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
.
5
,
Octo
b
e
r
2
0
1
8
,
p
p
.
3
3
4
1
~3
3
4
8
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v8
i
5
.
p
p
3
3
4
1
-
3348
3341
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
Initia
l O
p
ti
m
a
l
P
a
ra
m
ete
rs o
f
A
rti
ficial Neu
ra
l Ne
t
w
o
rk
and
Suppo
rt
Vect
o
r
R
eg
ress
io
n
E
dy
F
ra
din
a
t
a
1
,
Sa
k
e
s
un
S
u
t
hu
mm
a
no
n
2
,
Wa
n
na
ra
t
Su
ntia
m
o
rnt
ut
3
1
In
d
u
strial
En
g
i
n
e
e
rin
g
De
p
a
rtm
e
n
t,
S
y
iah
Ku
a
la Un
iv
e
rsity
,
Ba
n
d
a
Ac
e
h
,
In
d
o
n
e
sia
1,
2
In
d
u
strial
E
n
g
in
e
e
rin
g
De
p
a
rte
m
e
n
t,
P
rin
c
e
o
f
S
o
n
g
k
la Un
iv
e
rsity
,
Ha
t
y
a
i,
T
h
a
il
a
n
d
3
Co
m
p
u
ter E
n
g
in
e
e
rin
g
De
p
a
rte
m
e
n
t,
P
rin
c
e
o
f
S
o
n
g
k
la Un
iv
e
rsity
,
Ha
t
y
a
i,
T
h
a
il
a
n
d
Art
ic
le
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
1
6
,
2
0
1
8
R
ev
i
s
ed
Ma
r
3
,
2
0
1
8
A
cc
ep
ted
Ma
r
2
3
,
2
0
1
8
T
h
is
p
a
p
e
r
p
re
se
n
ts
a
rc
h
it
e
c
tu
re
o
f
b
a
c
k
p
ro
p
a
g
a
ti
o
n
A
rti
f
ici
a
l
Ne
u
ra
l
Ne
tw
o
rk
(
A
NN
)
a
n
d
S
u
p
p
o
r
t
V
e
c
to
r
Re
g
re
ss
io
n
(S
V
R)
m
o
d
e
ls
in
su
p
e
rv
ise
d
lea
rn
in
g
p
r
o
c
e
ss
f
o
r
c
e
m
e
n
t
d
e
m
a
n
d
d
a
tas
e
t.
T
h
is
st
u
d
y
a
i
m
s
to
i
d
e
n
ti
f
y
th
e
e
ffe
c
ti
v
e
n
e
ss
o
f
e
a
c
h
p
a
ra
m
e
ter
o
f
m
e
a
n
sq
u
a
re
e
rro
r
(M
S
E)
i
n
d
ica
to
rs
f
o
r
ti
m
e
se
ries
d
a
ta
s
e
t.
T
h
e
stu
d
y
v
a
ri
e
s d
iff
e
re
n
t
ra
n
d
o
m
sa
m
p
le i
n
e
a
c
h
d
e
m
a
n
d
p
a
ra
m
e
ter
in
th
e
n
e
tw
o
rk
o
f
A
N
N
a
n
d
su
p
p
o
rt
v
e
c
to
r
f
u
n
c
ti
o
n
a
s
we
ll
.
T
h
e
v
a
riatio
n
s
o
f
p
e
rc
e
n
t
d
a
tas
e
ts
f
ro
m
a
c
ti
v
a
ti
o
n
f
u
n
c
ti
o
n
,
lea
rn
in
g
ra
te
o
f
sig
m
o
id
a
n
d
p
u
re
li
n
,
h
id
d
e
n
lay
e
r,
n
e
u
r
o
n
s,
a
n
d
trai
n
in
g
f
u
n
c
ti
o
n
sh
o
u
l
d
b
e
a
p
p
li
e
d
f
o
r
A
NN
.
F
u
rth
e
rm
o
re
,
S
V
R
is
v
a
ried
in
k
e
rn
e
l
f
u
n
c
ti
o
n
,
l
o
st
f
u
n
c
ti
o
n
a
n
d
in
se
n
si
ti
v
it
y
to
o
b
tain
th
e
b
e
st
re
su
lt
f
ro
m
it
s
si
m
u
latio
n
.
T
h
e
b
e
st
re
su
lt
s
o
f
th
is
st
u
d
y
f
o
r
A
N
N
a
c
ti
v
a
ti
o
n
f
u
n
c
ti
o
n
is
S
ig
m
o
id
.
T
h
e
a
m
o
u
n
t
o
f
d
a
ta
in
p
u
t
is
1
0
0
%
o
r
9
6
o
f
d
a
ta,
1
5
0
lea
rn
i
n
g
ra
tes
,
o
n
e
h
id
d
e
n
la
y
e
r,
tri
n
lm
train
in
g
f
u
n
c
ti
o
n
,
1
5
n
e
u
ro
n
s
a
n
d
3
to
tal
lay
e
rs.
T
h
e
b
e
st
re
su
lt
s
f
o
r
S
V
R
a
re
six
v
a
riab
les
th
a
t
ru
n
i
n
o
p
ti
m
a
l
c
o
n
d
it
i
o
n
,
k
e
rn
e
l
f
u
n
c
ti
o
n
is
li
n
e
a
r,
lo
ss
f
u
n
c
ti
o
n
is
౬
-
in
se
n
si
ti
v
e
,
a
n
d
i
n
se
n
siti
v
it
y
w
a
s
1
.
T
h
e
b
e
tt
e
r
re
su
lt
s
f
o
r
b
o
th
m
e
th
o
d
s
a
re
six
v
a
riab
les
.
T
h
e
c
o
n
tri
b
u
t
io
n
o
f
th
is
stu
d
y
is
to
o
b
tai
n
th
e
o
p
ti
m
a
l
p
a
ra
m
e
ter
s f
o
r
sp
e
c
if
ic v
a
riab
les
o
f
A
NN
a
n
d
S
V
R.
K
ey
w
o
r
d
:
A
N
N
MSE
Op
ti
m
izatio
n
Su
p
er
v
i
s
ed
SVR
Co
p
y
rig
h
t
©
201
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
:
E
d
y
Fra
d
i
n
ata,
Dep
ar
te
m
en
t o
f
I
n
d
u
s
tr
ial
E
n
g
in
ee
r
in
g
,
S
y
ia
h
K
u
ala
Un
iv
er
s
it
y
,
T
eu
k
u
N
y
ak
A
r
ie
f
,
Dar
u
s
s
ala
m
,
B
an
d
a
A
ce
h
,
2
3
1
1
1
,
I
n
d
o
n
esia.
E
m
ail:
ed
i
n
ata6
9
@
g
m
ai
l.c
o
m
1.
I
NT
RO
D
UCT
I
O
N
A
r
ti
f
icial
Ne
u
r
al
Net
w
o
r
k
(
ANN)
is
a
s
tr
u
ctu
r
e
o
f
lear
n
in
g
s
y
s
te
m
s
w
h
er
e
it
is
in
s
p
ir
ed
b
y
liv
i
n
g
o
r
g
an
is
m
s
,
e
s
p
ec
iall
y
to
a
h
u
m
an
s
y
s
te
m
.
I
t
co
n
s
is
t
s
o
f
a
v
er
y
co
m
p
lex
n
et
w
o
r
k
th
a
t
is
e
q
u
i
p
p
ed
w
ith
s
o
m
e
n
eu
r
o
n
s
w
h
ich
ar
e
in
ter
co
n
n
e
c
ted
e
ac
h
o
th
e
r
,
th
ese
n
e
u
r
o
n
s
w
o
r
k
to
r
e
m
e
m
b
er
,
to
ca
lcu
lat
e,
to
g
en
er
alize
,
to
ad
ap
t,
to
g
et
lo
w
d
y
n
a
m
i
s
m
an
d
h
as
h
i
g
h
f
le
x
ib
ilit
y
.
SV
R
is
a
m
et
h
o
d
to
co
n
tr
ib
u
te
th
e
s
o
lu
tio
n
b
y
s
m
all
s
u
b
s
et
f
r
o
m
t
h
e
tr
ai
n
in
g
p
o
in
ts
w
h
er
e
p
r
o
d
u
ce
t
h
e
en
o
r
m
o
u
s
co
m
p
u
tatio
n
al
ad
v
a
n
ta
g
es.
T
h
e
e
-
i
n
s
e
n
s
i
tiv
e
lo
s
s
f
u
n
ctio
n
p
r
eten
d
s
t
h
e
ex
i
s
ten
c
e
o
f
th
e
g
lo
b
al
m
i
n
i
m
u
m
s
o
l
u
t
io
n
an
d
th
e
o
p
ti
m
izatio
n
b
o
u
n
d
[
1
]
.
Su
p
p
o
r
t
Ve
cto
r
R
e
g
r
ess
io
n
(
SVR
)
ca
n
i
m
p
r
o
v
e
v
ar
io
u
s
i
n
ter
esti
n
g
f
ea
t
u
r
es
a
n
d
p
r
o
d
u
ce
a
b
etter
p
er
f
o
r
m
a
n
ce
[
2
]
.
T
h
e
ca
lcu
la
tio
n
i
s
co
n
s
tr
u
c
ted
o
n
th
e
co
n
ce
p
tio
n
o
f
m
i
n
i
m
izatio
n
i
n
s
tr
u
ct
u
r
al
r
i
s
k
.
T
h
e
co
n
ce
p
t
o
f
p
er
f
o
r
m
a
n
ce
is
b
etter
th
an
t
h
e
tr
ad
itio
n
al
E
m
p
ir
ical
R
is
k
Mi
n
i
m
izat
io
n
(
E
R
M)
w
h
er
e
it
w
as
w
o
r
k
ed
i
n
co
n
v
en
t
io
n
al
n
e
u
r
al
n
et
w
o
r
k
s
[
3
]
.
A
ct
u
all
y
,
S
VM
h
a
s
t
h
e
p
u
r
p
o
s
e
to
s
o
l
v
e
th
e
clas
s
i
f
icatio
n
co
n
d
itio
n
,
b
u
t
latel
y
it
ca
n
b
e
u
s
ed
in
th
e
r
eg
r
es
s
io
n
d
o
m
ai
n
.
Or
ig
i
n
all
y
,
it
w
a
s
d
esig
n
ed
f
o
r
s
o
lv
in
g
p
atter
n
r
ec
o
g
n
itio
n
.
