I
AE
S
I
nte
rna
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
9
,
No
.
3
,
Sep
tem
b
er
2020
,
p
p
.
480
~
487
I
SS
N:
2252
-
8938
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
ai.
v
9
.i
3
.
p
p
480
-
4
8
7
480
Jou
r
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Applica
tion o
f
art
ificial neura
l
net
w
o
rk
t
o
pr
edict
am
o
unt
of
ca
rried w
eig
ht
o
f
ca
rg
o
t
ra
in in rail
t
ra
nspo
rtation
sy
ste
m
Siti
Na
s
uh
a
Z
u
bir
1
, S
.
Sa
rif
a
h Ra
dia
h Sha
rif
f
2
,
Siti
M
er
i
a
m
Z
a
ha
ri
3
1,
2,
3
Ce
n
tre f
o
r
S
tatisti
c
s a
n
d
De
c
isio
n
S
c
ie
n
c
e
S
tu
d
ies
,
F
a
c
u
lt
y
o
f
Co
m
p
u
ter an
d
M
a
th
e
m
a
ti
c
a
l
S
c
ie
n
c
e
s,
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
A
RA
,
S
h
a
h
A
la
m
,
S
e
lan
g
o
r,
M
a
lay
sia
2
M
a
la
y
sia
In
stit
u
te o
f
T
ra
n
sp
o
rt,
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
A
R
A
,
S
h
a
h
A
la
m
,
S
e
lan
g
o
r
,
M
a
lay
sia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
F
eb
1
,
2
0
20
Re
v
i
s
ed
A
pr
2
2
,
2
0
20
A
cc
ep
ted
M
a
y
24
,
2
0
20
De
ra
il
m
e
n
ts
o
f
c
a
rg
o
h
a
v
e
f
re
q
u
e
n
tl
y
o
c
c
u
rre
d
in
M
a
la
y
sia
n
trai
n
se
rv
ice
s
d
u
ri
n
g
th
e
las
t
d
e
c
a
d
e
.
M
a
n
y
f
a
c
to
rs
c
o
n
tri
b
u
te
to
th
is
in
c
id
e
n
t
,
e
sp
e
c
iall
y
it
s
to
tal
a
m
o
u
n
t
o
f
c
a
rried
w
e
ig
h
t.
It
is
f
o
u
n
d
t
h
a
t
se
v
e
re
d
e
ra
il
m
e
n
ts
c
a
u
se
d
a
m
a
g
e
to
b
o
th
li
v
e
s
a
n
d
p
ro
p
e
rti
e
s
e
v
e
r
y
y
e
a
r.
If
th
e
a
m
o
u
n
t
o
f
c
a
rried
w
e
ig
h
t
o
f
c
a
r
g
o
train
c
o
u
l
d
b
e
a
c
c
u
ra
tely
f
o
re
c
a
ste
d
in
a
d
v
a
n
c
e
,
th
e
n
it
s
d
e
tri
m
e
n
tal
e
ff
e
c
t
c
o
u
ld
b
e
g
re
a
tl
y
m
in
i
m
ize
d
.
T
h
is
p
a
p
e
r
p
re
se
n
ts
t
h
e
a
p
p
li
c
a
ti
o
n
o
f
A
rti
f
icia
l
Ne
u
ra
l
Ne
t
w
o
rk
(A
NN
)
to
p
re
d
ict
th
e
a
m
o
u
n
t
o
f
c
a
rried
w
e
i
g
h
t
o
f
c
a
rg
o
train
,
w
it
h
K
T
M
B
u
se
d
a
s
th
e
stu
d
y
c
a
s
e
.
A
s
th
e
re
a
r
e
m
a
n
y
ty
p
e
s
o
f
c
a
rg
o
b
e
in
g
c
a
rried
b
y
K
T
M
B,
th
is
stu
d
y
f
o
c
u
se
s
o
n
l
y
o
n
c
e
m
e
n
t
th
a
t
b
e
in
g
c
a
rried
in
tw
e
lv
e
(1
2
)
d
if
fe
re
n
t
ro
u
tes
.
In
th
is
stu
d
y
,
A
rti
f
icia
l
N
e
u
ra
l
Ne
t
w
o
rk
(
A
N
N)
h
a
s
b
e
e
n
in
c
o
rp
o
ra
ted
f
o
r
d
e
v
e
lo
p
i
n
g
a
p
re
d
ictiv
e
m
o
d
e
l
w
it
h
th
re
e
(
3
)
d
if
f
e
re
n
t
train
in
g
a
lg
o
rit
h
m
s,
Lev
e
n
b
e
rg
-
M
a
rq
u
a
rd
t
(L
M
),
Qu
ick
P
r
o
p
a
g
a
ti
o
n
(Q
P
)
a
n
d
Co
n
ju
g
a
te
G
r
a
d
ien
t
De
sc
e
n
t
(CG
D).
T
h
e
b
e
st
train
in
g
a
l
g
o
rit
h
m
is
se
lec
ted
to
p
re
d
ict
th
e
a
m
o
u
n
t
o
f
c
a
rried
we
ig
h
t
b
y
c
o
m
p
a
rin
g
th
e
e
rro
r
m
e
a
su
re
s
o
f
a
ll
th
e
train
i
n
g
a
lg
o
rit
h
m
w
h
ich
a
re
Ro
o
t
M
e
a
n
S
q
u
a
re
d
Err
o
r
(RM
S
E)
a
n
d
M
e
a
n
A
b
so
lu
te
P
e
rc
e
n
tag
e
Err
o
r
(M
A
P
E).
T
h
e
o
b
tain
e
d
re
su
lt
s
in
d
ica
ted
th
a
t
t
h
e
A
N
N
tec
h
n
iq
u
e
is
su
it
a
b
le
f
o
r
p
re
d
ictin
g
th
e
a
m
o
u
n
t
o
f
c
a
rried
w
e
i
g
h
t.
K
ey
w
o
r
d
s
:
A
r
ti
f
icial
n
e
u
r
al
n
et
w
o
r
k
C
ar
r
ied
w
ei
g
h
t
C
o
n
j
u
g
a
te
g
r
ad
ien
t d
esce
n
t
L
e
v
en
b
er
g
-
m
ar
q
u
ar
d
t
Qu
ic
k
p
r
o
p
ag
atio
n
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
S.
