
h
s
t
a
n
d
a
r
d
R
B
F
N
f
o
r
a
c
c
u
ra
c
i
e
s
.
F
i
n
a
l
l
y
,
s
e
c
t
i
o
n
5
c
o
n
c
l
u
d
e
s
t
h
e
fi
n
d
i
n
g
s
a
n
d
di
s
c
us
s
e
d
s
om
e
f
ut
ur
e
w
or
k
t
ha
t
w
oul
d
he
l
p
i
n
i
m
pr
ovi
ng
t
he
pr
opos
e
d
i
m
pr
ove
d
R
B
F
N
.
3.
ME
T
H
O
D
O
L
O
G
Y
3.
1.
R
ad
i
al
b
as
i
s
f
u
n
c
t
i
on
n
e
t
w
or
k
(R
B
F
N
)
RBF
N
i
s
a
n
i
n
t
e
l
l
i
g
e
n
t
i
n
t
e
r
p
o
l
a
t
i
o
n
t
e
c
h
n
i
q
u
e
f
o
r
m
o
d
e
l
i
n
g
a
l
i
n
e
a
r
o
r
n
o
n
l
i
n
e
a
r
m
u
l
t
i
d
i
m
e
n
s
i
o
n
a
l
da
t
a
.
R
B
F
N
i
s
of
t
e
n
us
e
d
f
or
pr
e
di
c
t
i
on
pr
obl
e
m
s
.
I
t
s
ke
r
ne
l
ha
s
t
w
o
pa
r
a
m
e
t
e
r
s
;
t
he
c
e
nt
e
r
a
nd
i
t
s
r
a
di
us
.
Th
e
s
e
t
w
o
p
a
r
a
m
e
t
e
r
s
c
a
n
d
e
t
e
r
m
i
n
e
d
t
h
r
o
u
g
h
u
n
s
u
p
er
v
i
s
ed
o
r
s
u
p
er
v
i
s
ed
l
ear
n
i
n
g
as
p
r
o
p
o
s
ed
b
y
M
o
o
d
y
an
d
D
ar
k
en
i
n
1
9
8
9
[1
]
.
RBF
N
i
s
a
t
y
p
e
o
f
f
e
e
d
-
fo
rw
a
rd
n
e
u
ra
l
n
e
t
w
o
rk
.
T
h
e
n
e
t
w
o
rk
s
t
ru
c
t
u
re
c
o
n
s
i
s
t
s
o
f
th
r
e
e
-
la
y
e
r
n
e
tw
o
r
k
s
im
ila
r
to
m
u
lti
-
la
y
e
r
f
e
e
d
-
fo
rw
ar
d
n
et
w
o
r
k
.
I
n
t
h
e
n
et
w
o
r
k
s
t
r
u
ct
u
r
e
o
f
R
B
F
N
as
s
h
o
w
n
in
F
ig
u
r
e
1
,
th
e
f
ir
s
t
la
y
e
r
is
th
e
in
p
u
t
la
y
e
r
a
n
d
is
c
o
m
p
o
s
e
d
o
f
s
ig
n
a
l
s
o
u
r
c
e
n
o
d
e
s
.
T
h
e
s
e
c
o
n
d
la
y
e
r
is
a
hi
dde
n
l
a
ye
r
,
a
nd
t
he
num
be
r
of
node
s
of
t
he
l
a
ye
r
i
s
de
t
e
r
m
i
ne
d
by
t
he
na
t
ur
e
of
t
he
pr
obl
em
t
o
b
e
s
o
l
v
ed
an
d
t
h
e
ch
ar
act
er
i
s
t
i
cs
o
f
t
h
e
s
p
eci
f
i
c
p
r
o
b
l
em
.
T
h
e
t
r
an
s
f
er
f
u
n
ct
i
o
n
o
f
t
h
e
n
eu
r
o
n
s
i
n
t
h
i
s
l
ay
er
i
s
t
h
e
r
ad
i
al
ba
s
i
s
f
unc
t
i
on,
w
hi
c
h
i
s
a
gl
oba
l
f
unc
t
i
on
t
ha
t
r
e
s
pons
e
s
a
s
t
he
f
or
w
a
r
d
n
et
w
o
r
k
t
r
an
s
f
o
r
m
at
i
o
n
f
u
n
ct
i
o
n
.
Th
e
t
h
i
r
d
l
a
y
e
r
i
s
t
he
out
put
l
a
ye
r
,
w
hi
c
h
r
e
s
ponds
t
o
t
he
i
nput
l
a
ye
r
.
Th
e
m
a
i
n
i
d
e
a
o
f
t
h
e
R
B
F
N
i
s
t
o
u
s
e
t
h
e
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
a
s
t
h
e
"
b
a
s
e
"
o
f
t
h
e
h
i
d
d
e
n
l
a
y
e
r
u
n
i
t
to
b
u
ild
th
e
h
id
d
e
n
la
y
e
r
.
I
n
th
e
h
id
d
e
n
la
y
e
r
,
th
e
in
p
u
t
v
e
c
to
r
is
tr
a
n
s
f
o
r
m
e
d
.
T
h
u
s
,
lo
w
-
di
m
e
ns
i
ona
l
in
p
u
t
da
t
a
c
a
n
c
onve
r
t
t
o
a
hi
gh
-
di
m
e
ns
i
ona
l
s
pa
c
e
.
S
uc
h
a
ppr
oa
c
h
c
a
n
t
r
a
ns
f
or
m
a
l
i
ne
a
r
l
y
i
ns
e
pa
r
a
bl
e
pr
obl
e
m
i
n
a
l
o
w
-
di
m
e
ns
i
ona
l
s
pa
c
e
i
nt
o
a
l
i
ne
a
r
l
y
s
e
pa
r
a
bl
e
i
n
a
hi
gh
-
di
m
e
ns
i
ona
l
s
pa
c
e
,
t
he
r
e
by
s
ol
vi
ng
t
he
r
e
l
a
t
e
d
pr
obl
e
m
.
R
B
F
N
t
r
a
i
ni
ng
i
s
s
i
m
pl
e
a
n
d
ha
s
f
a
s
t
l
e
a
r
ni
ng
c
onve
r
ge
nc
e
.
I
t
c
a
n
a
ppr
oxi
m
a
t
e
a
ny
nonl
i
ne
a
r
fu
n
c
t
i
o
n
.
T
h
e
re
fo
re
,
t
h
e
R
B
F
N
h
a
s
a
e
x
t
e
n
s
i
v
e
a
p
p
l
i
c
a
t
i
o
n
s
i
n
p
a
t
t
e
rn
re
c
o
g
n
i
t
i
o
n
,
i
m
a
g
e
p
ro
c
e
s
s
i
n
g
,
pr
e
di
c
t
i
ons
a
nd
nonl
i
ne
a
r
c
ont
r
ol
.
Fr
o
m
t
h
e
p
e
r
s
p
e
c
t
i
v
e
o
f
f
u
n
c
t
i
o
n
a
p
p
r
o
x
i
m
a
t
i
o
n
,
R
B
FN
a
re
l
o
c
a
l
ap
p
r
o
x
i
m
at
i
o
n
s
.
I
f
t
h
e
n
u
m
b
er
o
f
n
eu
r
al
u
n
i
t
s
i
n
t
h
e
h
i
d
d
en
l
ay
er
r
each
es
a
cer
t
ai
n
l
ev
el
,
t
h
e
n
et
w
o
r
k
can
ap
p
r
o
x
i
m
at
e
an
y
co
n
t
i
n
u
o
u
s
f
u
n
ct
i
o
n
w
i
t
h
ar
b
i
t
r
ar
y
p
r
eci
s
i
o
n
.
I
n
ad
d
i
t
i
o
n
,
d
u
e
t
o
R
B
F
N
ad
o
p
t
s
a
l
i
n
ear
ma
p
p
i
n
g
r
e
l
a
t
i
o
n
s
h
i
p
b
e
t
w
e
e
n
t
h
e
o
u
t
p
ut
l
a
ye
r
a
nd
t
he
hi
dde
n
l
a
ye
r
;
t
he
ne
t
w
or
k
c
a
n
a
voi
d
t
he
c
om
pl
i
c
a
t
e
d
ba
c
k
-
pr
opa
ga
t
i
on
ope
r
a
t
i
on
a
s
i
n
ba
c
k
-
pr
opa
ga
t
i
on
ne
ur
a
l
ne
t
w
or
k.