Dete
r
m
in
a
tio
n
o
f
h
y
p
er
p
la
n
e
is
s
ep
ar
atin
g
th
e
p
o
s
itiv
e
a
n
d
n
e
g
ati
v
e
e
n
v
ir
o
n
m
en
t
v
alu
e
o
f
th
e
m
.
T
h
is
m
et
h
o
d
is
v
er
y
co
m
m
a
n
d
u
s
ed
in
f
u
n
d
a
m
e
n
tal
r
is
k
m
in
i
m
izatio
n
an
d
n
u
m
er
ical
lear
n
in
g
t
h
eo
r
y
[
4
]
.
T
h
e
lear
n
in
g
a
n
d
tr
ain
i
n
g
er
r
o
r
r
ate
w
er
e
u
s
ed
f
o
r
test
i
n
g
i
n
t
h
e
l
i
m
ited
d
ata
er
r
o
r
.
A
NN
an
d
S
VR
ar
e
th
e
m
et
h
o
d
s
th
at
co
u
ld
b
e
r
u
n
d
ata
to
f
i
n
d
th
e
b
est
m
o
d
el
w
ith
t
h
eir
d
ata
’
s
ch
ar
ac
ter
is
tic
[
5
]
.
T
h
e
m
o
d
el
w
ill
r
ep
r
ese
n
t
t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
5
,
Octo
b
er
2
0
1
8
:
3
3
4
1
–
3
3
4
8
3342
co
n
d
itio
n
o
f
d
ata
ac
cu
r
ac
y
[
6
]
.
T
h
e
m
o
d
el
w
a
s
tr
ea
ted
t
h
e
b
est
ac
cu
r
ac
y
i
f
t
h
e
co
m
b
in
atio
n
o
f
th
e
h
id
d
e
n
la
y
er
,
t
h
e
n
eu
r
o
n
,
th
e
ac
ti
v
at
io
n
f
u
n
ctio
n
a
n
d
t
h
e
k
in
d
o
f
tr
ain
i
n
g
f
u
n
ctio
n
co
n
tr
ib
u
te
th
e
s
m
aller
Me
a
n
Sq
u
ar
e
E
r
r
o
r
(
MSE
)
b
el
o
n
g
t
o
th
e
k
i
n
d
s
o
f
d
ata
t
h
at
f
o
r
ec
asted
w
h
er
e
it
w
a
s
co
m
p
ar
ed
to
th
e
o
r
ig
i
n
al
d
ata.
T
h
e
co
m
b
in
a
tio
n
o
f
p
ar
a
m
eter
s
th
at
r
u
n
i
s
ca
lled
ar
ch
itect
u
r
e
[
7
]
.
T
h
is
p
ap
er
is
m
ai
n
l
y
p
r
o
p
o
s
e
an
A
NN
a
n
d
SV
R
ap
p
r
o
ac
h
to
c
h
o
o
s
e
t
h
e
b
e
s
t
f
it
o
f
p
ar
am
eter
s
b
ef
o
r
e
it
co
u
ld
b
e
u
s
ed
to
th
e
s
p
ec
if
ic
s
tep
s
o
f
t
h
e
n
e
t
w
o
r
k
.
A
N
N’
s
p
ar
a
m
eter
s
ar
e
a
v
ar
iet
y
i
n
p
er
ce
n
t
o
f
d
ata,
th
e
h
id
d
en
la
y
er
,
th
e
n
e
u
r
o
n
,
th
e
tr
a
n
s
f
er
f
u
n
ctio
n
,
an
d
tr
ain
i
n
g
f
u
n
ctio
n
.
So
m
e
p
ar
a
m
eter
s
o
f
S
VR
ar
e:
ke
r
n
el
f
u
n
ctio
n
,
lo
s
t
f
u
n
ctio
n
an
d
in
s
e
n
s
i
tiv
it
y
.
T
h
e
ar
ch
itec
tu
r
e
w
i
ll
in
f
l
u
en
ce
t
h
e
r
esu
l
t
o
f
m
ea
s
u
r
e
m
en
t
o
f
a
n
et
w
o
r
k
.
2
.
M
E
T
H
O
DO
L
O
G
Y
T
h
e
m
et
h
o
d
o
lo
g
y
w
i
ll
b
r
ief
t
h
e
v
ie
w
s
tep
o
f
ar
ch
itec
tu
r
e,
ea
ch
p
ar
a
m
eter
th
at
r
ep
r
ese
n
ti
n
g
b
o
th
m
et
h
o
d
s
b
et
w
ee
n
A
NN
an
d
S
VR
.
I
n
t
h
is
r
e
s
ea
r
ch
w
ill
f
o
cu
s
i
n
a
b
ac
k
p
r
o
p
ag
atio
n
n
et
w
o
r
k
a
n
d
eise
n
s
i
tiv
e
to
SVR
[
8
]
.
T
h
e
p
r
o
ce
s
s
o
f
m
et
h
o
d
o
lo
g
y
is
ill
u
s
tr
ated
at
Fi
g
u
r
e
1
.
Fig
u
r
e
1
.
Me
th
o
d
o
lo
g
y
o
f
s
tu
d
y
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
3
.
1
.
Da
t
a
ex
peri
m
ent
T
h
e
v
ar
iab
les
d
eter
m
i
n
ato
f
d
e
m
an
d
ar
e
GDP
g
r
o
w
t
h
(
D1
)
;
P
o
p
u
latio
n
(
D2
)
,
A
p
o
ten
tia
l
cu
s
to
m
er
(
D3
)
,
;
P
r
ice
(
D4
)
;
Sales
(
D5
)
;
A
d
v
er
tis
in
g
(
D6
)
;
Q
u
ali
t
y
(
D7
)
;
E
x
p
ec
tatio
n
f
u
t
u
r
e
p
r
ice
(
D8
)
;
P
r
ef
er
en
ce
p
r
ice,
(
T
r
en
d
s
ea
s
o
n
al)
(
D9
)
[
9
]
.
T
h
e
f
lu
ctu
a
tio
n
s
o
f
d
ata
s
h
o
w
t
h
e
ch
ar
ac
ter
is
tic
o
f
ti
m
e
s
er
ies
d
ata
s
et
in
m
o
n
t
h
l
y
b
asis
o
r
in
8
y
ea
r
s
ce
m
en
t
d
e
m
a
n
d
[
10
]
.
3
.
2
.
Desig
n o
f
ANN
pa
ra
m
et
er
s
3
.
2
.
1
.
T
est
o
f
inp
ut
v
a
ria
ble
T
h
e
d
if
f
er
e
n
ce
v
ar
iab
le
h
a
s
b
ee
n
ca
lc
u
lated
ab
o
v
e
w
i
th
s
el
ec
ted
d
ata
co
r
r
elatio
n
to
d
e
m
a
n
d
an
d
t
h
e
to
tal
d
ata
s
et
.
T
h
is
e
x
p
er
i
m
e
n
t
s
h
o
w
s
th
e
in
f
l
u
e
n
ce
o
f
th
e
a
m
o
u
n
t
o
f
i
n
p
u
t
v
ar
iab
les
w
it
h
s
i
g
m
o
id
a
s
a
tr
a
n
s
f
er
f
u
n
ctio
n
T
ab
le
1
.
T
a
b
le
1
th
e
v
ar
iab
les
f
r
o
m
2
v
ar
iab
les
w
er
e
v
ar
ied
to
6
v
ar
iab
les.
W
h
en
t
h
e
a
m
o
u
n
t
o
f
v
ar
iab
les
i
n
cr
ea
s
e,
th
e
M
SE
ten
d
s
to
d
ec
r
ea
s
ed
.
T
h
e
s
m
al
l
est
w
as
6
v
ar
iab
les,
w
i
th
t
h
e
MSE
3
.
78e
-
6
(
P
o
s
t
p
r
o
ce
s
s
in
g
v
alu
e
b
u
t
it
w
as
n
o
t
r
ep
le
b
ac
k
to
t
h
e
in
itial
s
c
ale,
it
i
s
el
ig
ib
le
to
co
m
p
ar
e
d
ea
ch
o
t
h
er
)
.
T
h
e
p
u
r
p
o
s
e
m
o
d
el
o
f
A
N
N
is
s
h
o
w
n
i
n
Fi
g
u
r
e
2
(
a)
an
d
Fig
u
r
e
2
(
b
)
.
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
I
n
itia
l O
p
tima
l P
a
r
a
mete
r
s
o
f A
r
tifi
cia
l Neu
r
a
l
N
etw
o
r
k
a
n
d
S
u
p
p
o
r
t V
ec
to
r
R
eg
r
ess
io
n
(
E
d
y
F
r
a
d
in
a
t
a
)
3343
T
ab
le
1
.
T
est R
u
n
t
h
e
Am
o
u
n
t
o
f
Var
iab
les
A
mo
u
n
t
o
f
v
a
r
i
a
b
l
e
s
M
S
E
2
(
D
3
,
D
4
)
4
.
2
0
e
-
6
4
(
D
3
,
D
4
,
D
5
,
D
6
)
6
.
2
2
e
-
6
6
(
D
3
,
D
4
,
D
5
,
D
6
,
D
7
,
D
9
)
3
.
7
8
e
-
6
Fig
u
r
e
2
.
T
h
e
p
u
r
p
o
s
e
co
n
ce
p
t
o
f
b
ac
k
p
r
o
p
ag
atio
n
n
e
u
r
al
n
et
w
o
r
k
3
.
2
.
2
.
T
est
o
f
ent
ra
nce
da
t
a
s
et
Six
v
ar
iab
le
i
n
p
u
t
d
ata
w
a
s
v
ar
ied
f
r
o
m
4
0
%
to
1
0
0
%
th
e
n
m
ea
s
u
r
ed
th
e
MSE
,
t
h
e
r
esu
l
ted
ca
n
b
e
s
ee
n
in
T
ab
le
2
.
T
ab
le
2
w
h
e
n
p
er
ce
n
t
o
f
d
ata
i
n
cr
ea
s
e
th
e
MSE
d
ec
r
ea
s
e.
T
h
e
m
i
n
i
m
u
m
MSE
r
es
u
lt
s
1
0
0
%
o
f
d
ata.
T
ab
le
2
.
Var
y
in
g
P
er
ce
n
t I
n
p
u
t o
f
Data
P
e
r
c
e
n
t
D
a
t
a
F
e
e
d
D
a
t
a
M
S
E
4
0
%
38
8
.