Sar
if
a
h
R
ad
ia
h
Sh
ar
i
f
f
,
Ma
la
y
s
ia
I
n
s
ti
tu
te
o
f
T
r
an
s
p
o
r
t,
Un
i
v
er
s
iti T
ek
n
o
lo
g
i M
A
R
A
,
Sh
a
h
A
la
m
,
Selan
g
o
r
,
Ma
la
y
s
i
a
.
E
m
ail: s
h
ar
i9
9
0
@
u
it
m
.
ed
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
C
ar
g
o
o
r
f
r
eig
h
t
r
ef
er
s
to
g
o
o
d
s
o
r
p
r
o
d
u
ct
th
at
ar
e
tr
an
s
f
er
r
ed
o
f
d
is
tr
ib
u
ted
g
e
n
er
all
y
f
o
r
co
m
m
er
cial
g
ai
n
.
No
w
ad
a
y
s
,
ca
r
g
o
tr
an
s
p
o
r
t
ca
n
b
e
ca
r
r
ied
o
n
w
ater
,
air
o
r
lan
d
.
Mo
s
t
wid
el
y
u
s
ed
to
ca
r
r
y
ca
r
g
o
is
r
o
ad
tr
an
s
p
o
r
t.
Dif
f
er
en
t
f
o
r
m
o
f
w
e
ig
h
t
a
n
d
v
eh
ic
le
ar
e
u
s
ed
to
tr
an
s
p
o
r
t
ca
r
g
o
ar
o
u
n
d
.
R
o
ad
tr
an
s
p
o
r
t
h
a
s
m
an
y
ad
v
an
tag
e
s
li
k
e
it
ca
n
d
o
d
o
o
r
to
d
o
o
r
d
eliv
er
y
o
n
to
p
o
f
h
a
v
i
n
g
s
e
v
er
al
t
y
p
e
o
f
v
eh
ic
les
li
k
e
tr
u
ck
s
,
b
u
s
s
es,
lo
r
r
y
,
ca
r
s
an
d
s
o
o
n
.
Ho
w
e
v
er
,
s
o
m
e
b
u
l
k
y
ite
m
s
li
k
e
s
u
g
ar
,
ce
m
en
ts
,
ch
ar
co
als
th
at
n
ee
d
to
b
e
tr
an
s
f
er
r
ed
in
l
ar
g
e
v
o
lu
m
e
ar
e
m
o
v
ed
u
s
in
g
tr
ain
o
r
r
ail
tr
an
s
p
o
r
t.
Oth
er
t
h
a
n
k
n
o
w
n
as
ab
le
to
ca
r
r
y
p
a
s
s
e
n
g
er
s
,
tr
ain
is
a
ls
o
ca
p
ab
le
o
f
tr
an
s
p
o
r
tin
g
lar
g
e
v
o
lu
m
e
o
f
ite
m
s
s
u
c
h
as
w
ater
,
ce
m
e
n
t,
s
teel,
w
o
o
d
an
d
co
al.
Gen
er
ally
,
tr
ain
ca
r
g
o
h
a
s
a
d
ir
ec
t
r
o
u
te
to
its
d
esti
n
atio
n
.
Un
d
er
th
e
r
ig
h
t
co
n
d
itio
n
,
ca
r
g
o
tr
an
s
p
o
r
t
b
y
r
ail
is
m
o
r
e
e
co
n
o
m
ic
an
d
m
o
r
e
p
r
o
d
u
ctiv
e
co
m
p
ar
ed
to
r
o
a
d
tr
an
s
p
o
r
t,
esp
ec
iall
y
w
h
en
tr
a
n
s
p
o
r
tin
g
ite
m
s
i
n
lar
g
e
v
o
l
u
m
e
o
v
er
lo
n
g
d
is
tan
ce
.
T
h
e
ch
o
ice
o
f
m
o
d
e
o
f
tr
an
s
p
o
r
tatio
n
d
ep
en
d
s
v
er
y
m
u
c
h
o
n
ca
r
r
ied
w
ei
g
h
t.
C
ar
r
ied
w
ei
g
h
t
is
an
i
m
p
o
r
ta
n
t
m
atter
i
n
tr
a
n
s
p
o
r
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
p
p
lica
ti
o
n
o
f a
r
tifi
cia
l n
eu
r
a
l n
etw
o
r
k
to
p
r
ed
ict
a
mo
u
n
t
o
f…
(
S
iti N
a
s
u
h
a
Zu
b
ir
)
481
s
y
s
te
m
an
d
n
ee
d
to
b
e
co
n
s
id
er
ed
.
I
n
th
e
lo
g
is
tic
tr
a
n
s
p
o
r
tatio
n
s
y
s
te
m
,
th
e
a
m
o
u
n
t
o
f
w
e
ig
h
t
ca
r
r
ied
is
v
er
y
i
m
p
o
r
tan
t
to
en
s
u
r
e
th
a
t
all
th
e
g
o
o
d
s
ar
r
iv
e
s
a
f
el
y
at
t
h
e
d
esti
n
atio
n
i
n
ti
m
e.
Usi
n
g
tr
ain
as
m
o
d
e
o
f
tr
an
s
p
o
r
tatio
n
i
s
b
en
e
f
icia
l
to
th
e
e
n
v
ir
o
n
m
en
t
as
it
i
s
li
m
itin
g
g
r
ee
n
h
o
u
s
e
g
a
s
e
m
i
s
s
io
n
s
,
i
n
cr
ea
s
i
n
g
f
u
el
ef
f
icien
c
y
a
n
d
r
ed
u
cin
g
its
ca
r
b
o
n
f
o
o
tp
r
in
t
[
1
]
.
KT
MB
Fre
ig
h
t
Ser
v
ice
h
as
th
r
ee
t
y
p
e
s
o
f
tr
ai
n
s
er
v
ices
:
T
r
ain
C
o
n
te
n
a
Ser
v
ice,
T
r
ain
C
ar
g
o
C
o
n
v
en
t
io
n
al
Ser
v
ice
an
d
T
r
ain
L
an
d
b
r
id
g
e
Ser
v
ice.
I
n
2
0
1
7
,
K
T
MB
h
as
ex
p
er
ien
ce
d
th
r
ee
m
aj
o
r
d
er
ail
m
en
t
s
.
O
n
Au
g
u
s
t
2
1
,
2
0
1
7
,
a
ca
r
g
o
tr
ain
cr
a
s
h
ed
at
J
alan
K
u
cin
g
,
ca
u
s
in
g
d
ela
y
s
f
o
r
a
f
e
w
d
a
y
s
[
2
]
.