T
he
r
e
f
or
e
,
i
n
R
B
F
N
,
t
he
s
pe
e
d
of
ope
r
a
t
i
on
a
nd
t
he
a
c
c
ur
a
c
y
of
nonl
i
ne
a
r
f
i
t
t
i
ng
a
r
e
i
m
pr
ove
d.
D
ur
i
ng
t
he
ope
r
a
t
i
on
t
o
s
o
lv
e
a
p
r
o
b
le
m
,
th
e
RBF
N
h
a
s
t
h
e
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
o
f
t
h
e
t
r
a
i
n
i
n
g
s
a
m
p
l
e
s
c
o
r
r
e
s
p
o
n
d
i
n
g
t
o
t
h
e
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
m
a
p
p
i
n
g
,
re
s
u
l
t
i
n
g
i
n
a
l
a
rg
e
a
m
o
u
n
t
o
f
n
e
t
w
o
rk
c
o
m
p
u
t
a
t
i
o
n
,
a
n
d
m
a
y
c
a
u
s
e
p
ro
b
l
e
m
s
i
n
s
o
l
v
i
n
g
t
h
e
n
e
t
w
o
rk
s
we
i
g
h
t
s
.
3.
2.
RB
F
N
mo
d
e
l
a
n
d
le
a
r
n
in
g
a
lg
o
r
it
h
m
Th
e
R
B
F
N
b
y
d
e
f
a
u
l
t
u
s
e
s
t
h
e
G
a
u
s
s
i
a
n
f
u
n
c
t
i
o
n
a
s
t
h
e
k
e
r
ne
l
f
unc
t
i
on
of
t
he
hi
dde
n
uni
t
(1
).
(
)
2
2
i
ik
i
k
XC
k
re
σ
−
−
∑
=
(1
)
Wh
e
r
e
k
r
is
th
e
o
u
tp
u
t
v
a
lu
e
o
f
th
e
k
-
th
G
a
u
s
s
ia
n
u
n
it
o
f
th
e
h
id
d
e
n
la
y
e
r
;
C
is
th
e
i
-
th
in
p
u
t
va
r
i
a
bl
e
va
l
ue
of
t
he
c
e
nt
e
r
of
t
he
ke
r
ne
l
of
t
he
k
-
th
G
a
u
s
s
ia
n
u
n
it;
k
σ
is
th
e
s
p
r
e
a
d
o
f
th
e
k
e
r
n
e
l
o
f
th
e
k
-
th
Ga
u
s
s
i
a
n
u
n
i
t
.
Th
e
s
u
p
e
r
v
i
s
e
d
l
e
a
r
n
i
n
g
r
u
l
e
u
s
e
d
t
o
a
d
j
u
s
t
t
h
e
l
i
n
k
w
e
i
g
h
t
i
n
g
v
a
l
u
e
b
e
t
w
e
e
n
t
h
e
h
i
d
d
e
n
u
n
i
t
an
d
t
h
e
out
put
uni
t
,
a
nd
t
he
t
hr
e
s
hol
d
va
l
ue
of
t
he
out
put
uni
t
.
T
he
di
f
f
i
c
ul
t
y
w
i
t
h
R
B
F
N
de
t
e
r
m
i
ne
s
t
he
num
be
r
of
hi
dde
n
uni
t
s
a
nd
t
he
c
e
nt
e
r
of
t
he
G
a
us
s
i
a
n
f
unc
t
i
on
a
nd
t
he
r
a
di
us
pa
r
a
m
e
t
e
r
s
.
T
he
t
w
o
pa
r
a
m
e
t
e
r
s
,
c
e
nt
e
r
an
d
r
ad
i
u
s
,
can
d
et
er
m
i
n
ed
b
y
s
u
p
e
rv
i
s
e
d
o
r
u
n
s
u
p
e
rv
i
s
e
d
l
e
a
rn
i
n
g
.
T
h
e
s
u
p
e
rv
i
s
e
d
l
e
a
rn
i
n
g
o
f
R
B
F
N
i
s
si
m
i
l
a
r
t
o
b
a
c
k
-
pr
opa
ga
t
i
on,
a
nd
t
he
l
e
a
r
ni
ng
r
ul
e
s
c
a
n
de
r
i
ve
d
by
m
i
ni
m
i
z
i
ng
t
he
s
um
of
s
qua
r
e
s
e
r
r
or
s
a
nd
th
e
g
r
a
d
ie
n
t
d
e
s
c
e
n
t
m
e
th
o
d
.
U
n
s
u
p
e
r
v
is
e
d
le
a
r
n
in
g
u
s
e
s
th
e
K
-
me
a
n
s
a
l
g
o
r
i
t
h
m
t
o
fi
n
d
t
h
e
c
l
u
s
t
e
r
c
e
n
t
e
r
o
f
th
e
s
a
m
p
le
a
s
th
e
c
e
n
te
r
o
f
th
e
G
a
u
s
s
ia
n
f
u
n
c
tio
n
o
f
e
a
c
h
h
id
d
e
n
u
n
it.
Th
o
u
g
h
R
B
F
N
h
a
s
a
n
e
x
t
e
n
s
i
v
e
ap
p
l
i
cat
i
o
n
s
i
n
p
at
t
er
n
r
eco
g
n
i
t
i
o
n
,
p
r
ed
i
ct
i
o
n
s
,
an
d
co
n
t
r
o
l
s
y
s
t
em
,
t
h
e
al
g
o
r
i
t
h
m
s
s
t
i
l
l
a
d
i
s
ad
v
an
t
ag
e
o
n
an
in
d
e
p
e
n
d
e
n
t
v
a
r
i
ab
l
es
h
av
e
t
h
e
s
am
e
p
o
s
i
t
i
o
n
,
s
o
t
h
e
co
n
t
o
u
r
o
f
t
h
e
k
er
n
el
f
u
n
ct
i
o
n
i
s
ci
r
cu
l
ar
.
H
o
w
ev
er
,
each
in
d
e
p
e
n
d
e
n
t
v
a
r
ia
b
le
h
a
s
d
if
f
e
r
e
n
t
in
f
lu
e
n
c
e
o
n
th
e
d
e
p
e
n
d
e
n
t
v
a
r
ia
b
le
,
a
n
d
th
e
c
o
n
to
u
r
o
f
th
e
k
e
r
n
e
l
f
u
n
c
tio
n
sh
o
u
l
d
b
e
e
l
l
i
p
t
i
c
a
l
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
IS
S
N
:
2252
-
8938
r
An
i
m
p
r
o
v
e
d
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
in
n
e
tw
o
r
k
s
w
e
ig
h
ts
a
d
ju
s
tm
e
n
t fo
r
…
(
Li
m
E
n
g
A
i
k
)
67
3.
3.
RB
F
N
im
p
le
m
e
n
t
a
t
io
n
, a
dv
a
nt
a
g
e
s
a
nd
l
i
m
i
t
a
t
i
o
n
Th
e
i
m
p
l
e
m
e
n
t
a
t
i
o
n
o
f
t
h
e
R
B
F
n
e
u
r
a
l
n
e
t
w
o
r
k
c
o
n
s
i
s
t
s
o
f
t
w
o
p
a
r
t
s
:
t
h
e
n
e
t
w
o
r
k
s
t
r
u
c
t
u
r
e
p
a
r
t
a
n
d
th
e
a
lg
o
r
ith
m
p
a
r
a
m
e
te
r
p
a
r
t.
T
h
e
n
e
tw
o
r
k
s
tr
u
c
tu
r
e
is
d
e
s
ig
n
e
d
to
d
e
te
r
m
in
e
th
e
n
u
m
b
e
r
o
f
n
o
d
e
s
in
th
e
hi
dde
n
l
a
ye
r
of
t
he
ne
t
w
or
k.
Th
e
p
a
r
t
o
f
t
h
e
a
l
g
o
r
i
t
h
m
p
a
r
a
m
e
t
e
r
s
i
s
d
e
t
e
r
m
i
n
e
d
f
o
r
t
h
e
t
h
r
e
e
i
m
p
o
r
t
a
n
t
pa
r
a
m
e
t
e
r
s
of
t
he
da
t
a
c
e
nt
e
r
of
t
he
r
a
di
a
l
ba
s
i
s
f
unc
t
i
on,
t
he
s
pr
e
a
d
c
ons
t
a
nt
a
nd
t
he
out
put
l
a
ye
r
w
e
i
ght
.