0
3
e
-
7
5
0
%
48
7
.
7
9
e
-
7
6
0
%
58
7
.
8
8
e
-
7
7
0
%
68
7
.
7
7
e
-
7
8
0
%
78
6
.
6
1
e
-
7
9
0
%
88
5
.
6
8
e
-
7
1
0
0
%
96
4
.
7
3
e
-
7
3
.
2
.
3
.
T
est
diff
er
ence
o
f
a
ct
iv
a
t
i
o
n f
un
ct
io
n
T
h
e
test
f
o
r
th
is
ac
tiv
atio
n
f
u
n
ctio
n
t
h
r
ea
ted
2
k
in
d
s
o
f
ac
tiv
atio
n
f
u
n
ctio
n
.
T
h
ey
w
er
e
s
i
g
m
o
id
an
d
p
u
r
elin
.
T
h
is
ac
tiv
at
io
n
ai
m
ed
to
p
u
r
s
u
e
th
e
ac
tiv
a
ted
o
f
t
h
e
d
ata
to
p
r
o
ce
s
s
th
eir
r
a
n
g
e.
T
ab
le
3
s
h
o
w
s
if
th
e
v
ar
iab
le
i
n
cr
ea
s
ed
t
h
e
MSE
o
f
s
i
g
m
o
id
te
n
d
to
th
e
d
ec
r
ea
s
ed
w
ea
th
er
h
a
s
a
p
ea
k
at
th
e
4
v
a
r
iab
les.
Var
iab
le
6
is
s
m
alles
t f
o
r
s
i
g
m
o
id
.
T
ab
le
3
.
R
u
n
w
it
h
d
if
f
er
en
t
Activ
atio
n
Fu
n
ctio
n
No
V
a
r
i
a
b
l
e
s
M
S
E
S
i
g
mo
i
d
P
u
r
e
l
i
n
1
2
4
.
2
0
e
-
6
5
.
3
0
e
-
6
2
4
6
.
2
2
e
-
6
1
.
1
3
e
-
6
3
6
3
.
7
8
e
-
6
4
.
2
6
e
-
6
3
.
2
.
4
.
T
est
o
f
lea
rning
ra
t
e
T
h
e
lear
n
in
g
r
ate
tr
ied
s
o
m
e
k
in
d
s
o
f
r
ate:
5
0
,
1
0
0
,
1
5
0
an
d
2
0
0
.
It
c
an
b
e
s
ee
n
in
T
ab
le
4
.
T
ab
le
4
s
h
o
w
s
t
h
at
lear
n
i
n
g
r
ate
in
cr
ea
s
ed
to
co
n
tr
ib
u
te
th
e
i
m
p
ac
t
o
f
th
e
MSE
d
ec
r
ea
s
ed
at
p
o
in
t
1
5
0
.
T
h
is
p
o
in
t
w
a
s
co
n
tr
ib
u
ted
th
e
s
m
al
lest
er
r
o
r
w
it
h
0
.
0
0
0
1
8
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
5
,
Octo
b
er
2
0
1
8
:
3
3
4
1
–
3
3
4
8
3344
T
ab
le
4
.
R
u
n
w
it
h
a
d
if
f
er
en
t
L
ea
r
n
i
n
g
R
ate
N
e
u
r
o
n
M
S
E
50
0
.
0
0
0
2
0
1
1
0
0
0
.
0
0
0
2
2
5
1
5
0
0
.
0
0
0
1
8
9
2
0
0
0
.
0
0
0
2
1
0
3
.
2
.
5
.
T
est
o
f
a
hid
den la
y
er
Hid
d
en
la
y
er
s
et
tr
y
w
it
h
1
,
2
an
d
3
,
(
f
r
o
m
:
Si
g
m
o
id
,
9
6
d
at
a)
.
T
h
e
test
w
it
h
r
an
d
o
m
b
lo
ck
in
g
,
s
ee
i
n
T
ab
le
5
.
T
h
e
T
ab
le
5
s
h
o
w
s
t
h
e
g
r
o
u
p
la
y
er
s
co
m
b
i
n
e
i
n
th
r
e
e
o
b
s
er
v
atio
n
s
.
I
t
te
n
d
s
to
d
ec
r
ea
s
e
in
t
h
eir
p
o
o
l.
T
h
en
f
r
o
m
t
h
e
la
y
er
1
MSE
is
th
e
s
m
alle
s
t a
t 0
.
0
0
0
2
4
8
.
T
h
e
1
lay
er
w
il
l b
e
u
s
ed
[
11
]
.
T
ab
le
5
.
R
u
n
th
e
H
id
d
en
L
a
y
e
r
O
b
se
r
v
a
t
i
o
n
(
G
r
o
u
p
)
L
a
y
e
r
R
e
su
l
t
(
M
S
E)
1
1
0
.
0
0
0
3
4
9
2
1
0
.
0
0
0
2
4
8
3
1
0
.
0
0
0
3
7
8
4
2
0
.
0
0
0
2
5
6
5
2
0
.
0
0
0
3
3
9
6
2
0
.
0
0
0
3
6
3
7
3
0
.
0
0
0
3
1
3
8
3
0
.
0
0
0
3
7
9
3
0
.
0
0
0
4
4
5
3
.
2
.
6
.
T
est
t
he
a
m
o
un
t
s
o
f
neuro
n
i
n t
he
L
a
y
er
T
h
e
test
am
o
u
n
t
o
f
n
eu
r
o
n
wer
e
test
ed
w
ith
3
d
if
f
er
en
t
n
e
u
r
o
n
s
,
6
,
8
an
d
1
0
.
I
t
s
h
o
w
s
i
n
T
ab
le
6
.
T
ab
le
6
s
h
o
w
s
t
h
at
th
e
a
m
o
u
n
t
o
f
n
e
u
r
o
n
w
as
in
cr
ea
s
ed
w
ill
co
n
tr
ib
u
te
t
h
e
MSE
w
a
s
d
ec
r
ea
s
ed
an
d
1
0
n
eu
r
o
n
s
w
er
e
th
e
b
est co
n
tr
ib
u
ted
to
er
r
o
r
.
T
ab
le
6
.
R
u
n
w
it
h
a
d
if
f
er
en
t
Am
o
u
n
t o
f
Ne
u
r
o
n
A
mo
u
n
t
o
f
N
e
u
r
o
n
M
S
E
6
0
.
0
0
4
8
1
8
0
.
0
0
4
5
2
10
0
.
0
0
3
1
0
3
.
2
.
7
.
T
est
o
f
net
w
o
rk
t
ra
ini
ng
f
un
ct
io
n
T
h
e
v
ar
io
u
s
n
et
w
o
r
k
tr
ain
i
n
g
f
u
n
ctio
n
s
ar
e
ap
p
lied
in
th
is
ex
p
er
i
m
en
t
to
s
ee
th
e
ef
f
ec
t
iv
it
y
ea
c
h
n
et
w
o
r
k
tr
ain
i
n
g
f
u
n
ct
io
n
,
i
t
c
an
b
e
s
ee
n
i
n
T
ab
le
7
.
T
ab
le
7
s
h
o
w
s
t
h
e
v
ar
iet
y
o
f
n
et
w
o
r
k
tr
ai
n
i
n
g
f
u
n
ctio
n
s
an
d
th
e
b
es
t tr
ain
in
g
f
u
n
ctio
n
i
s
T
r
ain
l
m
a
n
d
th
e
s
ec
o
n
d
i
s
T
r
ain
g
d
m
.
T
h
e
s
t
u
d
y
i
s
tr
ied
w
it
h
s
i
x
v
ar
iab
les a
n
d
s
h
o
w
s
in
th
e
T
ab
le
1
th
at
th
e
a
m
o
u
n
t
o
f
v
ar
iab
le
s
ar
e
i
n
cr
ea
s
ed
th
e
M
SE
d
ec
r
ea
s
ed
w
it
h
m
i
n
i
m
u
m
3
.
7
8
e
-
6
.
T
h
e
v
ar
iab
le
w
ill
i
n
f
l
u
en
ce
t
h
e
r
esu
lt
o
u
tp
u
t
o
f
p
r
ed
ictio
n
,
in
th
i
s
r
esear
ch
s
ix
v
ar
iab
le
s
ar
e
b
etter
am
o
u
n
t
th
an
th
e
s
m
aller
d
ataset,
t
h
is
i
s
v
er
y
r
ea
s
o
n
ab
le
f
o
r
n
e
u
r
al
n
et
w
o
r
k
p
o
w
er
f
u
l
to
s
i
m
u
la
te
n
o
n
lin
ea
r
b
elo
n
g
t
h
e
n
u
m
b
er
o
f
d
if
f
er
en
t v
ar
iab
les i
n
h
o
r
izo
n
ter
m
s
o
f
ti
m
e
[
12
]
.
T
h
en
at
T
ab
le
2
s
h
o
w
s
th
e
a
m
o
u
n
t
o
f
d
ata
i
n
cr
ea
s
e
w
h
i
le
t
h
e
MSE
d
ec
r
ea
s
e
a
n
d
th
e
b
est
p
er
ce
n
tag
e
is
1
0
0
%
o
r
9
6
a
m
o
u
n
t
s
o
f
d
at
a
w
it
h
MSE
4
.
7
3
e
-
7
.
T
h
is
is
r
ea
s
o
n
ab
le
f
o
r
th
e
b
ig
g
er
d
ata
s
h
o
u
ld
i
m
p
r
o
v
e
t
h
e
b
etter
r
esu
lt
o
f
p
r
ed
ictio
n
f
r
o
m
t
h
e
o
u
tp
u
t
p
atter
n
o
f
n
eu
r
al
n
et
w
o
r
k
f
o
r
th
is
ch
ar
ac
ter
i
s
ti
c
o
f
d
ataset.
T
h
is
is
r
elev
an
t
to
th
e
t
h
eo
r
y
o
f
n
e
u
r
al
n
et
w
o
r
k
t
h
at
n
eu
r
al
n
et
w
o
r
k
is
b
etter
w
o
r
k
i
n
g
w
it
h
b
i
g
d
ata
th
e
n
s
m
aller
,
b
ec
au
s
e
t
h
e
s
m
aller
d
ata
c
o
u
ld
n
o
t
d
o
th
e
tr
ai
n
in
g
p
r
o
ce
s
s
m
o
r
e
ac
cu
r
atel
y
a
n
d
th
e
b
ias
w
i
ll
b
e
h
ig
h
er
[
13
]
,
[
14
]
.