On
Sep
te
m
b
er
2
3
,
2
0
1
7
,
K
T
MB
’
s
ca
r
g
o
tr
ain
s
n
ap
p
ed
elec
tr
ical
ca
b
le
b
etw
ee
n
R
a
w
a
n
g
an
d
Ku
an
g
s
tatio
n
s
an
d
f
o
r
cin
g
KT
MB
to
clo
s
e
a
ll
tr
ac
k
s
f
o
r
t
w
o
d
a
y
s
[
3
]
.
On
No
v
e
m
b
er
2
3
,
2
0
1
7
,
o
n
ce
ag
ain
a
n
o
th
er
ca
r
g
o
tr
ain
ac
cid
en
t
s
o
cc
u
r
r
ed
w
h
e
n
t
w
el
v
e
ca
r
g
o
tr
ai
n
s
tr
a
v
elli
n
g
s
o
u
t
h
w
a
r
d
b
et
w
ee
n
t
h
e
N
atio
n
al
B
an
k
Statio
n
an
d
Ku
a
la
L
u
m
p
u
r
Statio
n
s
l
ip
p
ed
d
u
e
to
h
ea
v
y
w
ei
g
h
t
a
n
d
o
v
er
s
ized
lo
ad
s
ca
r
r
ied
b
y
t
h
e
ca
r
g
o
tr
ai
n
s
.
As
a
r
esu
lt,
KT
M
an
d
E
T
S
s
er
v
ices
w
er
e
d
is
r
u
p
ted
o
n
s
e
v
er
al
r
o
u
tes
ar
o
u
n
d
th
e
K
lan
g
Valle
y
.
O
n
e
o
f
t
h
e
m
aj
o
r
ca
u
s
es
o
f
th
i
s
tr
ag
ed
y
is
th
e
o
v
er
lo
ad
in
g
o
f
th
e
ca
r
g
o
tr
ain
’
s
w
a
g
o
n
[
4
]
.
I
n
r
ec
en
t a
cc
id
en
t th
a
t o
cc
u
r
r
ed
o
n
2
1
J
u
l
y
2
0
1
9
ca
r
g
o
tr
ain
th
at
ca
r
r
ied
3
0
w
a
g
o
n
s
o
f
ce
m
en
t.
Du
r
in
g
th
e
d
er
ail
m
en
t,
KT
M
B
n
ee
d
ed
to
r
elo
ca
te
all
th
e
w
ag
o
n
s
as
s
o
o
n
as
p
o
s
s
ib
le
b
ec
au
s
e
all
th
e
KT
MB
’
s
s
er
v
ice
s
w
er
e
ef
f
ec
ted
[
5
]
.
T
h
e
d
er
ailm
e
n
t
h
ap
p
en
ed
d
u
e
to
m
a
n
y
f
ac
to
r
s
an
d
o
n
e
o
f
th
e
m
o
s
t
s
i
g
n
if
ican
t
f
ac
t
o
r
s
is
th
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t.
Ha
v
i
n
g
th
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
eig
h
t
p
lan
n
ed
to
m
atc
h
t
h
e
tr
ac
k
c
ap
ab
ilit
y
ca
n
a
v
o
id
d
er
ail
m
en
t
o
cc
u
r
r
en
ce
s
.
A
r
tific
ial
Neu
r
al
Net
w
o
r
k
(
ANN)
is
a
p
o
p
u
lar
m
et
h
o
d
u
s
ed
b
y
o
th
er
p
r
ev
io
u
s
r
esear
ch
er
s
to
p
r
ed
ict
ca
r
r
ied
w
ei
g
h
t.
I
n
t
h
is
s
tu
d
y
,
t
h
e
ca
r
g
o
tr
ain
ca
r
r
ied
w
e
ig
h
t
w
il
l b
e
p
r
ed
icted
.
T
h
e
p
r
ev
io
u
s
r
esear
ch
o
u
tco
m
es
d
e
m
o
n
s
tr
ated
t
h
at
th
e
A
NN
is
a
n
e
f
f
i
c
i
en
t
o
p
tio
n
s
tr
ateg
y
i
n
p
r
ed
ictio
n
[6
-
10]
.
T
h
is
i
s
s
u
p
p
o
r
ted
b
y
[
1
1
-
12]
w
h
o
p
r
o
p
o
s
ed
th
at
A
N
N
is
th
e
b
est
m
o
d
el
co
m
p
ar
ed
to
A
d
ap
tiv
e
Neu
r
o
-
Fu
zz
y
I
n
f
er
en
ce
S
y
s
te
m
(
A
N
FIS)
.
T
h
is
s
tu
d
y
co
m
p
ar
ed
th
e
m
o
d
els
w
it
h
Am
er
ica
n
C
o
n
cr
ete
I
n
s
tit
u
te
an
d
I
r
an
ian
C
o
n
cr
et
e
I
n
s
tit
u
t
e
e
m
p
ir
ical
co
d
es.
A
s
a
r
es
u
l
t,
th
e
p
r
ed
ictio
n
o
f
A
N
N
is
b
etter
t
h
an
ANFI
S
m
o
d
el.
I
n
[
1
3
]
d
ev
elo
p
ed
a
d
ec
i
s
io
n
s
u
p
p
o
r
t
s
y
s
te
m
th
a
t
ca
n
f
o
r
ec
ast
d
e
m
an
d
i
n
elec
tr
o
n
ic
r
etails
in
d
u
s
tr
y
at
T
u
r
k
e
y
b
y
u
s
in
g
A
N
N
tec
h
n
iq
u
es
s
u
ch
a
s
Gr
ad
ien
t
Desce
n
t
(
GD)
,
th
e
C
o
n
j
u
g
a
te
Gr
ad
ien
t
Descen
t
(
GC
D)
,
Qu
ick
P
r
o
p
ag
atio
n
(
QP
)
an
d
L
M
m
eth
o
d
s
.
Ho
w
ev
er
,
in
m
u
lti
–
s
tag
e
s
u
p
p
l
y
–
ch
ain
ar
ea
t
h
e
ap
p
licatio
n
o
f
t
h
ese
ar
ti
f
icial
tec
h
n
iq
u
e
s
t
ill h
av
e
s
e
v
er
e
lack
.