Si
n
c
e
t
h
e
n
u
m
b
e
r
o
f
h
i
d
d
e
n
l
a
y
e
r
n
o
d
e
s
i
s
c
o
n
s
i
s
t
e
n
t
w
i
t
h
t
h
e
n
u
m
b
e
r
o
f
sa
m
p
l
e
s,
a
n
d
t
h
e
c
e
n
t
e
r
o
f
t
h
e
r
a
d
i
a
l
ba
s
i
s
f
unc
t
i
on
i
s
t
he
s
a
m
pl
e
i
t
s
e
l
f
,
onl
y
t
he
s
pr
e
a
d
c
ons
t
a
nt
a
nd
t
he
out
put
l
a
ye
r
w
e
i
ght
ne
e
d
t
o
be
de
t
e
r
m
i
ne
d
wh
e
n
e
s
t
a
b
l
i
s
h
i
n
g
t
h
e
n
e
t
wo
r
k
.
Ho
we
v
e
r
,
wh
e
n
b
u
i
l
d
i
n
g
a
n
R
B
F
N,
t
h
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m
n
e
e
d
s
t
o
s
o
l
v
e
mo
r
e
pr
obl
e
m
s
,
i
nc
l
udi
ng
t
he
de
t
e
r
m
i
na
t
i
on
of
t
he
num
be
r
of
hi
dde
n
l
a
ye
r
node
s
,
t
he
de
t
e
r
m
i
na
t
i
on
of
t
he
da
t
a
c
e
nt
e
r
of
t
he
r
a
di
a
l
ba
s
i
s
f
unc
t
i
on,
t
he
s
pr
e
a
d
c
ons
t
a
nt
,
a
nd
t
he
c
or
r
e
c
t
i
on
of
t
he
w
e
ig
h
ts
o
f
th
e
out
put
l
a
ye
r
.
Th
e
e
s
t
a
b
l
i
s
h
m
e
n
t
a
n
d
t
r
a
i
n
i
n
g
o
f
t
h
e
ra
d
i
a
l
b
a
s
i
s
fu
n
c
t
i
o
n
n
e
u
ra
l
n
e
t
w
o
rk
i
s
c
o
m
p
o
s
e
d
o
f
t
w
o
st
a
g
e
s.
T
h
e
f
i
r
st
p
h
a
se
p
r
i
m
a
r
i
l
y
d
e
t
e
r
m
i
n
e
s
t
h
e
d
a
t
a
c
e
n
t
e
r
o
f
t
h
e
r
a
d
i
a
l
b
a
si
s
f
u
n
c
t
i
o
n
i
n
t
h
e
h
i
d
d
e
n
l
a
y
e
r
,
ty
p
ic
a
lly
u
s
e
d
a
s
th
e
c
e
n
te
r
'
s
s
e
lf
-
or
ga
ni
z
i
ng
s
e
l
e
c
t
i
on
m
e
t
hod
f
or
uns
upe
r
vi
s
e
d
l
e
a
r
ni
n
g
pr
oc
e
s
s
e
s
.
T
he
i
nput
sa
m
p
l
e
s
a
r
e
c
l
u
st
e
r
e
d
t
o
d
e
t
e
r
m
i
n
e
t
h
e
c
e
n
t
e
r
o
f
t
h
e
r
a
d
i
a
l
b
a
si
s
f
u
n
c
t
i
o
n
o
f
t
h
e
h
i
d
d
e
n
l
a
y
e
r
n
o
d
e
s.
Th
e
R
B
F
N
h
a
s
i
t
s
a
d
a
p
t
i
v
e
a
b
i
l
i
t
y
a
n
d
f
a
u
l
t
t
o
l
e
r
a
n
c
e
.
Th
e
r
e
f
o
r
e
,
w
h
i
l
e
d
e
a
l
i
n
g
w
i
t
h
t
h
e
c
o
m
p
l
e
x
ev
al
u
at
i
o
n
o
f
n
o
n
l
i
n
ear
i
t
y
,
t
h
e
fa
c
t
o
rs
t
h
a
t
h
a
v
e
n
o
s
i
g
n
i
fi
c
a
n
t
i
n
fl
u
e
n
c
e
o
n
t
h
e
e
v
a
l
u
a
t
i
o
n
re
s
u
l
t
s
c
a
n
ef
f
ect
i
v
el
y
el
i
m
i
n
at
ed
.
T
h
e
R
B
F
N
i
s
i
m
p
r
o
v
ed
v
i
a
t
h
e
s
u
p
er
v
i
s
ed
l
ear
n
i
n
g
r
u
l
es
ar
e
u
s
ed
t
o
ad
j
u
s
t
t
h
e
l
i
n
k
we
i
g
h
t
i
n
g
v
a
l
u
e
b
e
t
we
e
n
t
h
e
h
i
d
d
e
n
u
n
i
t
a
n
d
t
h
e
o
u
t
p
u
t
u
n
i
t
,
a
n
d
t
h
e
t
h
r
es
h
o
ld
o
f
th
e
o
u
tp
u
t
u
n
it.
Th
e
l
e
a
r
n
i
n
g
r
u
l
e
i
s
d
e
r
i
v
e
d
b
y
m
i
n
i
m
i
z
i
n
g
t
h
e
s
u
m
o
f
s
q
u
a
r
e
d
e
r
r
o
r
s
a
n
d
t
h
e
g
r
a
d
i
e
n
t
d
e
s
c
e
n
t
m
e
t
h
o
d
.
3.
4.
Im
p
r
o
v
e
d
R
B
F
N
(
IR
B
F
N
)
RBF
N
w
e
r
e
r
e
p
o
r
t
e
d
h
a
s
a
n
e
x
t
e
n
s
i
v
e
a
p
p
l
i
c
a
t
i
o
n
s
a
n
d
i
t
s
a
l
g
o
r
i
t
h
m
s
h
a
v
e
m
a
n
y
d
i
f
f
e
r
e
n
t
v
a
r
i
a
n
t
s
[1
1
,
1
9
,
6
6
,
67]
.
Y
e
h
a
n
d
C
h
e
n
[6
8
]
pr
opos
e
d
a
n
i
m
pr
ove
d
R
B
F
N
w
i
t
h
ke
r
ne
l
s
ha
pe
pa
r
a
m
e
t
e
r
s
t
o
de
r
i
ve
i
t
s
le
a
r
n
in
g
r
u
le
s
in
s
u
p
e
r
v
is
e
d
le
a
r
n
in
g
,
w
h
ic
h
is
s
u
p
e
r
io
r
to
c
o
n
v
e
n
tio
n
a
l
R
B
F
N
.
I
t
is
p
r
o
v
e
d
th
a
t
a
f
te
r
th
e
ke
r
ne
l
f
unc
t
i
on
i
s
gi
ve
n
t
he
w
e
i
ght
;
t
he
s
ha
pe
of
t
he
ke
r
ne
l
f
unc
t
i
on
i
s
e
l
l
i
pt
i
c
a
l
,
w
hi
c
h
i
s
m
or
e
r
e
a
s
ona
bl
e
th
a
n
th
e
s
h
a
p
e
o
f
th
e
c
ir
c
u
la
r
k
e
r
n
e
l
f
u
n
c
tio
n
w
h
ic
h
h
a
s
th
e
s
a
m
e
p
o
s
itio
n
a
s
a
ll
th
e
in
d
e
p
e
n
d
e
n
t
v
a
r
ia
b
le
s
o
f
th
e
c
o
n
v
e
n
tio
n
a
l
R
B
F
N
.
T
h
e
r
e
f
o
r
e
,
th
is
p
r
o
p
o
s
e
d
im
p
r
o
v
e
d
R
B
F
N
is
r
e
a
s
o
n
a
b
ly
h
a
s
s
u
p
e
r
io
r
f
ittin
g
d
e
g
r
e
e
th
a
n
c
o
n
v
e
n
tio
n
a
l R
B
F
N
.
Du
r
i
n
g
t
h
e
f
o
r
wa
r
d
p
r
o
p
a
g
a
t
i
o
n
o
f
p
r
o
p
o
s
e
d
n
e
t
wo
r
k
,
wh
i
c
h
is
a
ls
o
k
n
o
w
n
a
s
r
e
c
a
ll
s
ta
g
e
,
th
e
c
a
lc
u
la
tio
n
o
f
th
e
h
id
d
e
n
la
y
e
r
o
u
tp
u
t v
e
c
to
r
s
r
e
q
u
i
r
e
s
(2
)
a
n
d
(3
).
(
)
ex
p
kk
hn
e
t
=
−
(2
)
(
)
2
2
2
,
1
,
2
,
3
,
..