T
ab
le
3
s
h
o
w
s
th
e
test
o
f
a
ctiv
atio
n
f
u
n
ct
io
n
s
ar
e
v
ar
ie
d
w
ith
Si
g
m
o
id
an
d
P
u
r
elin
,
th
e
b
est
ac
tiv
atio
n
f
u
n
ct
io
n
is
s
i
g
m
o
id
o
n
s
ix
v
ar
iab
les
co
m
p
ar
e
to
ea
ch
o
th
er
o
n
th
e
s
a
m
e
a
m
o
u
n
t
o
f
v
ar
iab
les
s
o
it
w
o
u
ld
b
e
u
s
ed
f
o
r
th
e
p
ar
a
m
e
ter
s
to
k
ee
p
s
m
o
o
th
l
y
r
u
n
n
in
g
to
ex
ec
u
te
d
ata
o
n
th
e
r
a
n
g
e
o
f
0
to
1
.
T
ab
le
4
s
h
o
w
s
t
h
e
d
if
f
er
en
t
lear
n
i
n
g
r
ate
f
r
o
m
5
0
to
2
0
0
an
d
th
e
b
e
s
t
o
n
e
w
as
at
1
5
0
.
T
h
is
lear
n
i
n
g
r
ate
w
ill
h
elp
th
e
d
ata
to
p
r
o
ce
s
s
in
th
e
o
v
er
lap
o
f
th
e
r
ea
l
d
ata
b
ef
o
r
e
test
in
g
th
e
d
ata
to
p
r
ed
ictio
n
.
I
n
t
h
is
s
ec
tio
n
,
th
e
r
u
le
d
ef
in
e
s
th
e
n
e
t
w
o
r
k
w
ei
g
h
t
o
n
tr
ial
an
d
er
r
o
r
b
y
an
ep
o
ch
.
T
h
e
er
r
o
r
is
u
p
d
ated
to
s
u
p
er
v
is
ed
lear
n
i
n
g
u
n
til
f
o
u
n
d
th
e
s
m
aller
n
et
w
o
r
k
er
r
o
r
[
15
]
.
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
I
n
itia
l O
p
tima
l P
a
r
a
mete
r
s
o
f A
r
tifi
cia
l Neu
r
a
l
N
etw
o
r
k
a
n
d
S
u
p
p
o
r
t V
ec
to
r
R
eg
r
ess
io
n
(
E
d
y
F
r
a
d
in
a
t
a
)
3345
T
ab
le
5
s
h
o
w
s
th
e
v
ar
ietie
s
o
f
la
y
er
an
d
t
h
e
o
b
s
er
v
atio
n
o
f
th
e
b
est M
SE
w
ith
0
.
0
0
0
2
4
8
w
it
h
la
y
er
1
,
th
e
la
y
er
to
h
elp
t
h
e
d
ata
r
u
n
n
in
g
o
n
t
h
e
o
p
ti
m
al
to
k
ee
p
t
h
e
o
v
er
f
it
tin
g
p
r
o
ce
s
s
,
b
ec
a
u
s
e
i
f
to
o
m
an
y
la
y
er
w
il
l
n
o
t
tak
e
th
e
lo
n
g
ti
m
e
p
r
o
ce
s
s
.
I
t
w
ill
also
o
cc
u
r
e
t
h
e
o
v
er
f
it
tin
g
w
i
th
th
e
d
ata
w
ea
t
h
er
th
e
la
y
er
s
h
o
u
ld
b
e
o
b
tain
th
e
b
etter
r
esu
l
t b
u
t t
h
e
o
v
er
f
itti
n
g
w
il
l b
e
s
to
p
p
ed
o
n
th
e
p
r
o
ce
s
s
o
f
f
o
r
ec
asti
n
g
d
ataset.
I
n
t
h
is
s
tep
,
th
e
n
u
m
b
er
o
f
w
ei
g
h
t
s
d
o
th
e
iter
ate
ca
lcu
latio
n
to
th
e
h
id
d
en
la
y
er
s
p
ar
t.
T
h
e
n
u
m
b
er
s
o
f
w
eig
h
t
s
ar
e
d
ep
en
d
in
g
o
n
th
e
s
ize
o
f
tr
ain
i
n
g
s
et
to
t
h
e
i
n
d
iv
id
u
al
r
ef
lec
t
io
n
o
f
d
ata
a
n
d
u
s
e
f
o
r
ac
tu
a
l
f
o
r
ec
asti
n
g
d
ata
s
et.
On
t
h
e
tab
le
s
h
o
w
s
,
t
h
at
m
o
r
e
a
m
o
u
n
t
o
f
h
id
d
en
la
y
er
s
co
n
tr
ib
u
tes
u
n
s
ati
s
f
y
r
es
u
lt
a
f
t
er
it
w
as
r
u
n
m
o
r
e
a
m
o
u
n
t
o
f
h
id
d
en
la
y
er
s
.
So
m
e
r
ec
o
m
m
e
n
d
atio
n
s
f
r
o
m
o
t
h
er
r
esear
ch
er
s
v
er
y
co
m
m
o
n
to
u
s
e
o
n
e
o
f
h
id
d
en
la
y
er
is
b
etter
th
a
n
m
o
r
e
i
n
a
p
r
o
ce
s
s
o
f
it.
T
ab
le
6
f
r
o
m
t
h
is
e
x
p
er
i
m
e
n
t
s
h
o
w
th
e
―
d
i
f
f
er
en
ce
s
‖
a
m
o
u
n
t
o
f
d
eter
m
i
n
atio
n
n
e
u
r
o
n
s
in
h
id
d
en
la
y
er
,
th
e
s
tep
w
as
s
tar
ti
n
g
f
r
o
m
t
h
e
s
m
alle
s
t
n
u
m
b
er
to
h
i
g
h
er
n
u
m
b
er
o
f
n
eu
r
o
n
s
w
h
er
e
th
e
co
n
tr
ib
u
tio
n
o
f
th
e
n
eu
r
o
n
s
i
g
n
i
f
ica
n
tl
y
to
g
e
t
th
e
s
m
aller
M
SE.
T
h
e
r
an
d
o
m
n
u
m
b
er
s
a
m
p
le
ar
e
tak
e
n
f
r
o
m
6
to
1
0
an
d
th
is
d
eter
m
in
at
io
n
o
b
tain
t
h
e
b
est
MSE
0
.
0
0
3
1
0
w
ith
1
0
n
eu
r
o
n
s
[
16
]
.
T
ab
le
7
tells
th
e
v
ar
iatio
n
o
f
tr
ai
n
i
n
g
f
u
n
ctio
n
w
h
er
e
t
h
e
b
es
t
r
es
u
lt
is
T
r
ain
l
m
w
i
th
MSE
0
.
0
0
0
2
3
4
in
al
g
o
r
ith
m
o
f
L
e
v
en
b
er
g
Ma
r
q
u
ad
t.
I
n
th
i
s
ca
s
e,
T
r
ain
l
m
is
b
etter
t
h
an
s
ig
m
o
id
w
ea
t
h
er
s
i
g
m
o
id
is
m
o
r
e
co
m
m
o
n
is
u
s
ed
to
tr
a
in
b
ac
k
p
r
o
p
ag
atio
n
alg
o
r
ith
m
[
17
]
.
T
ab
le
7
.
Dif
f
er
en
t
n
et
w
o
r
k
tr
a
in
i
n
g
f
u
n
ct
io
n
T
r
a
i
n
i
n
g
F
u
n
c
t
i
o
n
M
S
E
T
r
a
i
n
l
m
0
.
0
0
0
2
3
4
T
r
a
i
n
g
d
m
0
.
0
0
0
4
2
8
T
r
a
i
n
i
n
g
d
a
0
.
0
0
0
8
1
7
T
r
a
i
n
g
d
x
0
.
0
0
1
1
0
Fro
m
th
e
d
is
c
u
s
s
io
n
p
ar
t,
it
c
an
co
n
cl
u
d
e
th
at
t
h
e
r
esu
lt
will
b
e
g
iv
e
n
t
h
e
b
est
f
i
t
o
f
d
ata
if
u
s
e
th
e
s
elec
ted
v
ar
iab
les
,
m
ea
n
i
n
g
t
h
at
t
h
e
r
es
u
lt
f
r
o
m
ea
c
h
tr
ai
n
in
g
f
u
n
ctio
n
w
il
l
b
e
i
n
f
lu
e
n
c
ed
s
ig
n
i
f
ica
n
tl
y
a
n
d
r
ed
u
ce
th
e
o
v
er
f
itti
n
g
p
r
o
ce
s
s
to
o
b
tain
th
e
o
p
ti
m
al
co
n
d
itio
n
.
3
.
3
.
Desig
n o
f
SVR’
s
pa
ra
m
et
er
s
T
h
er
e
ar
e
s
o
m
e
p
ar
a
m
eter
s
in
SVR
to
co
n
s
tr
u
ct
th
e
S
VM
f
o
r
p
r
ed
ictin
g
.
Ho
w
e
v
er
,
th
e
t
w
o
d
o
m
i
n
a
n
t
r
elev
an
t
ar
e
e
-
in
s
en
s
iti
v
it
y
a
n
d
k
er
n
el
f
u
n
ctio
n
b
ec
a
u
s
e
b
o
th
p
ar
a
m
eter
s
co
u
ld
b
e
in
cr
ea
s
ed
th
e
e
-
m
ea
n
a
n
d
d
ec
r
ea
s
ed
th
e
er
r
o
r
an
d
in
cr
ea
s
in
g
t
h
e
ac
cu
r
ac
y
o
f
th
e
p
r
o
ce
s
s
o
f
d
ata.
I
t
ca
n
d
ec
r
ea
s
e
th
e
n
u
m
b
er
o
f
SV
s
lead
in
g
to
d
ata
co
m
p
r
ess
io
n
.
T
h
e
p
ar
am
eter
s
o
f
SVR
a
r
e
k
er
n
el
f
u
n
ctio
n
,
ε
-
i
n
s
e
n
s
i
tiv
e
lo
s
s
f
u
n
ct
io
n
,
in
s
e
n
s
itiv
it
y
,
a
n
u
p
p
er
b
o
n
d
.