T
h
er
e
ar
e
m
o
r
e
s
t
u
d
ies
f
o
cu
s
in
g
o
n
h
o
w
th
e
p
r
ed
icti
v
e
ab
ilit
y
ca
n
b
e
i
n
f
lu
e
n
ce
d
b
y
t
h
e
tr
ain
i
n
g
an
d
test
i
n
g
alg
o
r
it
h
m
.
A
cc
o
r
d
in
g
to
[
1
4
]
in
th
e
ir
s
t
u
d
y
,
A
NN
is
u
s
ed
to
p
r
ed
ict
ca
r
r
ied
w
eig
h
t
a
n
d
th
r
ee
(
3
)
class
es
o
f
A
NN
ar
e
u
s
ed
w
h
ic
h
ar
e
in
cr
e
m
en
ta
l
b
ac
k
p
r
o
p
ag
atio
n
alg
o
r
ith
m
(
I
B
P
)
,
Gen
etic
al
g
o
r
ith
m
(
G
A
)
a
n
d
L
ev
en
b
er
g
-
Ma
r
q
u
ar
d
t a
l
g
o
r
ith
m
(
L
M)
.
T
h
e
p
r
ed
ictin
g
p
er
f
o
r
m
an
ce
o
f
t
h
e
t
h
r
ee
alg
o
r
ith
m
w
as
co
m
p
ar
ed
.
T
h
is
s
tu
d
y
w
a
s
ap
p
lied
i
n
a
n
a
u
t
o
m
o
b
ile
in
d
u
s
tr
y
,
I
r
an
K
h
o
d
r
o
C
o
m
p
a
n
y
(
I
KC
O)
as
to
ap
p
r
o
p
r
iately
p
r
o
v
id
e
th
e
m
ac
h
in
er
y
r
eso
u
r
ce
s
,
l
ab
o
r
an
d
tr
an
s
p
o
r
t
s
y
s
te
m
d
em
a
n
d
.
A
NN
w
as
u
s
ed
to
te
s
t
t
h
e
w
ee
k
l
y
d
at
a
o
f
ca
r
r
ied
w
ei
g
h
t
b
ased
o
n
t
h
e
o
b
s
er
v
atio
n
o
f
t
h
e
n
u
m
b
er
o
f
v
eh
ic
les
an
d
f
u
e
l
co
n
s
u
m
p
tio
n
.
A
t
th
e
en
d
o
f
th
e
s
t
u
d
y
,
I
B
P
g
iv
e
th
e
o
p
ti
m
u
m
tr
ain
in
g
alg
o
r
it
h
m
.
As
f
o
r
i
m
p
r
o
v
e
m
e
n
t,
[
1
5
]
u
s
ed
th
e
s
a
m
e
v
ar
iab
le
as
t
h
e
p
r
ev
io
u
s
r
esear
c
h
to
p
r
ed
ict
th
e
ca
r
r
ied
w
e
ig
h
t.
I
n
s
tead
o
f
u
s
i
n
g
G
A
a
n
d
I
B
P
,
Qu
ic
k
P
r
o
p
ag
atio
n
(
QP
)
an
d
B
atch
B
ac
k
p
r
o
p
ag
atio
n
(
B
B
P
)
ar
e
u
s
ed
a
n
d
QP
ex
h
ib
it
s
th
e
b
etter
p
er
f
o
r
m
a
n
ce
.
Hen
ce
,
th
i
s
p
ap
er
p
r
esen
ts
t
h
e
ap
p
licatio
n
o
f
A
r
ti
f
ici
al
Neu
r
al
Net
w
o
r
k
(
A
NN)
to
p
r
ed
ict
th
e
am
o
u
n
t
o
f
ca
r
r
ied
o
f
ca
r
g
o
tr
ain
,
u
s
in
g
t
h
r
ee
tr
ain
i
n
g
al
g
o
r
ith
m
s
:
L
e
v
en
b
er
g
-
Ma
r
q
u
ar
d
t
alg
o
r
ith
m
(
L
M)
as
a
w
ell
p
er
f
o
r
m
ed
alg
o
r
it
h
m
to
p
r
ed
ict
d
if
f
er
en
t
s
et
o
f
ca
r
r
ied
w
ei
g
h
t
d
ata,
C
o
n
j
u
g
ate
Gr
ad
ien
t
Desce
n
t
(
GC
D)
as
a
w
el
l
p
er
f
o
r
m
al
g
o
r
ith
m
f
o
r
p
r
ed
ict
io
n
o
f
o
th
er
s
et
s
o
f
d
ata
an
d
Qu
ick
P
r
o
p
ag
atio
n
(
QP
)
as a
n
e
w
al
g
o
r
ith
m
u
s
ed
t
o
p
r
ed
ict
ca
r
r
ied
w
ei
g
h
t.
2.
RE
S
E
ARCH
M
E
T
H
O
D
S
A
N
N
is
a
m
a
th
e
m
atica
l
m
o
d
el
o
r
co
m
p
u
tatio
n
a
l
m
o
d
el
b
ased
o
n
t
h
e
n
e
u
r
al
n
et
w
o
r
k
s
o
r
ca
lled
an
i
m
itatio
n
o
f
b
i
o
lo
g
ical
n
e
u
r
al
s
y
s
te
m
.
I
t
is
a
n
ad
ap
tiv
e
s
y
s
te
m
as
it
co
u
ld
m
o
d
i
f
y
th
e
s
tr
u
ct
u
r
e
b
ased
o
n
th
e
i
n
f
o
r
m
atio
n
eith
er
i
n
te
r
n
al
o
r
ex
ter
n
al
t
h
at
f
lo
w
t
h
r
o
u
g
h
th
e
n
et
w
o
r
k
[
1
6
]
.
T
h
is
m
o
d
el
i
s
a
f
lex
ib
l
e
co
m
p
u
ti
n
g
f
r
a
m
e
w
o
r
k
an
d
a
u
n
iv
er
s
al
ap
p
r
o
x
i
m
ato
r
.
I
t
c
an
b
e
ap
p
lied
to
a
w
id
e
r
an
g
e
o
f
p
r
o
b
lem
lik
e
a
ti
m
e
s
er
ies
f
o
r
ec
asti
n
g
w
it
h
a
h
i
g
h
d
eg
r
ee
o
f
ac
c
u
r
ac
y
.
ANN
r
ep
licates
t
h
e
b
io
lo
g
ical
n
eu
r
o
n
s
tr
u
ctu
r
e
b
y
cr
ea
tin
g
a
s
i
m
p
le
p
r
o
ce
s
s
i
n
g
u
n
it
ca
lled
ar
ti
f
icial
n
eu
r
o
n
s
.