.,
ik
i
ik
i
k
k
Vx
C
ne
t
k
N
σ
−
==
∑
(3
)
wh
e
r
e
i
x
is
th
e
i
-
th
in
p
u
t
v
a
lu
e
;
k
h
is
th
e
o
u
tp
u
t
v
a
lu
e
o
f
th
e
k
-
th
G
a
u
s
s
ia
n
u
n
it
o
f
th
e
h
id
d
e
n
la
y
e
r
;
k
ne
t
is
th
e
n
e
t
v
a
lu
e
o
f
th
e
k
-
th
G
a
u
s
s
ia
n
u
n
it
o
f
th
e
h
id
d
e
n
la
y
e
r
;
ik
C
is
th
e
i
-
th
in
p
u
t
v
a
r
ia
b
le
v
a
lu
e
o
f
th
e
c
e
n
te
r
o
f
th
e
k
e
r
n
e
l
o
f
th
e
k
-
th
G
a
u
s
s
ia
n
u
n
it;
ik
V
is
th
e
w
e
ig
h
te
d
v
a
lu
e
o
f
th
e
i
-
th
in
p
u
t
v
a
r
ia
b
le
o
f
th
e
k
e
r
n
e
l
o
f
th
e
k
-
th
G
a
u
s
s
ia
n
u
n
it,
r
e
p
r
e
s
e
n
tin
g
th
e
im
p
o
r
ta
n
c
e
o
f
th
e
in
p
u
t
v
a
r
ia
b
le
;
k
σ
is
th
e
r
a
d
iu
s
o
f
th
e
k
e
r
n
e
l
o
f
th
e
k
-
th
Ga
u
s
s
i
a
n
u
n
i
t
.
Du
r
i
n
g
t
h
e
n
e
t
wo
r
k
t
r
a
i
n
i
n
g
p
r
o
c
e
s
s
,
k
σ
ma
y
a
d
j
u
s
t
e
d
t
o
0
,
c
au
s
i
n
g
t
h
e
d
en
o
m
i
n
at
o
r
t
o
b
e
0
.
In
o
rd
e
r
to
a
v
o
id
th
is
tr
o
u
b
le
,
(3
)
i
s
m
o
d
i
fi
e
d
t
o
t
h
e
fo
l
l
o
w
i
n
g
fo
rm
u
l
a
.
(
)
2
22
,
1
,
2
,
3
,
.
.
.
,
kk
i
k
i
i
k
i
ne
t
Q
V
x
C
k
N
=
−
=
∑
(4
)
wh
e
r
e
k
Q
is
th
e
r
e
c
ip
r
o
c
a
l o
f
th
e
r
a
d
iu
s
o
f
th
e
k
e
r
n
e
l o
f
th
e
k
-
th
Ga
u
s
s
i
a
n
u
n
i
t
.
In
c
a
l
c
u
l
a
t
i
n
g
t
h
e
r
es
p
o
n
d
ed
o
u
t
p
u
t
v
ect
o
r
,
(5
)
i
s
u
s
e
d
.
Evaluation Warning : The document was created with Spire.PDF for Python.
r
IS
S
N
:
2252
-
8938
IJ
-
AI
Vo
l
.
8
, N
o
.
1,
M
a
r
c
h
201
9
:
6
3
–
7
6
68
(
)
1
1e
x
p
j
j
y
ne
t
=
+
−
(5
)
jk
j
k
j
k
ne
t
W
h
θ
=
−
∑
(6
)
wh
e
r
e
j
ne
t
is
th
e
n
e
t
v
a
lu
e
o
f
th
e
j
-
th
u
n
it
o
f
th
e
o
u
tp
u
t
la
y
e
r
;
j
y
is
th
e
o
u
tp
u
t
v
a
lu
e
o
f
th
e
j
-
th
out
put
uni
t
;
kj
W
is
th
e
c
o
n
n
e
c
tio
n
w
e
ig
h
t
b
e
tw
e
e
n
th
e
k
-
th
u
n
it
o
f
th
e
h
id
d
e
n
la
y
e
r
a
n
d
th
e
j
-
th
u
n
it
o
f
th
e
out
put
l
a
ye
r
;
j
θ
is
th
e
th
r
e
s
h
o
ld
o
f
th
e
j
-
th
u
n
it o
f
th
e
o
u
tp
u
t la
ye
r
.
No
w
f
o
r
b
a
c
k
wa
r
d
p
r
o
p
a
g
a
t
i
o
n
,
wh
i
c
h
i
s
t
h
e
l
e
a
r
n
i
n
g
s
t
a
g
e
,
a
s
t
h
i
s
s
u
p
e
r
v
i
s
e
d
l
e
a
r
n
i
n
g
a
i
m
s
t
o
re
d
u
c
e
t
h
e
d
i
ffe
re
n
c
e
b
e
t
w
e
e
n
t
h
e
t
a
rg
e
t
o
u
t
p
u
t
v
a
l
u
e
o
f
t
h
e
n
e
t
w
o
rk
o
u
t
p
u
t
u
n
i
t
a
n
d
t
h
e
re
s
p
o
n
d
e
d
o
u
t
p
u
t
va
l
ue
,
t
he
qua
l
i
t
y
of
l
e
a
r
ni
ng
i
s
ge
ne
r
a
l
l
y
e
xpr
e
s
se
d
b
y
t
h
e
f
o
l
l
o
w
i
n
g
c
o
st
f
u
n
c
t
i
o
n
:
(
)
2
1
2
jj
Et
y
=
−
∑
(7
)
wh
e
r
e
j
t
is
th
e
ta
r
g
e
t
o
u
tp
u
t
v
a
lu
e
o
f
th
e
j
-
th
o
u
tp
u
t
u
n
it
o
f
th
e
tr
a
in
in
g
s
a
m
p
le
o
u
tp
u
t
la
y
e
r
;
j
y
is
th
e
r
e
s
p
o
n
d
e
d
o
u
tp
u
t v
a
lu
e
o
f
th
e
j
-
th
o
u
tp
u
t u
n
it o
f
th
e
tr
a
in
in
g
s
a
m
p
le
o
u
tp
u
t la
y
e
r
.
Si
n
c
e
t
h
e
c
o
s
t
f
u
n
c
t
i
o
n
i
s
a
f
u
n
c
t
i
o
n
o
f
t
h
e
r
e
s
p
o
n
d
e
d
o
u
t
p
u
t
v
a
l
u
e
,
a
n
d
t
h
e
r
e
s
p
o
n
d
e
d
o
u
t
p
u
t
v
a
l
u
e
is
a
f
u
n
c
tio
n
o
f
th
e
n
e
tw
o
r
k
c
o
n
n
e
c
tio
n
w
e
ig
h
t
v
a
lu
e
a
n
d
th
e
p
a
r
am
et
er
s
o
f
t
h
e
cen
t
er
an
d
r
ad
i
u
s
o
f
t
h
e
Ga
u
s
s
i
a
n
u
n
i
t
,
t
h
e
c
o
s
t
f
u
n
c
t
i
o
n
i
s
a
f
u
n
c
t
i
o
n
t
h
a
t
l
i
n
k
s
t
h
e
we
i
g
h
t
v
a
l
u
e
a
n
d
t
h
e
Ga
u
s
s
i
a
n
u
n
i
t
p
a
r
a
m
e
t
e
r
.
Th
e
r
e
f
o
r
e
,
a
d
j
u
s
t
i
n
g
t
h
e
l
i
n
k
w
e
i
g
h
t
v
a
l
u
e
a
n
d
t
h
e
G
a
u
s
s
i
a
n
u
n
i
t
p
a
r
a
m
e
t
e
r
c
a
n
c
h
a
n
g
e
t
h
e
s
i
z
e
o
f
t
h
e
c
o
s
t
f
unc
t
i
on.
I
n
or
de
r
t
o
m
i
ni
m
i
z
e
t
he
c
os
t
f
unc
t
i
on,
t
he
gr
a
di
e
nt
de
s
c
e
nt
m
e
t
hod
i
s
us
e
d
t
o
a
dj
us
t
t
he
ne
t
w
or
k
lin
k
w
e
ig
h
tin
g
v
a
lu
e
a
n
d
th
e
G
a
u
s
s
ia
n
u
n
it p
a
r
a
m
e
te
r
.
Th
e
f
o
l
l
o
w
i
n
g
i
s
d
i
v
i
d
e
d
i
n
t
o
t
w
o
p
a
r
t
s
t
o
d
e
r
i
v
e
t
h
e
c
o
r
r
e
c
t
i
o
n
f
o
r
m
u
l
a
.
1.