T
h
e
test
is
u
s
i
n
g
t
h
e
d
i
f
f
er
e
n
t
a
m
o
u
n
t
o
f
d
ata.
T
h
e
d
ata
w
i
ll
b
e
u
s
ed
6
v
ar
iab
les.
Ker
n
el
F
u
n
c
tio
n
:
L
i
n
ea
r
,
P
o
l
y
n
o
m
ial,
R
ad
ial
B
asis
F
u
n
c
tio
n
,
T
an
g
en
t
H
y
p
er
b
o
lic,
an
d
L
o
s
s
f
u
n
ctio
n
’
s
p
ar
am
eter
s
ar
e
e
-
in
s
en
s
iti
v
e,
Qu
ad
r
atic,
L
ap
lace
an
d
H
u
b
er
.
I
n
s
en
s
iti
v
it
y
is
1
.
K
ern
el
F
u
n
ctio
n
is
th
e
class
i
f
icatio
n
p
r
o
b
le
m
s
in
o
p
tim
al
co
n
d
itio
n
σ
ca
n
b
e
co
m
p
u
ted
b
ased
o
n
Fis
h
er
d
is
cr
i
m
i
n
atio
n
.
I
t
is
also
to
r
eg
r
ess
io
n
th
e
p
r
o
b
lem
s
in
th
e
b
asic
o
f
s
ca
le,
s
p
ac
e
th
eo
r
y
,
an
d
it
is
d
em
o
n
s
tr
ated
th
e
ex
is
ten
ce
o
f
a
ce
r
tain
r
an
g
e
o
f
σ
,
w
it
h
i
n
t
h
e
g
e
n
er
aliza
tio
n
p
er
f
o
r
m
a
n
ce
is
s
tab
le
.
A
ce
r
tain
i
m
p
o
r
tan
t
i
n
t
h
e
r
an
g
e
o
f
σ
ca
n
b
e
r
ea
ch
ed
v
ia
d
y
n
a
m
ic
e
v
al
u
ati
o
n
.
I
n
co
n
cl
u
s
io
n
,
t
h
e
lo
w
er
b
o
u
n
d
o
f
a
n
iter
at
in
g
s
tep
s
ize
o
f
σ
i
s
g
i
v
e
n
.
Lo
s
s
fu
n
ctio
n
is
th
e
r
elatio
n
s
h
ip
f
u
n
ct
io
n
b
et
w
ee
n
er
r
o
r
an
d
t
h
e
p
en
a
lt
y
to
t
h
at
er
r
o
r
.
T
h
e
d
if
f
er
e
n
ce
s
o
f
lo
s
s
f
u
n
ctio
n
w
ill
p
r
o
d
u
ce
t
h
e
d
if
f
er
en
ce
s
o
f
SV
R
.
L
o
s
s
f
u
n
ctio
n
ɛ
-
in
s
en
s
iti
v
e
is
th
e
v
er
y
co
m
m
o
n
.
T
h
e
ex
p
er
i
m
e
n
t star
t
s
f
r
o
m
t
h
e
6
v
ar
iab
les an
d
m
ea
s
u
r
e
th
e
r
es
u
l
t o
f
b
o
th
p
ar
am
e
ter
s
,
s
u
ch
as
:
3
.
3
.
1
.
T
est
o
f
k
er
nel f
un
ct
io
n a
nd
l
o
s
s
f
un
ct
io
n.
T
h
e
k
er
n
el
f
u
n
c
tio
n
an
d
lo
s
s
f
u
n
ct
io
n
w
er
e
test
ed
w
it
h
lin
ea
r
,
p
o
ly
n
o
m
ial
f
o
r
Ker
n
el,
an
d
e
-
in
s
e
n
s
itiv
e
f
o
r
lo
s
s
f
u
n
ctio
n
.
I
t
ca
n
b
e
s
ee
n
in
T
ab
le
8
.
T
ab
le
8
,
th
e
lin
ea
r
is
b
etter
th
an
p
o
l
y
n
o
m
ial
i
n
a
Ker
n
el
F
u
n
ct
io
n
.
I
t
w
a
s
t
h
e
b
est ch
o
ice
f
o
r
MSE
.
T
h
e
o
th
er
s
id
e
lo
s
s
f
u
n
ct
io
n
is
b
etter
f
o
r
e
in
s
e
n
s
itiv
e.
T
ab
le
8
.
Ru
n
d
if
f
er
en
t K
er
n
el
Fu
n
ctio
n
a
n
d
L
o
s
s
F
u
n
ct
io
n
G
a
u
ssi
a
n
K
e
r
n
e
l
F
u
n
c
t
i
o
n
L
o
ss
F
u
n
c
t
i
o
n
S
t
a
t
i
st
i
c
L
i
n
e
a
r
P
o
l
y
n
o
m
M
e
a
n
s
0
.
2
0
0
7
0
.
2
0
0
7
M
S
E
0
.
0
0
2
1
0
.
0
2
5
7
SD
0
.
0
0
1
8
0
.
0
0
1
8
S
t
a
t
i
st
i
c
e
-
i
n
se
n
si
t
i
v
e
M
e
a
n
s
0
.
2
0
0
7
M
S
E
0
.
0
0
1
8
SD
0
.
0
0
2
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
5
,
Octo
b
er
2
0
1
8
:
3
3
4
1
–
3
3
4
8
3346
3
.
3
.
2
.
T
est
t
he
“
up
per
bo
n
d”
C
h
o
o
s
e
e
-
is
e
n
s
iti
v
it
y
as
a
f
o
cu
s
o
n
a
v
ar
iet
y
o
f
v
ar
iab
les,
th
e
test
w
it
h
Up
B
2
an
d
3
f
r
o
m
T
ab
le
9
,
as
f
o
llo
w
:
T
ab
le
9
th
e
u
p
p
er
b
o
n
d
2
an
d
3
ar
e
n
o
ch
an
g
ed
at
al
.
I
t
ca
n
b
e
ch
o
s
en
n
u
m
b
er
2
,
m
ea
n
s
t
h
at
t
h
e
r
esu
lt
o
f
th
e
ei
n
s
e
n
s
i
tiv
e
w
h
et
h
er
it
w
a
s
ch
a
n
g
ed
,
it
w
o
u
ld
b
e
n
o
im
p
ac
t to
th
e
ei
n
s
e
n
s
it
iv
e.
T
ab
le
9
.
R
u
n
w
it
h
a
d
if
f
er
en
t
Up
p
er
B
o
n
d
in
e
-
i
s
en
s
iti
v
e
U
p
B
=
2
e
i
n
se
n
si
t
i
v
e
M
e
a
n
s
0
.
2
0
0
7
M
S
E
0
.
0
0
2
1
SD
0
.
0
0
1
8
U
p
B
=
3
e
i
n
se
n
si
t
i
v
e
M
e
a
n
s
0
.
2
0
0
7
M
S
E
0
.
0
0
2
1
SD
0
.
0
0
1
8
3
.
3
.
3
.
Chec
k
t
he
in
s
ens
it
iv
e
n
u
m
be
r
1
a
nd
2
T
h
is
test
w
a
s
v
ar
ied
o
f
th
e
i
n
s
e
n
s
i
tiv
e:
1
an
d
2
.
T
h
e
test
ed
ca
n
b
e
s
ee
n
at
T
ab
le
1
0
.
T
a
b
le
1
0
in
s
e
n
s
itiv
e
1
a
n
d
2
ar
e
tr
ied
w
it
h
e
-
in
s
en
s
iti
v
e
a
n
d
b
o
th
o
f
t
h
e
m
ar
e
th
e
b
est.
B
u
t
u
s
u
a
ll
y
b
etter
u
s
e
t
h
e
1
in
s
e
n
s
itiv
e.
T
h
is
al
s
o
s
h
o
w
s
n
o
ef
f
ec
t
to
th
e
r
es
u
lt
w
h
et
h
er
it
is
ch
a
n
g
ed
.
T
ab
le
8
s
h
o
w
s
th
e
v
ar
iatio
n
o
f
Ker
n
el
f
u
n
c
tio
n
an
d
lo
s
s
f
u
n
ctio
n
.
T
h
e
k
er
n
el
f
u
n
ctio
n
v
a
r
iatio
n
is
lin
ea
r
o
n
g
o
o
d
r
esu
lt
an
d
s
h
o
w
n
b
etter
th
an
p
o
l
y
n
o
m
ial
w
it
h
0
.
0
0
2
1
.
T
h
e
lo
s
s
f
u
n
ctio
n
i
s
s
m
all
en
o
u
g
h
to
b
e
u
s
ed
w
it
h
ei
n
s
en
s
it
iv
e
w
it
h
MS
E
0
.
0
0
1
8
.
T
h
e
k
er
n
el
f
u
n
c
tio
n
s
h
av
e
f
u
n
ctio
n
o
f
co
n
s
tr
u
c
tin
g
th
e
n
o
n
li
n
ea
r
d
ec
is
io
n
h
y
p
er
-
s
u
r
f
ac
e
o
n
t
h
e
i
n
p
u
t
s
p
ac
e
o
f
SVR
.
B
o
th
o
f
th
e
m
m
u
s
t
b
e
s
elec
ted
co
r
r
ec
tly
w
h
er
e
th
e
s
tr
u
ct
u
r
e
w
a
s
d
ef
in
ed
o
n
th
e
d
i
m
e
n
s
io
n
al
f
ea
t
u
r
e
s
p
ac
e
an
d
o
r
d
er
co
m
p
lex
to
en
d
s
o
l
u
tio
n
[
18
]
.
Oth
er
r
esear
ch
er
u
s
e
s
t
h
e
s
a
m
e
Gau
s
s
ia
n
k
er
n
el
f
u
n
ctio
n
f
o
r
p
r
ed
ict
th
e
p
er
f
o
r
m
an
ce
[
19
]
b
u
t
in
th
is
r
esear
c
h
tr
y
t
w
o
k
i
n
d
o
f
Gau
s
s
ian
k
e
r
n
el
f
u
n
ct
io
n
s
,
t
h
e
y
ar
e
lin
ea
r
a
n
d
p
o
l
y
n
o
m
ial.