A
n
ap
p
r
o
x
i
m
atio
n
o
f
th
e
3
-
d
i
m
en
s
io
n
a
l
in
ter
co
o
n
ec
ted
n
e
s
s
o
f
b
io
lo
g
i
ca
l
n
e
u
r
o
n
es
is
d
o
n
e
i
n
A
NN
b
y
m
ea
n
s
o
f
th
e
u
s
ag
e
o
f
la
y
e
r
s
.
Fi
g
u
r
e
1
s
h
o
ws
an
ANN
w
ith
i
n
p
u
t
n
o
d
es,
h
id
d
en
n
o
d
es,
an
d
o
n
e
o
u
tp
u
t
n
o
d
e.
T
h
e
h
id
d
en
n
o
d
es
w
ill
b
e
g
en
er
ated
u
s
in
g
th
e
d
if
f
er
e
n
t b
u
il
t
-
i
n
al
g
o
r
ith
m
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
4
8
0
–
487
482
Fig
u
r
e
1.
A
r
ti
f
ici
al
n
eu
r
al
n
et
w
o
r
k
m
o
d
el
(
A
N
N)
2
.
1
.
T
ra
ini
ng
a
lg
o
rit
hm
T
h
r
ee
b
u
ilt in
tr
ain
i
n
g
alg
o
r
it
h
m
s
ar
e
u
s
ed
an
d
co
m
p
ar
ed
.
a.
L
e
v
en
b
er
g
-
Ma
r
q
u
ar
d
t (
L
M)
I
t
is
a
h
ig
h
er
-
o
r
d
er
ad
ap
ti
v
e
alg
o
r
ith
m
a
n
d
it
m
in
i
m
izes
t
h
e
Me
an
Sq
u
ar
e
E
r
r
o
r
o
f
a
n
eu
r
al
n
et
w
o
r
k
[
1
7
]
.
L
M
alg
o
r
it
h
m
i
s
a
v
ar
iatio
n
o
f
Ne
w
to
n
’
s
m
e
th
o
d
th
at
i
s
d
esig
n
ed
f
o
r
m
i
n
i
m
izi
n
g
f
u
n
ctio
n
s
t
h
at
ar
e
s
u
m
s
o
f
s
q
u
ar
es
o
f
o
th
er
n
o
n
li
n
ea
r
f
u
n
ctio
n
s
.
L
M
al
g
o
r
ith
m
p
r
o
v
id
es
n
u
m
er
ical
s
o
l
u
tio
n
to
m
i
n
i
m
ized
n
o
n
-
li
n
ea
r
f
u
n
ctio
n
.
T
h
e
(
n
o
n
-
n
e
g
ati
v
e)
d
a
m
p
in
g
p
ar
a
m
eter
is
ad
j
u
s
t
ed
in
e
v
er
y
i
ter
atio
n
,
w
h
er
e
s
m
al
l
v
a
lu
e
s
o
f
th
e
al
g
o
r
ith
m
i
c
p
ar
am
e
ter
λ
r
es
u
lt
i
n
Ga
u
s
s
-
Ne
w
to
n
u
p
d
ate,
an
d
lar
g
e
v
a
l
u
es
o
f
λ
r
es
u
lt
i
n
a
g
r
ad
ien
t
d
esce
n
t
u
p
d
ate.
T
h
e
p
ar
am
eter
λ
i
s
in
itialized
to
b
e
lar
g
e
s
o
th
at
f
ir
s
t
u
p
d
ates
ar
e
s
m
all
s
tep
s
i
n
th
e
s
teep
est
d
esce
n
t
d
ir
ec
tio
n
.
I
f
an
y
iter
atio
n
h
ap
p
en
s
to
lead
to
a
p
o
o
r
ap
p
r
o
x
i
m
atio
n
,
t
h
en
λ
i
s
i
n
cr
ea
s
ed
.
T
h
er
ef
o
r
e,
f
o
r
lar
g
e
v
al
u
es
o
f
λ
,
th
e
s
tep
w
ill
b
e
ta
k
en
ap
p
r
o
x
i
m
atel
y
i
n
th
e
d
ir
ec
ti
o
n
o
f
t
h
e
g
r
ad
ien
t.
Oth
er
w
i
s
e,
as
th
e
s
o
lu
t
io
n
i
m
p
r
o
v
es,
λ
is
d
ec
r
ea
s
ed
,
th
e
L
M
m
et
h
o
d
ap
p
r
o
ac
h
es
th
e
Ga
u
s
s
-
Ne
w
to
n
m
et
h
o
d
,
an
d
th
e
s
o
l
u
tio
n
t
y
p
icall
y
ac
ce
ler
ates to
th
e
lo
ca
l
m
i
n
i
m
u
m
.
b.
C
o
n
j
u
g
a
te
Gr
ad
ien
t D
esce
n
t (
C
GD
)
T
h
e
C
GD
m
e
th
o
d
s
o
lv
es
s
y
s
t
e
m
s
o
f
li
n
ea
r
eq
u
atio
n
s
,
also
u
s
ed
to
s
o
lv
e
s
y
s
te
m
w
h
er
e
m
atr
ix
is
n
o
t
s
y
m
m
etr
ic,
n
o
t
p
o
s
it
iv
e
-
d
ef
in
ite,
an
d
s
ti
ll
n
o
t
s
q
u
ar
e
[
1
8
]
.
C
GD
i
s
an
ad
v
a
n
ce
d
m
et
h
o
d
f
o
r
tr
ain
i
n
g
m
u
l
ti
-
la
y
er
n
e
u
r
al
n
et
w
o
r
k
.
I
n
th
e
C
GD
m
et
h
o
d
,
th
e
li
n
e
i
s
n
o
t
s
ea
r
ch
ed
,
b
u
t
a
p
la
n
e
is
s
ea
r
ch
ed
.
A
p
la
n
e
i
s
f
o
r
m
u
lated
f
r
o
m
a
r
an
d
o
m
lin
ea
r
co
m
b
in
at
io
n
o
f
t
w
o
v
ec
to
r
s
.