Th
e
w
e
i
g
h
t
e
d
v
a
l
u
e
o
f
t
h
e
l
i
n
k
b
e
t
w
e
e
n
t
h
e
h
i
d
d
e
n
l
a
y
e
r
a
n
d
t
h
e
o
u
t
p
u
t
l
a
y
e
r.
2.
Th
e
P
a
r
a
m
e
t
e
r
s
o
f
t
h
e
G
a
u
s
s
i
a
n
u
n
i
t
i
n
h
i
d
d
e
n
l
a
y
e
r
.
To
f
i
n
d
t
h
e
w
e
i
g
h
t
e
d
v
a
l
u
e
o
f
t
h
e
l
i
n
k
b
e
t
w
e
e
n
t
h
e
h
i
d
d
e
n
l
a
y
e
r
a
n
d
t
h
e
o
u
t
p
u
t
l
a
y
e
r
,
t
h
e
c
o
r
r
e
c
t
i
o
n
we
i
g
h
t
o
f
t
h
e
l
i
n
k
we
i
g
h
t
b
e
t
we
e
n
t
h
e
n
e
t
wo
rk
o
u
t
p
u
t
l
a
y
e
r
a
n
d
t
h
e
h
i
d
d
e
n
l
a
y
e
r
i
s
t
h
e
s
a
m
e
a
s
t
h
e
i
n
v
e
rs
e
gr
a
di
e
nt
de
s
c
e
nt
a
l
gor
i
t
hm
,
but
i
t
i
s
s
t
i
l
l
f
ul
l
y
de
r
i
ve
d
he
r
e
.
F
i
r
s
t
,
ge
t
t
he
pa
r
t
i
a
l
di
f
f
e
r
e
nt
i
a
l
l
a
w
:
jj
kj
kj
j
j
kj
yn
e
t
EE
W
Wy
n
e
t
W
ηη
∂∂
∂∂
Δ
=
−
=
−
⋅
⋅
∂∂
∂
∂
(8
)
Wh
e
r
e
,
(
)
(
)
2
1
2
jj
jj
jj
E
ty
ty
yy
∂∂
=
−
=
−
−
∂∂
⎛⎞
⎜⎟
⎝⎠
∑
(9
)
(
)
(
)
j
jj
jj
y
fn
e
t
f
ne
t
ne
t
∂
∂
ʹ
==
∂∂
(1
0
)
j
kj
k
j
k
k
kj
kj
ne
t
Wh
h
WW
θ
∂
∂
=
−
=
∂∂
⎛⎞
⎜⎟
⎝⎠
∑
(1
1
)
Pl
a
c
e
(9
)
-
(1
1
)
i
n
t
o
(8
),
w
e
o
b
t
a
i
n
e
d
,
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
IS
S
N
:
2252
-
8938
r
An
i
m
p
r
o
v
e
d
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
in
n
e
tw
o
r
k
s
w
e
ig
h
ts
a
d
ju
s
tm
e
n
t fo
r
…
(
Li
m
E
n
g
A
i
k
)
69
(
)
(
)
(
)
kj
j
j
j
k
j
j
j
k
Wt
y
f
h
t
y
f
h
ηη
ʹʹ
Δ
=
−
⋅
−
−
⋅
⋅
=
⋅
−
⋅
⋅
(1
2
)
De
f
i
n
e
,
(
)
(
)
(
)
j
jj
j
j
j
j
j
jj
j
y
EE
ty
f
ty
f
ne
t
y
ne
t
δ
∂
∂∂
ʹʹ
=
−
=
−
⋅
=
−
−
−
⋅
=
−
⋅
∂∂
∂
(1
3
)
Th
e
r
e
f
o
r
e
,
(1
2
)
c
a
n
re
w
ri
t
t
e
n
a
s
,
kj
j
k
Wh
ηδ
Δ
=
⋅
⋅
(1
4
)
He
n
c
e
,
th
e
c
o
r
r
e
c
tio
n
o
f
th
r
e
s
h
o
ld
v
a
lu
e
a
ls
o
c
a
n
r
e
w
r
itte
n
a
s
,
jj
j
E
θη
η
δ
θ
∂
Δ
=
−
=
−
∂
(1
5
)
Fo
r
Pa
r
a
m
e
t
e
r
s
o
f
t
h
e
h
i
d
d
e
n
l
a
y
e
r
G
a
u
s
s
i
a
n
u
n
i
t
,
t
h
e
c
e
n
t
r
a
l
v
a
l
u
e
c
o
r
r
e
c
t
i
o
n
o
f
t
h
e
k
e
r
n
e
l
o
f
t
h
e
Ga
u
s
s
i
a
n
u
n
i
t
c
a
n
o
b
t
a
i
n
e
d
b
y
t
h
e
p
a
r
t
i
a
l
d
i
f
f
e
r
e
n
t
i
a
l
l
a
w:
k
ik
ik
k
ik
ne
t
EE
C
Cn
e
t
C
ηη
∂
∂∂
Δ
=
−
=
−
⋅
∂∂
∂
(1
6
)
W
he
r
e
,
kk
k
j
kk
k
k
k
hn
e
t
EE
E
f
ne
t
h
ne
t
ne
t
h
∂∂
∂∂
∂
ʹ
=
⋅
=
⋅
∂∂
∂
∂
∂
⎛⎞
⎜⎟
⎝⎠
∑
(1
7
)
Du
e
t
o
(1
3
),
j
j
E
ne
t
δ
∂
=
−
∂
(1
8
)
Th
i
s
t
u
r
n
(
1
7
)
i
n
t
o
,
(
)
jk
j
k
jk
j
jj
k
E
Wf
W
f
ne
t
δδ
∂
ʹʹ
=
−
⋅
=
−
⋅
∂
⎛⎞
⎛
⎞
⎜⎟
⎜
⎟
⎝⎠
⎝
⎠
∑∑
(1
9
)
No
w
we
d
e
f
i
n
e
,
kj
k
j
k
j
Wf
δδ
ʹ
=
⋅
⋅
⎛⎞
⎜⎟
⎝⎠
∑
(2
0
)
Th
e
r
e
f
o
r
e
,
k
k
E
ne
t
δ
∂
=
−
∂
(2
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
r
IS
S
N
:
2252
-
8938
IJ
-
AI
Vo
l
.
8
, N
o
.
1,
M
a
r
c
h
201
9
:
6
3
–
7
6
70
(
)
(
)
(
)
2
22
22
2
k
k
ik
i
ik
k
ik
i
ik
ii
ik
ik
ne
t
QV
x
C
QV
x
C
CC
∂
∂
=
−
=
−
−
∂∂
⎛⎞
⎜⎟
⎝⎠
∑∑
(2
2
)
No
w,
we
s
u
b
s
t
i
t
u
t
e
(
2
1
)
a
n
d
(
2
2
)
i
n
t
o
(
1
6
)
,
we
o
b
t
a
i
n
e
d
,
(
)
(
)
(
)
(
)
22
2
2
22
ik
k
k
ik
i
ik
k
k
ik
i
ik
i
CQ
V
x
C
Q
V
x
C
ηδ
η
δ
Δ
=
−
⋅
−
⋅
−
−
=
−
−
∑
(2
3
)
Si
m
i
l
a
r
l
y
,
t
h
e
m
u
t
u
a
l
c
o
r
r
e
c
t
i
o
n
o
f
t
h
e
r
a
d
i
u
s
o
f
t
h
e
k
e
r
n
e
l
o
f
t
h
e
G
a
u
s
s
i
a
n
u
n
i
t
c
a
n
o
b
t
a
i
n
b
y
a
pa
r
t
i
a
l
di
f
f
e
r
e
nt
i
a
l
l
a
w
:
k
k
kk
k
ne
t
EE
Q
Qn
e
t
Q
ηη
∂
∂∂
Δ
=
−
=
−
⋅
∂∂
∂
(2
4
)
wh
e
r
e
,
(
)
(
)
22
22
2
2
k
k
ik
i
ik
k
ik
i
ik
ii
kk
ne
t
QV
x
C
Q
V
x
C
QQ
∂
∂
=
−
=
−
∂∂
⎛⎞
⎜⎟
⎝⎠
∑∑
(2
5
)
By
s
u
b
s
t
i
t
u
t
i
n
g
(
2
1
)
a
n
d
(
2
5
)
i
n
t
o
(
2
4
)
,
w
e
o
b
t
a
i
n
e
d
,
(
)
(
)
(
)
22
22
22
kk
k
i
k
i
i
k
k
k
i
k
i
i
k
ii
QQ
V
x
C
Q
V
x
C
ηδ
η
δ
Δ
=
−
−
⋅
−
=
−
⎛⎞
⎜⎟
⎝⎠
∑∑
(2
6
)
Si
m
i
l
a
r
l
y
,
t
h
e
w
e
i
g
h
t
o
f
t
h
e
i
n
p
u
t
v
a
r
i
a
b
l
e
w
e
i
g
h
t
(
s
h
a
p
e
pa
r
a
m
e
t
e
r
)
of
t
he
G
a
us
s
i
a
n
uni
t
c
a
n
obt
a
i
ne
d
by
t
he
pa
r
t
i
a
l
di
f
f
e
r
e
nt
i
a
l
l
a
w
:
k
k
kk
k
ne
t
EE
V
Vn
e
t
V
ηη
∂
∂∂
Δ
=
−
=
−
⋅
∂∂
∂
(2
7
)
Wh
e
r
e
,
(
)
(
)
22
22
2
2
k
k
ik
i
ik
k
ik
i
ik
ii
kk
ne
t
QV
x
C
QV
x
C
VV
∂
∂
=
−
=
−
∂∂
⎛⎞
⎜⎟
⎝⎠
∑∑
(2
8
)
By
s
u
b
s
t
i
t
u
t
i
n
g
(2
1
)
a
n
d
(2
8
)
i
n
t
o
(2
7
),
w
e
o
b
t
a
i
n
e
d
,
(
)
(
)
(
)
22
22
22
kk
k
i
k
i
i
k
k
k
i
k
i
i
k
ii
VQ
V
x
C
Q
V
x
C
ηδ
η
δ
Δ
=
−
−
⋅
−
=
−
⎛⎞
⎜⎟
⎝⎠
∑∑
(2
9
)
Th
e
f
i
v
e
(
1
4
)
,(
1
5
)
, (
2
3
)
, (
2
6
)
a
n
d
(
2
9
)
a
r
e
th
e
le
a
r
n
in
g
r
u
le
s
o
f
th
e
im
p
r
o
v
e
d
R
B
F
n
e
u
r
a
l n
e
tw
o
r
k
.