T
ab
le
9
s
h
o
w
s
th
e
u
p
p
er
b
o
n
d
tr
y
w
it
h
2
a
n
d
3
n
u
m
b
er
s
,
b
u
t
i
t
s
h
o
w
s
t
h
at
n
o
ch
an
g
e
w
h
et
h
er
it
h
a
v
e
b
ee
n
ch
an
g
ed
f
o
r
b
o
th
n
u
m
b
er
s
,
it
ca
n
b
e
s
ee
n
i
n
T
ab
le
9
,
th
is
f
u
n
ct
io
n
to
k
ee
p
t
h
e
ac
cu
r
ac
y
i
n
th
e
h
y
p
er
p
lan
e
ar
ea
w
h
er
e
it
w
a
s
p
lace
d
o
n
th
e
p
o
in
ts
o
f
tr
ain
i
n
g
d
ataset
[
20
]
.
Gen
er
all
y
,
it
u
s
es
o
n
e
as th
e
u
p
p
er
b
o
n
d
s
f
o
r
th
e
ex
p
er
i
m
e
n
t
al.
T
ab
le
1
0
.
R
u
n
w
ith
d
i
f
f
er
en
t i
-
n
s
e
n
s
i
tiv
i
t
y
1
an
d
2
I
n
s
=1
ein
s
e
n
s
i
tiv
e
Me
an
s
0
.
2
0
0
7
MSE
0
.
0
0
2
1
SD
0
.
0
0
1
8
I
n
s
=2
ein
s
e
n
s
i
tiv
e
Me
an
s
0
.
2
0
0
7
MSE
0
.
0
0
2
1
SD
0
.
0
0
1
8
T
ab
le
1
0
s
h
o
w
s
t
h
e
in
s
e
n
s
iti
v
i
t
y
w
it
h
1
an
d
2
w
it
h
MSE
0
.
0
0
2
1
an
d
th
is
m
atter
al
s
o
n
o
ch
an
g
e
s
th
e
r
esu
lt
o
f
MSE
f
r
o
m
t
h
e
d
if
f
er
en
t
n
u
m
b
er
,
ch
o
o
s
e
in
s
en
s
it
y
1
,
th
e
in
s
en
s
iti
v
e
h
a
v
e
th
e
f
u
n
ctio
n
o
f
to
f
it
t
h
e
tr
ain
i
n
g
d
ata
f
r
o
m
T
ab
le
1
0
.
A
s
o
r
ig
i
n
all
y
,
t
h
e
p
u
r
p
o
s
e
u
s
e
s
v
m
w
as
f
o
r
s
o
lv
in
g
th
e
p
atter
n
r
ec
o
g
n
itio
n
ca
s
es,
b
u
t
latel
y
h
a
s
b
ee
n
e
x
t
en
d
ed
to
s
o
lv
e
n
o
n
l
in
ea
r
r
eg
r
ess
io
n
e
s
ti
m
atio
n
ca
s
es
s
u
c
h
as
in
ac
ad
e
m
ic
an
d
in
d
u
s
tr
ial
p
latf
o
r
m
s
e
-
in
s
en
s
iti
v
e
lo
s
s
f
u
n
ctio
n
[
21
]
.
Fo
r
s
v
r
t
h
e
r
esu
l
t
f
r
o
m
ea
c
h
p
ar
a
m
eter
w
il
l
b
e
in
f
lu
e
n
ce
d
s
ig
n
i
f
ica
n
tl
y
b
y
t
h
e
r
es
u
lt.
B
ec
au
s
e
t
h
e
s
v
r
w
ill
tr
a
n
s
f
o
r
m
t
h
e
d
ata
to
b
e
lin
ier
s
ep
ar
ab
le
in
th
e
f
ea
tu
r
e
s
p
ac
e
o
f
h
y
p
er
p
lan
e
to
b
e
th
e
b
est r
e
g
r
ess
io
n
.
T
h
is
m
et
h
o
d
h
as p
r
o
m
is
ed
t
h
e
g
o
o
d
m
et
h
o
d
s
in
t
h
e
f
u
t
u
r
e.
4.
CO
NCLU
SI
O
N
B
ased
o
n
t
h
is
s
t
u
d
y
,
t
h
i
s
i
s
t
h
e
in
it
ial
s
tep
to
t
h
e
n
ex
t
s
tep
f
o
r
th
e
f
u
t
u
r
e
ex
p
er
i
m
en
t
a
n
d
th
e
v
ar
ietie
s
o
f
p
ar
a
m
eter
s
o
f
d
e
m
an
d
co
u
l
d
b
e
in
f
lu
e
n
ce
d
o
n
t
h
e
ar
ti
f
ici
al
n
e
u
r
al
n
e
t
w
o
r
k
a
n
d
s
u
p
p
o
r
t
v
ec
to
r
r
eg
r
es
s
i
on
m
et
h
o
d
s
.
I
t
ca
n
b
e
co
n
cl
u
d
ed
as
f
o
llo
w
:
ANN
co
u
ld
b
e
a
n
e
f
f
ec
tiv
e
r
u
n
o
n
t
h
e
―
d
i
f
f
er
e
n
ce
s
‖
p
ar
a
m
eter
s
w
i
th
s
ix
i
n
p
u
t
v
ar
iab
le
s
,
ea
ch
co
n
d
itio
n
h
as
t
h
e
o
p
ti
m
al
p
o
in
t
i
ts
elf
.
T
h
e
r
esu
lt
o
f
t
h
is
s
tu
d
y
w
as
as
f
o
llo
w
:
th
e
ac
tiv
atio
n
f
u
n
ctio
n
w
as
Si
g
m
o
id
.
T
h
e
a
m
o
u
n
t
o
f
f
ee
d
d
at
a
w
a
s
1
0
0
%
o
r
9
6
,
1
5
0
lear
n
in
g
r
ate,
1
h
id
d
en
L
a
y
er
,
1
0
n
eu
r
o
n
s
,
tr
in
l
m
f
o
r
tr
ain
in
g
f
u
n
ctio
n
,
3
la
y
er
s
f
o
r
to
tal
lay
er
,
s
et
u
p
er
r
o
r
0
.
0
0
1
an
d
w
o
r
k
w
ith
a
n
et
w
o
r
k
o
f
f
ee
d
-
f
o
r
w
ar
d
b
ac
k
p
r
o
p
ag
atio
n
.
Fu
r
t
h
er
m
o
r
e,
t
h
e
SVR
as
w
e
ll
as
th
e
a
m
o
u
n
t
o
f
v
ar
iab
les
w
er
e
6
.
T
h
e
g
en
er
al
p
ar
am
e
ter
s
w
er
e
u
s
ed
w
i
th
li
n
ea
r
k
er
n
el
f
u
n
ct
io
n
,
e
-
i
n
s
en
s
iti
v
e
lo
s
s
f
u
n
ct
io
n
,
an
d
o
n
e
in
s
e
n
s
itiv
it
y
.
So
m
e
v
ar
iab
les
ar
e
tr
ied
in
th
is
s
t
u
d
y
f
o
r
s
v
r
b
u
t
n
o
t
s
h
o
w
t
h
e
s
i
g
n
if
ican
t
c
h
an
g
es,
m
ea
n
s
th
a
t
th
e
s
v
r
d
o
es n
o
t n
ee
d
to
id
en
ti
f
y
t
h
e
i
n
itial p
ar
a
m
e
ter
esp
ec
iall
y
f
o
r
u
p
p
er
b
o
n
d
an
d
in
s
en
s
itiv
e.
I
f
th
e
s
e
in
itial
co
n
d
itio
n
o
f
p
ar
a
m
eter
s
ar
e
u
s
ed
to
d
o
th
e
n
ex
t
s
tep
f
o
r
o
th
er
p
u
r
p
o
s
e
th
e
r
esu
lt
o
f
tr
ain
i
n
g
a
n
d
s
i
m
u
la
tio
n
s
h
o
u
l
d
b
e
q
u
ick
l
y
an
d
ea
s
y
to
g
et
t
h
e
o
p
ti
m
al
co
n
d
iti
o
n
o
f
n
e
t
w
o
r
k
p
r
o
ce
s
s
d
ata
d
u
e
to
th
e
ch
ar
ac
ter
is
tic
o
f
n
e
u
r
al
n
et
w
o
r
k
ab
le
to
w
o
r
k
i
n
n
o
n
lin
ea
r
it
y
a
n
d
p
r
o
d
u
ce
th
e
s
u
i
t
ab
le
task
f
o
r
o
th
er
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
I
n
itia
l O
p
tima
l P
a
r
a
mete
r
s
o
f A
r
tifi
cia
l Neu
r
a
l
N
etw
o
r
k
a
n
d
S
u
p
p
o
r
t V
ec
to
r
R
eg
r
ess
io
n
(
E
d
y
F
r
a
d
in
a
t
a
)
3347
p
u
r
p
o
s
es
p
er
f
o
r
m
an
ce
.
A
c
tu
al
l
y
t
h
er
e
is
n
o
s
p
ec
if
y
w
a
y
to
g
et
t
h
e
b
est
r
es
u
lt
o
f
th
e
n
et
w
o
r
k
p
r
o
ce
s
s
d
ata,
m
o
s
tl
y
,
i
t
i
s
d
o
it
w
it
h
tr
ial
a
n
d
er
r
o
r
b
u
t
at
least
w
it
h
t
h
is
s
t
u
d
y
t
h
e
w
a
y
to
d
ef
i
n
e
t
h
e
o
p
ti
m
al
co
n
d
it
io
n
f
ir
s
t
b
ef
o
r
e
to
d
o
m
a
n
y
tr
ia
l
an
d
er
r
o
r
m
eth
o
d
o
lo
g
y
.
T
h
is
s
tu
d
y
f
i
n
d
i
n
g
t
h
e
w
a
y
h
o
w
to
g
e
t
th
e
s
tar
ti
n
g
in
it
ial
co
n
d
itio
n
to
s
tar
t n
e
u
r
al
n
et
w
o
r
k
p
r
o
ce
s
s
.
A
N
Ns
a
n
d
S
VR
ar
e
v
er
y
tal
en
ted
m
eth
o
d
to
b
etter
p
er
f
o
r
m
an
ce
o
f
th
e
n
et
w
o
r
k
r
es
u
lt
f
o
r
m
an
y
p
u
r
p
o
s
es,
s
u
c
h
a
s
f
o
r
f
o
r
ec
as
tin
g
,
r
o
b
o
tic,
au
to
m
o
ti
v
e,
m
e
d
ical
eq
u
ip
m
e
n
t
’
s
a
n
d
m
a
n
y
th
i
n
g
s
else.