Fo
r
m
i
n
i
m
izi
n
g
q
u
ad
r
atic
f
u
n
ct
io
n
s
,
t
h
e
p
lan
e
s
ea
r
ch
r
eq
u
ir
es
o
n
l
y
t
h
e
s
o
lu
tio
n
o
f
a
t
w
o
b
y
t
w
o
s
e
ts
o
f
li
n
ea
r
eq
u
at
io
n
f
o
r
α
an
d
β.
So
lv
i
n
g
co
n
v
e
x
o
p
tim
izatio
n
p
r
o
b
le
m
s
u
s
i
n
g
C
GD.
(
)
=
1
2
2
+
1
3
2
+
1
5
(
1
)
Gr
ad
ien
t
Desce
n
t
Me
t
h
o
d
w
ill
tr
y
to
f
i
n
d
th
e
m
i
n
i
m
u
m
b
y
c
o
m
p
u
ti
n
g
th
e
g
r
ad
ien
t
o
f
(
)
at
th
e
in
itial
g
u
e
s
s
.
T
o
ac
h
iev
e
t
h
e
v
al
u
e
o
f
x
clo
s
e
to
o
p
tim
al
s
o
l
u
tio
n
th
e
w
h
o
l
e
p
r
o
ce
s
s
h
as to
iter
ate.
c.
Qu
ic
k
P
r
o
p
ag
atio
n
(
QP
)
T
h
e
Qu
ick
P
r
o
p
ag
atio
n
m
et
h
o
d
u
s
es t
h
e
f
o
llo
w
i
n
g
u
p
d
atin
g
eq
u
atio
n
:
+
1
=
+
(
2
)
W
h
er
e,
=
(
△
−
)
/
(
3
)
△
=
+
1
−
(
4
)
is
th
e
m
o
d
el
r
esp
o
n
s
e
f
o
r
th
e
ith
iter
atio
n
.
T
h
e
ap
p
r
o
x
im
at
io
n
o
f
t
h
e
J
ac
o
b
ian
m
atr
i
x
+
1
f
o
r
th
e
(
+
1
)
ℎ
iter
atio
n
is
ca
lc
u
lated
u
s
in
g
t
h
e
J
ac
o
b
ian
m
atr
ix
ap
p
r
o
x
i
m
atio
n
,
th
e
p
ar
a
m
eter
p
er
tu
r
b
atio
n
v
ec
to
r
an
d
th
e
ch
a
n
g
e
in
t
h
e
m
o
d
el
r
esp
o
n
s
e
△
f
o
r
th
e
ith
iter
atio
n
.
T
h
e
u
p
d
atin
g
m
atr
ix
is
a
r
an
k
o
n
e
m
atr
i
x
an
d
B
r
o
y
d
en
'
s
m
et
h
o
d
is
a
r
an
k
-
o
n
e
q
u
ick
p
r
o
p
ag
atio
n
m
et
h
o
d
.
T
h
e
alg
o
r
ith
m
cla
s
s
i
f
ied
to
th
e
g
r
o
u
p
o
f
th
e
s
ec
o
n
d
o
r
d
er
lear
n
in
g
m
et
h
o
d
w
h
ic
h
i
s
it
f
o
llo
w
s
a
q
u
ad
r
atic
ap
p
r
o
x
i
m
atio
n
o
f
th
e
p
r
ev
io
u
s
g
r
ad
ie
n
t
s
tep
an
d
th
e
cu
r
r
en
t g
r
ad
ien
t
[
1
9
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
p
p
lica
ti
o
n
o
f a
r
tifi
cia
l n
eu
r
a
l n
etw
o
r
k
to
p
r
ed
ict
a
mo
u
n
t
o
f…
(
S
iti N
a
s
u
h
a
Zu
b
ir
)
483
2
.
2
.
E
rr
o
r
m
ea
s
ures
A
cc
o
r
d
in
g
to
[
2
0
]
,
f
o
r
ec
asti
n
g
er
r
o
r
is
ab
o
u
t
m
ea
s
u
r
i
n
g
h
o
w
g
o
o
d
th
e
p
er
f
o
r
m
a
n
ce
o
f
a
m
o
d
el
its
el
f
co
m
p
ar
es to
t
h
e
o
n
e
o
f
u
s
i
n
g
t
h
e
p
ast d
ata.
a.
R
o
o
t M
ea
n
Sq
u
ar
ed
E
r
r
o
r
(
R
MSE
)
b.
Me
an
A
b
s
o
lu
te
P
er
ce
n
tag
e
E
r
r
o
r
(
MA
P
E
)
2
.
3
.
M
o
del v
a
lid
a
t
io
n
T
h
e
f
ir
s
t
s
ta
g
e
i
s
ca
lled
i
n
itia
l
d
ata
p
r
ep
ar
atio
n
.
Du
r
in
g
th
e
f
ir
s
t
s
ta
g
e,
th
e
d
ata
s
er
ies
w
i
ll
d
iv
id
ed
i
n
to
t
w
o
p
ar
ts
.
T
h
e
f
ir
s
t p
ar
t
k
n
o
w
n
as
w
ith
in
s
a
m
p
les o
r
f
itt
in
g
p
ar
ts
t
h
at
u
s
ed
to
est
i
m
ate
th
e
p
er
f
o
r
m
an
ce
o
f
f
o
r
ec
asti
n
g
m
o
d
el
[
2
1
]
.
Me
an
w
h
ile,
t
h
e
s
ec
o
n
d
p
ar
t
is
t
o
ev
alu
a
te
t
h
e
m
o
d
el
ca
lled
as
o
u
t
s
a
m
p
les
o
r
ev
alu
a
tio
n
p
ar
t.
I
n
th
is
s
t
u
d
y
,
th
e
d
ata
ar
e
p
ar
titi
o
n
ed
in
to
7
0
%
f
o
r
tr
ain
in
g
p
ar
t
w
h
e
r
e
as
th
e
3
0
%
f
o
r
v
alid
atio
n
p
ar
t.
T
h
er
e
ar
e
1
3
,
1
5
2
o
b
s
er
v
atio
n
.
I
n
th
e
s
ec
o
n
d
s
ta
g
e,
th
e
w
it
h
i
n
s
a
m
p
le
s
tatis
t
ics
is
u
s
ed
to
esti
m
ate
th
e
m
o
d
el
u
s
i
n
g
t
h
r
ee
b
u
ilt
in
alg
o
r
ith
m
s
,
L
M,
C
GD
an
d
QP
.