Cl
e
a
r
l
y
,
f
r
o
m
a
b
o
v
e
d
e
r
i
v
a
t
i
o
n
e
q
u
a
t
i
o
n
s
,
t
h
e
r
e
i
s
o
n
e
m
o
r
e
s
h
a
p
e
p
a
r
a
m
e
t
e
r
i
n
t
h
e
t
u
n
a
b
l
e
pa
r
a
m
e
t
e
r
t
ha
n
t
he
c
onve
nt
i
ona
l
R
B
F
N
.
T
he
s
upe
r
vi
s
e
d
l
e
a
r
ni
ng
r
ul
e
s
of
a
l
l
pa
r
a
me
t
e
r
s
a
r
e
u
n
i
f
o
r
ml
y
de
r
i
ve
d
by
us
i
ng
t
he
pr
i
nc
i
pl
e
of
s
um
s
qua
r
e
s
e
r
r
or
m
i
ni
m
i
z
a
t
i
on,
i
nc
l
udi
ng
t
he
c
e
nt
e
r
,
r
a
di
us
,
a
nd
s
ha
pe
pa
r
a
m
e
t
e
r
s
of
t
he
ke
r
ne
l
f
unc
t
i
on,
t
he
w
e
i
ght
i
ng
va
l
ue
of
t
he
ne
t
w
or
ks
be
t
w
e
e
n
t
he
hi
dde
n
uni
t
a
nd
t
he
out
put
uni
t
,
a
nd
t
he
t
hr
e
s
hol
d
of
t
he
out
put
uni
t
.
Th
e
R
B
F
N
i
s
o
f
t
e
n
u
s
e
d
f
o
r
p
r
e
d
i
c
t
i
o
n
a
n
d
c
l
a
s
s
i
f
i
c
a
t
i
o
n
p
r
o
b
l
e
m
s
,
b
u
t
i
t
h
a
s
t
h
e
d
i
s
a
d
v
a
n
t
a
g
e
t
h
a
t
al
l
i
n
d
ep
en
d
en
t
v
ar
i
ab
l
es
h
av
e
eq
u
al
s
t
at
u
s
,
w
i
t
h
t
h
e
s
h
ap
e
o
f
t
h
e
k
er
n
el
f
u
n
ct
i
o
n
i
s
ci
r
cu
l
ar
an
d
t
h
i
s
i
n
f
l
u
en
ce
out
c
om
e
t
he
ne
t
w
o
rk
s
.
S
i
n
c
e
t
h
e
w
e
i
g
h
t
s
o
f
e
a
c
h
i
n
p
u
t
a
re
d
i
ffe
re
n
t
,
s
o
t
h
e
s
h
a
p
e
o
f
t
h
e
k
e
rn
e
l
fu
n
c
t
i
o
n
sh
o
u
l
d
b
e
e
l
l
i
p
t
i
c
a
l
.
T
o
o
v
e
r
c
o
m
e
t
h
i
s
sh
o
r
t
c
o
m
i
n
g
,
a
n
i
m
p
r
o
v
e
d
R
B
F
N
w
i
t
h
k
e
r
n
e
l
sh
a
p
e
p
a
r
a
m
e
t
e
r
s
i
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
AI
IS
S
N
:
2252
-
8938
r
An
i
m
p
r
o
v
e
d
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
in
n
e
tw
o
r
k
s
w
e
ig
h
ts
a
d
ju
s
tm
e
n
t fo
r
…
(
Li
m
E
n
g
A
i
k
)
71
pr
opos
e
d,
a
nd
i
t
s
l
e
a
r
ni
ng
r
ul
e
s
a
r
e
de
r
i
ve
d
by
s
upe
r
vi
s
e
d
l
e
a
r
ni
ng
.
T
h
is
a
r
c
h
ite
c
tu
r
e
is
s
u
p
e
r
io
r
to
tr
a
d
itio
n
a
l
ra
d
i
a
l
b
a
s
i
s
fu
n
c
t
i
o
n
n
e
t
w
o
rk
s
.
Cl
e
a
r
l
y
,
f
r
o
m
F
i
g
u
r
e
2
t
h
a
t
b
a
s
e
d
o
n
t
h
e
RBF
N
,
t
h
e
p
r
o
p
o
s
e
d
I
RBF
N
h
a
s
t
h
e
w
e
i
g
h
t
o
f
V
a
d
d
e
d
be
t
w
e
e
n
t
he
i
nput
l
a
ye
r
a
nd
t
he
hi
dde
n
l
a
ye
r
t
o
i
ndi
c
a
t
e
t
he
i
nf
l
ue
nc
e
of
di
f
f
e
r
e
nt
i
nput
l
a
ye
r
uni
t
s
on
t
he
hi
dde
n
l
a
ye
r
uni
t
,
w
hi
c
h
l
a
t
e
r
us
e
f
or
c
a
l
c
ul
a
t
i
ng
t
he
w
e
i
ght
s
of
hi
dde
n
l
a
ye
r
t
o
out
put
l
a
ye
r
.
Fi
g
u
r
e
2
.
Th
e
a
rc
h
i
t
e
c
t
u
re
o
f
IR
B
F
N
4.
RE
S
UL
T
S
A
ND
DI
S
CUS
S
I
O
N
Th
e
p
r
o
p
o
s
e
d
I
R
B
F
N
w
a
s
t
e
s
t
e
d
u
s
i
n
g
4
n
o
n
l
i
n
e
a
r
f
u
n
c
t
i
o
n
f
r
o
m
l
i
t
e
r
a
t
u
r
e
s
,
w
h
i
c
h
a
r
e
S
a
n
t
n
e
r
e
t
a
l
.
[6
9
]
fu
n
c
tio
n
g
iv
e
n
in
(3
0
),
L
i
m
e
t
a
l
.
[7
0
]
fu
n
c
t
i
o
n
i
n
(3
1
)
,
D
e
tte
a
n
d
P
e
p
e
ly
s
h
e
v
[7
1
]
fu
n
c
t
i
o
n
i
n
(3
2
),
an
d
F
r
i
ed
m
an
[7
2
]
f
unc
t
i
on
i
n
(
33)
.
F
or
a
l
l
t
he
s
e
4
f
unc
t
i
ons
,
t
he
t
r
a
i
ni
ng
s
e
t
f
or
R
B
F
N
c
ons
i
s
t
s
of
400
s
e
t
s
of
ra
n
d
o
m
g
e
n
e
ra
t
e
d
d
a
t
a
p
o
i
n
t
s
a
n
d
t
e
s
t
s
e
t
c
o
m
p
ri
s
e
s
4
0
0
s
e
t
s
o
f
ra
n
d
o
m
g
e
n
e
ra
t
e
d
d
a
t
a
p
o
i
n
t
s
,
bot
h
i
n
r
a
nge
of
[
0,
1]
.