So
m
e
r
esear
ch
er
s
h
av
e
m
an
y
co
m
p
a
r
ed
th
e
p
er
f
o
r
m
a
n
ce
o
f
tr
ad
iti
o
n
al
m
eth
o
d
s
w
h
ich
is
s
t
u
d
y
i
n
s
tatis
t
ical
m
aj
o
r
to
th
ese
m
et
h
o
d
s
,
s
p
ec
iall
y
n
e
u
r
al
n
et
w
o
r
k
m
et
h
o
d
s
b
u
t
f
o
r
SVR
t
h
e
s
tu
d
y
s
til
l
li
m
ited
a
n
d
n
ee
d
to
d
ev
elo
p
m
o
r
e
k
n
o
w
led
g
e
f
i
n
d
i
n
g
m
a
n
y
in
f
o
r
m
at
io
n
ab
o
u
t t
h
is
m
et
h
o
d
s
.
B
o
th
o
f
th
ese
m
et
h
o
d
s
ar
e
v
er
y
s
p
ec
ial
ca
s
e
b
ec
au
s
e
th
e
y
d
o
n
o
t
n
ee
d
th
e
s
tatis
tical
tes
ti
n
g
m
et
h
o
d
s
p
ec
if
icall
y
.
L
i
n
ea
r
an
d
n
o
n
li
n
ea
r
ca
n
d
o
w
it
h
t
h
i
s
m
e
th
o
d
s
,
p
ar
a
m
etr
ic
an
d
n
o
n
p
ar
a
m
et
r
ic
as
w
ell.
E
v
en
,
th
ese
m
et
h
o
d
s
w
i
ll
b
e
b
etter
w
o
r
k
in
g
w
ith
t
h
e
b
ig
d
ataset,
b
ec
au
s
e
it
ca
n
ea
s
y
to
tr
ain
t
h
e
d
a
ta
s
et
an
d
g
i
v
e
th
e
b
etter
r
esu
lt
.
T
h
e
Su
g
g
esti
o
n
t
o
th
e
n
ex
t
s
tu
d
y
i
s
a
d
ev
e
lo
p
m
en
t
o
f
t
h
ese
o
p
ti
m
izatio
n
co
n
d
itio
n
’
s
p
ar
a
m
eter
s
f
o
r
A
NN
an
d
SVR
to
d
o
t
h
e
f
u
r
th
er
s
t
u
d
y
s
u
ch
as
f
o
r
ec
asti
n
g
o
f
d
eter
m
i
n
a
n
t
o
f
d
e
m
an
d
with
d
e
v
elo
p
m
e
n
t
o
f
o
th
er
m
et
h
o
d
o
r
h
y
b
r
id
m
e
th
o
d
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
is
r
esear
ch
w
as
u
n
d
er
s
c
h
o
lar
s
h
ip
o
f
2
0
1
2
Kem
r
i
s
tek
d
i
k
ti
o
f
I
n
d
o
n
esia
.
T
h
an
k
y
o
u
v
er
y
m
u
c
h
to
Ke
m
r
is
te
k
D
i
k
ti o
f
I
n
d
o
n
esia
an
d
P
r
in
ce
o
f
So
n
g
k
l
a
U
n
i
v
er
s
it
y
,
Hat
y
ai,
T
h
ailan
d
.
RE
F
E
R
E
NC
E
S
[1
]
T
.
B.
T
ra
f
a
li
s
a
n
d
B.
S
a
n
to
sa
,
"
P
re
d
icti
n
g
m
o
n
th
ly
f
lo
u
r
p
rice
s
th
ro
u
g
h
Ne
u
ra
l
Ne
tw
o
rk
s,
RB
F
s
a
n
d
S
V
R
"
,
In
telli
g
e
n
t
E
n
g
i
n
e
e
rin
g
S
y
ste
ms
T
h
ro
u
g
h
Arti
fi
c
i
a
l
Ne
u
ra
l
Ne
two
rk
s,
v
o
l.
1
1
,
p
p
.
7
4
5
-
7
5
0
,
2
0
0
1
.
[2
]
H.
Dru
c
k
e
r,
e
t
a
l
.,
"
S
u
p
p
o
rt
v
e
c
t
o
r
m
a
c
h
in
e
s
f
o
r
sp
a
m
c
a
teg
o
riza
ti
on
"
,
Ne
u
ra
l
Ne
two
rk
s,
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
,
v
o
l.
1
0
,
p
p
.
1
0
4
8
-
1
0
5
4
,
1
9
9
9
.
[3
]
K.
M
u
ll
e
r,
e
t
a
l
.,
"
A
n
I
n
tro
d
u
c
ti
o
n
t
o
K
e
rn
e
l
-
b
a
se
d
L
e
a
rn
in
g
A
l
g
o
rit
h
m
s
"
,
Ne
u
ra
l
Ne
two
rk
s,
IEE
E
T
ra
n
s
a
c
ti
o
n
s
o
n
,
v
o
l.
1
2
,
p
p
.
1
8
1
-
2
0
1
,
2
0
0
1
.
[4
]
I.
B.
T
ij
a
n
i
a
n
d
R.
Ak
m
e
li
a
wa
t
i,
"
S
u
p
p
o
rt
V
e
c
t
o
r
Re
g
re
ss
io
n
b
a
se
d
F
rictio
n
M
o
d
e
li
n
g
a
n
d
C
o
m
p
e
n
sa
ti
o
n
in
M
o
ti
o
n
C
o
n
t
ro
l
S
y
ste
m
"
,
En
g
in
e
e
rin
g
A
p
p
li
c
a
ti
o
n
s
o
f
Art
if
icia
l
In
t
e
ll
ig
e
n
c
e
,
v
o
l.
2
5
,
p
p
.
1
0
4
3
-
1
0
5
2
,
2
0
1
2
.
[5
]
B.
S
h
a
n
,
e
t
a
l
.
,
"
A
p
p
li
c
a
ti
o
n
o
f
O
n
li
n
e
S
V
R
o
n
th
e
Dy
n
a
m
ic
L
iq
u
id
L
e
v
e
l
S
o
f
t
S
e
n
sin
g
"
,
in
Co
n
tro
l
a
n
d
De
c
isio
n
Co
n
fer
e
n
c
e
(
CCDC
),
2
0
1
3
2
5
t
h
Ch
in
e
se
,
2
0
1
3
,
p
p
.
3
0
0
3
-
3
0
0
7
.
[6
]
H.
Ese
n
,
e
t
a
l
.,
"
M
o
d
e
li
n
g
a
G
ro
u
n
d
-
c
o
u
p
le
d
He
a
t
P
u
m
p
S
y
st
e
m
b
y
a
S
u
p
p
o
r
t
V
e
c
to
r
M
a
c
h
in
e
"
,
Ren
e
wa
b
le
En
e
rg
y
,
v
o
l.
3
3
,
p
p
.
1
8
1
4
-
1
8
2
3
,
2
0
0
8
.
[7
]
S
.
J.
Ha
n
so
n
,
e
t
a
l
.,
"
Co
m
b
in
a
to
r
ial
c
o
d
e
s
in
v
e
n
tral
te
m
p
o
ra
l
lo
b
e
f
o
r
o
b
jec
t
re
c
o
g
n
it
io
n
:
Ha
x
b
y
(2
0
0
1
)
re
v
isit
e
d
:
is
t
h
e
re
a
―
f
a
c
e
‖
a
re
a
?
"
,
Ne
u
ro
ima
g
e
,
v
o
l
.
2
3
,
p
p
.
1
5
6
-
1
6
6
,
2
0
0
4
.
[8
]
B.
S
a
n
t
o
sa
,
Da
ta
M
in
i
n
g
T
e
k
n
ik
P
e
m
a
n
f
a
a
tan
Da
ta
u
n
tu
k
Ke
p
e
rlu
a
n
Bisn
is
v
o
l
.
9
7
8
,
2
0
0
7
.
[9
]
"
De
ter
m
in
a
n
ts
o
f
De
m
a
n
d
"
,
in
h
tt
p
://
ma
rk
e
t.
su
b
wiki.
o
rg
/wi
k
i/
De
t
e
rm
in
a
n
ts_
o
f
_
d
e
ma
n
d
,
e
d
,
2
2
D
e
c
e
m
b
e
r
2
0
1
2
,
c
o
ll
e
c
ted
o
n
1
2
De
c
e
m
b
e
r
2
0
1
5
.
[1
0
]
E.
F
ra
d
i
n
a
ta,
e
t
a
l
.,
"
F
o
re
c
a
stin
g
D
e
ter
m
in
a
n
t
o
f
Ce
m
e
n
t
De
m
a
n
d
in
I
n
d
o
n
e
sia
w
it
h
A
rti
f
icia
l
Ne
u
ra
l
N
e
tw
o
rk
"
,
J
o
u
rn
a
l
o
f
Asia
n
S
c
ien
ti
fi
c
Res
e
a
rc
h
,
v
o
l.
5
,
p
p
.
3
7
3
-
3
8
4
,
2
0
1
5
.
[1
1
]
E.
F
ra
d
i
n
a
ta,
e
t
a
l
.,
"
A
NN
,
A
RIM
A
a
n
d
M
A
T
i
m
e
se
rie
s
M
o
d
e
l
f
o
r
F
o
re
c
a
stin
g
in
Ce
m
e
n
t
M
a
n
u
f
a
c
tu
rin
g
In
d
u
s
try
:
Ca
se
St
u
d
y
a
t
L
a
f
a
rg
e
Ce
m
e
n
t
In
d
o
n
e
sia
—
A
c
e
h
"
,
in
Ad
v
a
n
c
e
d
In
f
o
rm
a
ti
c
s:
C
o
n
c
e
p
t,
T
h
e
o
ry
a
n
d
Ap
p
li
c
a
t
io
n
(
ICAICT
A),
2
0
1
4
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
f
,
2
0
1
4
,
p
p
.
3
9
-
4
4
.
[1
2
]
P
.
Bu
n
n
o
o
n
,
"
El
e
c
tri
c
it
y
P
e
a
k
L
o
a
d
De
m
a
n
d
u
sin
g
De
-
n
o
isin
g
W
a
v
e
let
T
ra
n
s
f
o
rm
in
teg
ra
ted
w
it
h
Ne
u
ra
l
Ne
t
w
o
rk
M
e
th
o
d
s
"
,
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
,
v
o
l.
6
,
p
.
1
2
,
2
0
1
6
.
[1
3
]
M
.
N.