T
h
e
b
est
esti
m
ati
o
n
ap
p
r
o
ac
h
is
s
elec
ted
b
ased
o
n
th
e
o
u
tc
o
m
e
s
o
f
co
m
p
ar
in
g
th
eir
er
r
o
r
m
ea
s
u
r
e
s
p
er
f
o
r
m
an
ce
s
[
2
2
]
.
Fo
r
t
h
is
p
u
r
p
o
s
e,
R
MSE
a
n
d
M
A
P
E
ar
e
u
s
ed
[
2
3
-
24]
.
T
r
ain
in
g
al
g
o
r
ith
m
w
it
h
t
h
e
s
m
alle
s
t e
r
r
o
r
m
ea
s
u
r
e
is
d
ec
id
ed
to
b
e
a
b
le
to
p
r
o
d
u
ce
th
e
b
est f
it
m
o
d
el
.
Ha
v
i
n
g
co
m
p
leted
t
h
e
f
ir
s
t
a
n
d
s
ec
o
n
d
s
ta
g
es,
t
h
e
la
s
t
s
tag
e
is
to
u
s
e
th
e
b
es
t
f
it
m
o
d
el
t
o
f
o
r
ec
ast
th
e
a
m
o
u
n
t o
f
ca
r
r
ied
w
ei
g
h
t
b
y
ea
c
h
tr
ain
p
er
tr
ip
,
th
at
ca
n
h
elp
KT
MB
to
p
lan
f
o
r
its
f
u
t
u
r
e
o
p
er
atio
n
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
P
r
ed
ictiv
e
m
o
d
elin
g
u
s
i
n
g
A
r
ti
f
icial
Ne
u
r
al
Net
w
o
r
k
w
er
e
ca
r
r
ied
o
u
t
b
y
u
s
i
n
g
Al
y
u
d
a
Neu
r
o
in
tel
lig
e
n
ce
s
o
f
t
w
ar
e.
I
n
th
e
f
ir
s
t
s
ta
g
e,
d
ata
is
tr
ea
ted
f
o
r
its
m
is
s
in
g
v
a
lu
e
s
.
I
n
itia
ll
y
,
th
er
e
w
er
e
1
2
r
o
u
tes.
Sin
ce
,
th
e
m
i
s
s
i
n
g
v
al
u
e
s
f
o
r
s
o
m
e
r
o
u
tes
ar
e
m
o
r
e
th
an
1
5
%
[
2
5
]
,
t
h
en
,
t
h
o
s
e
r
o
u
tes
ar
e
o
m
itted
.
T
h
e
r
em
ai
n
i
n
g
t
w
o
r
o
u
tes
w
h
ich
ar
e
R
o
u
te
1
an
d
R
o
u
te
2
ar
e
f
u
r
th
er
an
al
y
ze
d
an
d
u
n
d
er
w
e
n
t
i
m
p
u
ta
tio
n
p
r
o
ce
s
s
b
y
u
s
i
n
g
I
B
M
SP
SS
Mo
d
eler
1
8
.
0
s
o
f
t
w
ar
e.
T
ab
le
1
s
h
o
w
s
t
h
e
s
u
m
m
ar
y
s
tatis
tic
f
o
r
v
ar
iab
les
in
R
o
u
te
1
an
d
R
o
u
te
2
b
ef
o
r
e
im
p
u
tat
io
n
.
T
h
er
e
ar
e
th
r
ee
co
n
tin
u
o
u
s
an
d
f
o
u
r
ca
te
g
o
r
ical
v
ar
iab
le
s
r
esp
ec
tiv
el
y
.
On
l
y
1
ca
te
g
o
r
ical
v
ar
iab
le
h
as
m
is
s
i
n
g
v
a
lu
e
w
h
ich
is
L
ab
o
r
in
R
o
u
te
1
a
n
d
R
o
u
te
2
.
W
h
ile,
t
h
er
e
ar
e
m
i
s
s
i
n
g
v
al
u
es
f
o
r
all
co
n
ti
n
u
o
u
s
v
ar
iab
le
w
h
ic
h
ar
e
T
o
tal
W
ag
o
n
,
T
o
n
n
ag
e
/KM
an
d
C
ar
r
ied
W
eig
h
t
f
o
r
b
o
th
r
o
u
tes.
T
h
er
e
f
o
r
e,
i
m
p
u
tat
io
n
ar
e
n
ee
d
ed
f
o
r
L
ab
o
r
an
d
T
o
t
al
W
ag
o
n
.
Ho
w
e
v
er
,
T
o
n
n
ag
e/KM
w
ill
n
o
t
u
n
d
er
g
o
i
m
p
u
ta
tio
n
p
r
o
ce
s
s
.
Fo
r
tar
g
et
v
ar
iab
le
w
h
ic
h
is
C
ar
r
ied
W
eig
h
t,
all
t
h
e
ca
s
es
w
it
h
t
h
e
m
i
s
s
i
n
g
v
al
u
e
ar
e
d
is
ca
r
d
e
d
f
o
r
b
o
th
r
o
u
tes.
T
ab
le
1
.
Su
m
m
ar
y
s
ta
tis
tic
s
f
o
r
v
ar
iab
les
S
t
a
t
i
o
n
R
o
u
t
e
1
R
o
u
t
e
2
V
a
r
i
a
b
l
e
Ty
p
e
V
a
l
i
d
M
i
s
i
n
g
V
a
l
i
d
M
i
ss
i
n
g
T
r
a
i
n
n
o
C
a
t
e
g
o
r
i
c
a
l
1
0
9
6
0
1
0
9
6
0
C
o
mp
a
n
y
/
C
u
s
t
o
me
r
C
a
t
e
g
o
r
i
c
a
l
1
0
9
6
0
1
0
9
6
0
D
i
st
a
n
c
e
C
a
t
e
g
o
r
i
c
a
l
1
0
9
6
0
1
0
9
6
0
L
a
b
o
u
r
Ca
t
e
g
o
r
i
c
a
l
9
2
2
1
7
4
8
4
4
2
5
2
T
o
t
a
l
W
a
g
o
n
C
o
n
t
i
n
u
o
u
s
9
0
5
1
9
1
9
0
0
1
9
6
T
o
n
n
a
g
e
/
K
M
C
o
n
t
i
n
u
o
u
s
1
0
0
9
87
9
3
4
1
6
2
C
a
r
r
i
e
d
W
e
i
g
h
t
C
o
n
t
i
n
u
o
u
s
1
0
0
9
87
9
3
4
1
6
2
3
.
1
.