[
]
(
)
e
x
p
(
1
.
4
)
c
o
s
(3
.
5
)
,
0
,
1
fx
x
x
x
π
=
−
∈
(3
0
)
(
)
(
)
(
)
(
)
(
)
[
]
[
]
11
2
1
3
0
5
s
in
5
4
e
x
p
5
1
0
0
,
0
,
1
,
1
,
2
.
6
i
fx
x
x
x
x
i
=+
+
−
−
∈
∀
=
(3
1
)
(
)
(
)
(
)
(
)
[
]
2
22
2
12
2
2
3
3
42
8
8
3
4
1
6
1
2
1
,
0
,
1
,
1
,
2
,
3
.
i
fx
x
x
x
x
x
x
x
i
=
−
+
−
+
−
+
+
−
∈
∀
=
(3
2
)
(
)
(
)
(
)
[
]
2
12
3
4
5
10
s
i
n
20
0.
5
10
5
,
0,
1
,
1
,
2,
3
,
4,
5.
i
fx
x
x
x
x
x
x
i
π
=+
−
+
+
∈
∀
=
(3
3
)
IR
B
F
N
w
a
s
a
l
s
o
t
e
s
t
e
d
o
n
4
re
a
l
-
wo
r
l
d
d
a
t
a
s
e
t
s
f
o
r
i
t
s
p
e
r
f
o
r
m
a
n
c
e
.
T
h
e
B
i
o
c
h
e
m
i
c
a
l
Ox
y
g
e
n
De
m
a
n
d
(
B
OD)
c
o
n
c
e
n
t
r
a
t
i
o
n
d
a
t
a
s
e
t
,
p
h
y
t
o
p
l
a
n
k
t
o
n
g
r
o
wt
h
r
a
t
e
a
n
d
d
e
a
t
h
r
a
t
e
d
a
t
a
s
e
t
a
n
d
a
i
r
p
o
l
l
u
t
a
n
t
da
t
a
s
e
t
w
e
r
e
obt
a
i
ne
d
f
r
om
A
i
k
a
nd
Z
a
i
nuddi
n
[7
3
]
.
A
n
o
th
e
r
w
h
ic
h
is
th
e
d
a
ta
s
e
t
f
r
o
m
f
o
r
e
x
f
o
r
E
U
R
U
S
D
pa
i
r
s
i
s
c
ol
l
e
c
t
e
d
f
r
om
X
M
M
e
t
a
t
r
a
de
r
4
da
t
a
ba
s
e
[1
0
]
.
T
h
e
B
O
D
d
a
ta
s
e
t
a
n
d
p
h
y
to
p
la
n
k
to
n
d
a
ta
s
e
t
b
o
th
co
n
s
i
s
t
s
o
f
1
0
0
s
et
s
o
f
d
at
a
an
d
t
h
e
t
es
t
s
et
co
m
p
r
i
s
es
1
0
0
s
et
s
o
f
d
at
a.
M
ean
w
h
i
l
e,
f
o
r
ai
r
p
o
l
l
u
t
an
t
d
at
as
et
,
th
e
tr
a
in
in
g
s
e
t
h
a
s
4
8
0
s
e
ts
o
f
d
a
ta
a
n
d
th
e
te
s
t
s
e
t
h
a
s
7
2
se
t
s
o
f
d
a
t
a
,
w
h
i
c
h
b
o
t
h
w
e
r
e
t
a
k
e
n
f
r
o
m
h
o
u
r
l
y
a
i
r
x
1
x
2
x
3
x
N
.
.
.
.
.
.
In
p
u
t l
a
y
e
r
Hi
d
d
e
n
L
a
y
e
r
o
f
RB
F
Ou
t
p
u
t
L
a
y
e
r
ψ
ψ
ψ
ψ
w
N
.
.
.
w
2
w
3
w
1
V
1
1
V
1
2
V
1
3
V
1
4
V
21
V
22
V
23
V
24
V
31
V
32
V
33
V
34
V
N1
V
N2
V
N3
V
N4
Evaluation Warning : The document was created with Spire.PDF for Python.
r
IS
S
N
:
2252
-
8938
IJ
-
AI
Vo
l
.
8
, N
o
.
1,
M
a
r
c
h
201
9
:
6
3
–
7
6
72
da
t
a
.
F
or
E
U
R
U
S
D
pa
i
r
s
,
t
he
t
r
a
i
ni
ng
s
e
t
c
ons
i
s
t
s
of
519
s
e
t
s
of
d
at
a
t
ak
en
f
r
o
m
y
ear
2
0
1
6
t
o
en
d
of
ye
a
r
2017.
T
he
t
e
s
t
s
e
t
c
ons
i
s
t
s
of
155
s
e
t
s
of
da
t
a
t
a
ke
n
f
r
om
e
a
r
l
y
ye
a
r
2018
t
o
A
ugus
t
2018.
Th
e
e
x
p
er
i
m
en
t
w
as
i
m
p
l
em
en
t
ed
b
y
u
s
i
n
g
t
h
e
n
ew
r
b
f
u
n
ct
i
o
n
b
ecau
s
e
i
t
r
ep
r
es
en
t
s
t
h
e
g
en
er
al
fo
rm
o
f
a
n
R
B
F
n
e
t
w
o
rk
.
F
u
rt
h
e
rm
o
re
,
t
h
e
p
ro
p
o
s
e
d
c
l
u
s
t
e
ri
n
g
m
e
t
h
o
d
h
a
s
b
e
e
n
i
m
p
l
e
m
e
n
t
e
d
b
y
u
s
i
n
g
MA
T
L
A
B
’
s
f
u
n
c
t
i
o
n
.
G
a
u
s
s
i
a
n
b
a
s
i
s
f
u
n
c
t
i
o
n
h
a
s
b
e
e
n
u
s
e
d
f
o
r
b
o
t
h
n
e
t
w
o
r
ks
w
i
t
h
ot
he
r
pa
r
a
m
e
t
e
r
s
s
uc
h
a
s
sp
r
e
a
d
w
a
s
se
t
t
o
d
e
f
a
u
l
t
v
a
l
u
e
,
so
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
t
h
e
p
r
o
p
o
se
d
n
e
t
w
o
r
k
c
a
n
e
v
a
l
u
a
t
e
d
e
f
f
e
c
t
i
v
e
l
y
[
8
]
.
Pe
r
f
o
r
m
a
n
c
e
o
f
s
t
a
n
d
a
r
d
R
B
FN
a
n
d
I
R
B
FN
i
n
t
h
i
s
e
x
p
e
r
i
m
e
n
t
h
a
s
b
e
e
n
m
e
a
s
u
r
e
d
b
y
c
o
m
p
a
r
i
n
g
t
h
e
co
m
p
u
t
at
i
o
n
t
i
m
e
t
ak
en
fo
r
t
ra
i
n
i
n
g
w
i
t
h
n
u
m
b
e
r
o
f
i
t
e
ra
t
i
o
n
t
a
k
e
n
fo
r
c
o
n
v
e
rg
e
n
c
e
a
n
d
t
h
e
R
o
o
t
M
e
a
n
Sq
u
a
r
e
d
E
r
r
o
r
(
R
M
SE
)
t
o
m
e
a
s
u
r
e
h
o
w
w
e
l
l
b
o
t
h
n
e
t
w
o
r
k
s
a
p
p
r
o
x
i
m
a
t
e
s
t
h
e
c
h
o
se
n
f
u
n
c
t
i
o
n
s
an
d
i
t
i
s
g
i
v
en
b
y.
(
)
2
1
1
n
ii
i
RM
S
E
O
P
n
=
=
−
∑
wh
e
r
e
n
i
s
t
h
e
n
u
m
b
e
r
o
f
p
r
e
d
i
c
t
e
d
r
e
s
p
o
n
d
e
r
;
i
O
is
th
e
ta
r
g
e
t
v
a
lu
e
f
o
r
tim
e
-
st
e
p
i
,
a
n
d
i
P
is
th
e
p
r
e
d
ic
te
d
va
l
ue
of
t
he
m
ode
l
a
t
t
i
m
e
-
st
e
p
i
.