Ra
o
,
e
t
a
l
.,
"
A
P
re
d
ictiv
e
M
o
d
e
l
f
o
r
M
in
in
g
Op
i
n
i
o
n
s
o
f
a
n
Ed
u
c
a
ti
o
n
a
l
Da
tab
a
se
Us
in
g
Ne
u
ra
l
Ne
tw
o
rk
s
"
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
ri
n
g
,
v
o
l.
5
,
2
0
1
5
.
[1
4
]
B.
-
H.
A
d
il
a
n
d
G
.
Yo
u
ss
e
f
,
"
H
y
b
rid
M
e
t
h
o
d
HV
S
-
M
RM
R
f
o
r
Va
riab
le
S
e
lec
ti
o
n
in
M
u
lt
il
a
y
e
r
A
rti
f
icia
l
Ne
u
ra
l
Ne
tw
o
rk
Cla
ss
i
f
ier
"
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
,
v
o
l
.
7
,
p
.
2
7
7
3
,
2
0
1
7
.
[1
5
]
I.
A
.
Ba
sh
e
e
r
a
n
d
M
.
Ha
jm
e
e
r,
"
A
rti
f
icia
l
Ne
u
ra
l
Ne
t
w
o
rk
s:
F
u
n
d
a
m
e
n
tals,
Co
m
p
u
ti
n
g
,
De
sig
n
,
a
n
d
A
p
p
li
c
a
ti
o
n
"
,
J
o
u
rn
a
l
o
f
M
icr
o
b
i
o
lo
g
ica
l
M
e
th
o
d
s,
v
o
l
.
4
3
,
p
p
.
3
-
3
1
,
2
0
0
0
.
[1
6
]
B.
S
u
ti
j
o
,
e
t
a
l
.
,
"
F
o
re
c
a
stin
g
To
u
rism
Da
ta
u
sin
g
N
e
u
ra
l
Ne
t
wo
rk
s
-
M
u
lt
isc
a
le
A
u
to
re
g
re
ss
i
v
e
M
o
d
el
"
,
J
u
rn
a
l
M
a
tem
a
ti
k
a
&
S
a
in
s,
v
o
l.
1
6
,
p
p
.
3
5
-
4
2
,
2
0
1
1
.
[1
7
]
M
.
T
.
Ha
g
a
n
a
n
d
H.
B.
De
m
u
th
,
"
Ne
u
ra
l
n
e
tw
o
rk
s
f
o
r
c
o
n
t
ro
l,
"
in
Ame
ric
a
n
Co
n
tro
l
Co
n
fer
e
n
c
e
,
1
9
9
9
.
Pro
c
e
e
d
in
g
s
o
f
t
h
e
1
9
9
9
,
1
9
9
9
,
v
o
l.
3
,
p
p
.
1
6
4
2
-
1
6
5
6
.
[1
8
]
K.
Du
a
n
,
e
t
a
l
.
,
"
Ev
a
lu
a
ti
o
n
o
f
S
i
m
p
le
P
e
rf
o
r
m
a
n
c
e
M
e
a
su
re
s
f
o
r
T
u
n
in
g
S
V
M
Hy
p
e
rp
a
ra
m
e
t
e
rs"
,
Ne
u
ro
c
o
mp
u
ti
n
g
,
v
o
l
.
5
1
,
p
p
.
4
1
-
5
9
,
2
0
0
3
.
[1
9
]
A
.
S
m
o
la,
e
t
a
l
.
,
"
A
s
y
m
p
to
ti
c
a
ll
y
Op
ti
m
a
l
Ch
o
ice
o
f
ε
-
lo
ss
f
o
r
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
i
n
e
s
"
,
in
ICANN
9
8
,
e
d
:
S
p
r
in
g
e
r,
1
9
9
8
,
p
p
.
1
0
5
-
1
1
0
.
[2
0
]
N.
Cristi
a
n
in
i
a
n
d
J.
S
h
a
w
e
-
T
a
y
lo
r,
"
A
n
In
tro
d
u
c
ti
o
n
t
o
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
in
e
s an
d
o
th
e
r
Ke
rn
e
l
-
b
a
se
d
L
e
a
rn
in
g
M
e
th
o
d
s"
,
Ca
m
b
rid
g
e
Un
iv
e
rsit
y
P
re
ss
,
2
0
0
0
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
5
,
Octo
b
er
2
0
1
8
:
3
3
4
1
–
3
3
4
8
3348
[2
1
]
V
.
V
a
p
n
ik
,
e
t
a
l
.,
"
S
u
p
p
o
rt
V
e
c
to
r
M
e
th
o
d
f
o
r
F
u
n
c
ti
o
n
A
p
p
ro
x
im
a
ti
o
n
,
Re
g
re
ss
io
n
Est
im
a
ti
o
n
,
a
n
d
S
ig
n
a
l
P
r
o
c
e
ss
in
g
"
,
in
Ad
v
a
n
c
e
s in
Ne
u
r
a
l
In
fo
rm
a
t
io
n
Pro
c
e
ss
in
g
S
y
ste
ms
9
,
1
9
9
6
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Ed
y
Fra
d
i
n
a
t
a
re
c
e
iv
e
d
h
is
b
a
c
h
e
lo
r
a
n
d
m
a
ste
r
d
e
g
re
e
a
t
In
stit
u
te
T
e
k
n
o
lo
g
i
S
e
p
u
l
u
h
No
p
e
m
b
e
r
-
S
u
ra
b
a
y
a
(I
T
S
).
Cu
rre
n
tl
y
,
H
e
is
stu
d
y
in
g
f
o
r
P
h
D
p
ro
g
ra
m
a
t
P
rin
c
e
o
f
S
o
n
g
k
la
Un
iv
e
rsit
y
,
Ha
t
y
a
i,
T
h
a
il
a
n
d
.
He
h
a
s
so
m
e
w
o
rk
e
x
p
e
rin
c
e
s
a
t
P
e
tr
o
c
h
e
m
ica
l
p
lan
t,
c
h
e
m
ica
l,
ING
O/U
Ns
.
In
tere
stin
g
re
se
a
rc
h
s
a
re
in
Op
ti
m
a
li
z
a
ti
o
n
li
n
e
a
r
a
n
d
n
o
n
-
li
n
e
a
r,
Da
ta
m
i
n
in
g
a
n
d
h
e
u
risti
c
,
b
ig
d
a
ta,
S
CM
,
P
e
rf
.
M
g
n
t,
M
CDM,
c
h
e
m
ica
l
p
ro
c
e
ss
a
n
d
m
a
n
u
f
a
c
tu
re
in
d
u
stry
,
G
IS
,
e
tc.
S
a
k
e
s
u
n
S
u
t
h
u
m
m
a
n
o
n
re
c
e
i
v
e
d
h
is
M
.
B.
A
(Bu
sin
e
ss
A
d
m
i
n
istratio
n
),
P
rin
c
e
o
f
S
o
n
g
k
la
Un
iv
e
rsit
y
,
B.
En
g
.
(I
n
d
u
strial
E
n
g
in
e
e
rin
g
),
M
a
h
a
Na
k
o
rn
P
r
in
c
e
o
f
S
o
n
g
k
la
Un
iv
e
rsity
,
th
e
n
h
e
c
o
n
ti
n
u
e
d
h
is
stu
d
y
f
o
r
M
.
S
c
.
a
n
d
P
h
.
D.
(in
In
d
u
strial
E
n
g
in
e
e
rin
g
),
Un
iv
e
rsit
y
o
f
M
ia
m
i,
F
io
rid
a
,
USA
.
He
in
tere
stin
g
re
se
a
r
c
h
o
n
En
g
in
e
e
rin
g
Eco
n
o
m
ics
,
P
r
o
d
u
c
ti
o
n
a
n
d
Op
e
ra
ti
o
n
s
M
a
n
a
g
e
m
e
n
t,
Qu
a
li
ty
M
a
n
a
g
e
m
e
n
t,
L
o
g
isti
c
s an
d
S
u
p
p
ly
Ch
a
in
M
a
n
a
g
e
m
e
n
t,
e
tc
.
Wa
n
n
a
r
a
t
S
u
n
tia
m
o
r
n
t
u
t
,
h
e
w
a
s
f
ro
m
1
A
p
ril
1
9
9
8
-
3
0
Ju
n
e
1
9
9
9
Re
se
a
rc
h
e
r
a
t
Em
b
e
d
d
e
d
S
y
st
e
m
L
a
b
,
Co
m
p
u
ter
En
g
in
e
e
rin
g
De
p
t.
KMIT
L
1
A
p
ril
1
9
9
8
–
2
0
0
0
M
a
ste
r.
En
g
(c
o
m
)
a
t
Ch
u
lalo
n
g
k
ro
n
U
n
iv
e
rsity
.
1
Au
g
u
st
1
9
9
9
–
P
re
se
n
t,
L
e
c
tu
re
r,
Co
m
p
u
ter
En
g
in
e
e
ri
n
g
De
p
t.
P
ri
n
c
e
o
f
S
o
n
g
k
la
Un
iv
e
rsit
y
.
1
Ja
n
u
a
ry
2
0
0
2
w
a
s
a
s
P
h
.
D.
a
t
Un
iv
e
rsity
o
f
M
a
n
c
h
e
ste
r,
"
L
o
w
-
P
o
w
e
r
A
s
y
n
c
h
ro
n
o
u
s
Dig
it
a
l
S
i
g
n
a
l
P
r
o
c
e
ss
o
r.
No
w
,
sh
e
is
De
p
a
rtm
e
n
t
He
a
d
a
n
d
A
ss
o
c
iate
De
p
a
rtme
n
t
He
a
d
f
o
r
S
tu
d
e
n
t
Aff
a
irs
a
t
P
rin
c
e
o
f
S
o
n
g
k
la
Un
iv
e
rsity
.
S
h
e
is
in
tere
stin
g
re
se
a
rc
h
in
De
sig
n
a
n
d
V
e
rif
ica
ti
o
n
M
icr
o
p
r
o
c
e
ss
o
r
Us
in
g
V
HD
L
o
n
F
P
GA
,
Tes
ti
n
g
a
n
d
V
e
rif
ica
ti
o
n
,
A
s
y
n
c
h
ro
n
o
u
s De
sig
n
a
n
d
L
o
w
-
P
o
w
e
r
Circu
it
De
sig
n
,
e
tc.
Evaluation Warning : The document was created with Spire.PDF for Python.