Th
e
n
u
m
b
e
r
o
f
c
e
n
t
e
r
s
f
o
r
R
B
F
N
i
s
f
i
x
e
d
t
o
1
0
c
e
n
t
e
r
s
f
o
r
a
l
l
8
da
t
a
s
e
t
s
t
r
a
i
ni
ng.
F
or
a
i
r
pol
l
ut
a
nt
pr
obl
e
m
,
t
he
pol
l
ut
a
nt
m
oni
t
or
e
d
i
nc
l
ude
s
c
a
r
bon
m
onoxi
de
,
ni
t
r
i
c
oxi
de
,
ni
t
r
oge
n
di
oxi
de
,
oz
one
,
a
nd
oxi
de
s
of
ni
t
r
oge
n.
F
or
e
xpe
r
i
m
e
nt
a
l
pur
pos
e
s
,
hour
l
y
upda
t
e
d
a
i
r
qua
l
i
t
y
da
t
a
obt
a
i
ne
d
f
r
om
A
i
k
a
nd
Z
a
i
nuddi
n
[7
3
]
ha
s
be
e
n
us
e
d
t
o
pr
e
di
c
t
t
he
t
r
e
nd
of
i
nt
e
r
e
s
t
e
d
pol
l
ut
a
nt
s
f
or
N
i
t
r
i
c
O
xi
de
,
N
i
t
r
oge
n
D
i
oxi
de
a
nd
O
xi
de
s
of
N
i
t
r
oge
n.
W
hi
l
e
f
or
P
hyt
opl
a
nkt
on
pr
obl
e
m
,
gr
ow
th
r
a
te
a
n
d
d
e
a
th
r
a
te
h
a
v
e
b
e
e
n
u
s
e
d
a
s
th
e
in
te
r
e
s
te
d
va
l
ue
s
.
A
s
f
or
t
he
B
O
D
pr
obl
e
m
,
t
he
B
O
D
c
onc
e
nt
r
a
t
i
on
ha
s
be
e
n
t
a
ke
n
a
s
t
he
i
nt
e
r
e
s
t
e
d
va
l
ue
.
F
i
na
l
l
y,
t
he
fo
re
x
E
U
R
U
S
D
d
a
t
a
s
e
t
c
o
n
s
i
s
t
s
o
f
t
h
re
e
v
a
ri
a
b
l
e
s
i
s
t
a
k
e
n
c
o
n
s
i
d
e
re
d
fo
r
t
ra
i
n
i
n
g
,
w
h
i
c
h
a
re
th
e
d
a
ily
hi
ghe
s
t
pr
i
c
e
,
da
i
l
y
l
ow
e
s
t
pr
i
c
e
a
nd
ope
n
pr
i
c
e
,
w
hi
l
e
t
he
c
l
os
e
pr
i
c
e
i
s
us
e
d
f
or
pr
e
di
c
t
i
on.
Ta
b
l
e
1
.
Pe
r
f
o
r
m
a
n
c
e
o
f
I
RBF
N
a
n
d
S
t
a
n
d
a
r
d
RBF
N
pr
e
di
c
t
i
on
r
e
s
ul
t
s
f
or
da
t
a
s
e
t
s
Da
t
a
s
e
t
St
a
n
d
a
r
d
R
B
FN
I
RBF
N
Av
e
r
a
g
e
o
f
R
M
S
E
St
a
n
d
a
r
d
D
e
v
i
a
t
i
o
n
o
f
RM
S
E
Av
e
r
a
g
e
o
f
R
M
S
E
St
a
n
d
a
r
d
D
e
v
i
a
t
i
o
n
o
f
RM
S
E
Sa
n
t
n
e
r
0.
160980
0.
132510
0.
095935
0.
021740
Li
m
0.
151190
0.
053404
0.
094039
0.
011824
De
t
t
e
1.
938260
1.
318204
0.
952562
0.
175806
Fr
i
e
d
m
a
n
1.
333590
0.
700492
0.
162113
0.
058887
BO
D
0.
000252
3.
87e
-
07
0.
000247
4.
29e
-
05
Ph
y
t
o
p
l
a
n
k
t
o
n
0.
004980
0.
000933
0.
004035
0.
000800
Ai
r
P
o
l
l
u
t
a
n
t
4.
203630
2.
860265
0.
371875
0.
109962
EU
R
U
SD
0.
031649
0.
004812
0.
028432
0.
004942
Ta
b
l
e
2
.
Pe
r
c
e
n
t
a
g
e
o
f
I
m
p
r
o
v
e
m
e
n
t
f
o
r
I
RBF
N
o
v
e
r
S
t
a
n
d
a
r
d
RBF
N
b
y
RM
S
E
Da
t
a
s
e
t
Pe
r
c
e
n
t
a
g
e
o
f
Im
p
ro
v
e
m
e
n
t
(%
)
Sa
n
t
n
e
r
40.
41
Li
m
37.
80
De
t
t
e
50.
85
Fr
i
e
d
m
a
n
87.
84
BO
D
1.
80
Ph
y
t
o
p
l
a
n
k
t
o
n
18.
98
Ai
r
P
o
l
l
u
t
a
n
t
91.
15
EU
R
U
SD
10.
16
Re
s
u
l
t
s
f
r
o
m
T
a
b
l
e
1
s
h
o
w
s
t
h
a
t
I
RBF
N
n
e
t
w
o
r
k
s
o
u
t
p
e
r
f
o
r
m
s
t
a
n
d
a
r
d
RBF
N
i
n
a
v
e
r
a
g
e
RM
S
E
a
n
d
st
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
o
f
RM
S
E
.
A
l
l
r
e
s
u
l
t
s
o
f
a
v
e
r
a
g
e
RM
S
E
i
n
T
a
b
l
e
1
a
r
e
c
a
l
c
u
l
a
t
e
d
u
s
i
n
g
t
e
n
t
i
m
e
s
r
u
n
o
f
each
n
et
w
o
r
k
s
.
F
r
o
m
T
ab
l
e
1
,
I
R
B
F
N
n
et
w
o
r
k
s
u
r
p
as
s
es
t
h
e
s
t
an
d
ar
d
R
B
F
i
n
accu
r
acy
an
d
n
et
w
o
r
k
ar
ch
i
t
ect
u
r
e
b
y
u
s
i
n
g
t
r
ai
n
i
n
g
s
et
w
h
i
ch
co
n
s
i
s
t
s
o
n
l
y
8
1
.
8
%
,
7
1
.
8
%
,
6
8
.
5
%
,
an
d
65.
8%
of
t
ot
a
l
da
t
a
s
e
t
s
i
z
e
fo
r
S
a
n
t
n
e
r
d
a
t
a
s
e
t
,
L
i
m
d
a
t
a
s
e
t
,
D
e
t
t
e
d
a
t
a
s
e
t
,
a
n
d
F
ri
e
d
m
a
n
d
a
t
a
s
e
t
,
re
s
p
e
c
t
i
v
e
l
y
.
W
h
i
l
e
IR
B
F
N
n
e
t
w
o
rk
tr
a
in
in
g
f
o
r
r
e
a
l
-
wo
r
l
d
d
a
t
a
s
e
t
i
n
v
o
l
v
e
s
t
h
e
B
OD
d
a
t
a
s
e
t
,
P
h
y
t
o
p
l
a
n
k
t
o
n
d
a
t
a
s
e
t
,
Ai
r
p
o
l
l
u
t
a
n
t
d
a
t
a
s
e
t
a
n
d
fo
re
x
E
U
R
U
S
D
d
a
ta
s
e
t
u
s
e
d
o
n
ly
8
1
.8
%
,
7
1
.8
%
,
6
8
.5
%
a
n
d
6
6
.8
%
,
r
e
s
p
e
c
tiv
e
ly
.
T
h
is
m
e
a
n
s
th
a
t,
it
is
pos
s
i
bl
e
t
o
s
ui
t
a
bl
e
num
be
r
of
da
t
a
s
e
t
s
uc
h
t
ha
t
,
i
t
w
i
l
l
pr
ovi
de
a
ne
t
w
or
k
w
i
t
h
r
e
duc
e
d
c
om
pl
e
xi
t
y,
f
a
s
t
e
r
tr
a
in
in
g
tim
e
a
n
d
im
p
r
o
v
e
d
a
c
c
u
r
a
c
y
.
Ta
b
l
e
2
s
h
o
w
s
t
h
e
r
e
s
u
l
t
s
o
f
t
h
e
p
e
rc
e
n
t
a
g
e
o
f
i
m
p
ro
v
e
m
e
n
t
o
f
R
M
S
E
Evaluation Warning : The document was created with Spire.PDF for Python.