Desig
nin
g
t
he
net
wo
rk
I
n
o
r
d
er
to
ch
o
o
s
e
th
e
b
est
tr
ain
i
n
g
a
lg
o
r
it
h
m
,
th
e
b
es
t
n
et
w
o
r
k
ar
c
h
itect
u
r
e
is
d
e
f
in
ed
f
ir
s
t.
Fo
r
R
o
u
te
1
s
h
o
w
n
in
T
ab
le
2
,
th
er
e
w
er
e
8
iter
atio
n
s
in
f
i
n
d
in
g
t
h
e
b
est
n
e
t
w
o
r
k
ar
c
h
itect
u
r
e.
Ho
w
e
v
er
,
6
n
et
w
o
r
k
ar
ch
itec
tu
r
e
s
w
h
ich
i
s
in
r
ed
co
lo
r
ed
h
as
b
ee
n
r
e
m
o
v
ed
i
n
o
r
d
e
r
to
av
o
id
o
v
er
f
it
p
r
o
b
lem
w
h
en
t
h
e
n
u
m
b
er
o
f
h
id
d
en
n
o
d
e
s
is
g
r
ea
ter
th
a
n
th
e
n
u
m
b
er
o
f
i
n
p
u
t
n
o
d
es
.
Fro
m
t
h
e
r
es
u
lts
,
it
w
as
f
o
u
n
d
th
at
t
h
e
b
est
ar
ch
itectu
r
e
i
s
[
5
-
5
-
1
]
m
o
d
el
s
in
ce
it
g
i
v
es
t
h
e
lar
g
es
t
f
it
n
es
s
v
al
u
e,
lo
w
est
tes
t e
r
r
o
r
an
d
lo
w
e
s
t
A
I
C
.
T
ab
le
3
also
s
h
o
w
s
t
h
at
t
h
e
b
est ar
ch
itec
tu
r
e
f
o
r
R
o
u
te
2
i
s
also
[
5
-
5
-
1
]
m
o
d
el.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
4
8
0
–
487
484
T
ab
le
2
.
A
r
ch
itect
u
r
e
n
et
w
o
r
k
o
f
A
N
N
at
r
o
u
te
1
ID
A
r
c
h
i
t
e
c
t
u
r
e
F
i
t
n
e
ss
T
e
st
Er
r
o
r
A
I
C
C
o
r
r
e
l
a
t
i
o
n
R
-
S
q
u
a
r
e
d
1
[5
-
1
-
1]
0
.
0
8
5
9
1
1
.
6
4
3
9
-
2
8
2
5
.
6
0
9
8
0
.
9
9
4
1
0
.
9
8
8
2
2
[5
-
13
-
1]
0
.
2
3
7
1
4
.
2
1
7
3
-
3
9
1
8
.
6
4
5
7
0
.
9
9
9
7
0
.
9
9
9
4
3
[5
-
8
-
1]
0
.
2
0
5
7
4
.
8
6
2
6
-
3
6
3
5
.
6
8
2
1
0
.
9
9
9
4
0
.
9
9
8
8
4
[5
-
5
-
1]
0
.
2
0
4
3
4
.
8
9
4
9
-
3
5
7
1
.
4
9
6
8
0
.
9
9
8
7
0
.
9
9
7
4
5
[5
-
11
-
1]
0
.
1
7
5
1
5
.
7
0
9
7
-
3
5
5
1
.
4
6
4
2
0
.
9
9
9
2
0
.
9
9
8
4
6
[5
-
9
-
1]
0
.
1
5
3
8
6
.
5
0
3
6
-
3
5
1
5
.
9
0
8
4
0
.
9
9
9
2
0
.
9
9
8
4
7
[5
-
6
-
1]
0
.
1
7
1
6
5
.
8
2
7
6
-
3
6
3
3
.
2
9
1
2
0
.
9
9
9
2
0
.
9
9
8
4
8
[5
-
7
-
1]
0
.
1
8
2
8
5
.
4
696
-
3
6
5
5
.
1
7
6
3
0
.
9
9
9
2
0
.
9
9
8
4
T
ab
le
3
.
A
r
ch
itect
u
r
e
n
et
w
o
r
k
o
f
A
N
N
at
r
o
u
te
2
ID
A
r
c
h
i
t
e
c
t
u
r
e
F
i
t
n
e
ss
T
e
st
Er
r
o
r
A
I
C
C
o
r
r
e
l
a
t
i
o
n
R
-
S
q
u
a
r
e
d
1
[5
-
11
-
1]
0
.
6
8
6
5
1
.
4
5
6
6
-
3
8
9
6
.
9
1
7
8
0
.
9
9
9
8
0
.
9
9
9
6
2
[5
-
9
-
1]
0
.
6
2
6
2
1
.
5
9
6
9
-
3
8
6
6
.
8
7
9
3
0
.
9
9
9
7
0
.
9
9
9
4
3
[5
-
12
-
1]
0.
6
0
8
5
1
.
6
4
3
5
-
3
7
9
4
.
2
9
0
6
0
.
9
9
9
7
0
.
9
9
9
4
4
[5
-
10
-
1]
0
.
5
9
1
4
1
.
6
9
0
8
-
3
8
5
5
.
7
8
0
6
0
.
9
9
9
7
0
.
9
9
9
4
5
[5
-
13
-
1]
0
.
5
2
8
8
1
.
8
9
0
9
-
3
7
3
9
.
4
9
9
2
0
.
9
9
9
7
0
.
9
9
9
4
6
[5
-
8
-
1]
0
.
4
0
0
2
2
.
4
9
8
9
-
3
6
4
5
.
9
1
4
9
0
.
9
9
9
4
0
.
9
9
8
8
7
[5
-
5
-
1]
0
.
1
2
1
7
8
.
2
1
6
4
-
2
8
8
6
.
5
1
6
2
0
.
9
9
7
1
0
.
9
9
4
2
8
[5
-
1
-
1]
0
.
1
1
1
7
8
.
9
5
4
0
-
2
8
9
7
.
0
0
7
5
0
.
9
9
5
5
0
.
9
9
1
0
T
h
e
f
itn
e
s
s
o
f
tr
ai
n
i
n
g
alg
o
r
ith
m
s
is
a
ls
o
d
o
n
e
in
w
h
i
ch
T
ab
le
4
s
h
o
w
s
t
h
at
L
M
p
r
o
d
u
ce
s
th
e
s
m
al
lest
v
al
u
e
o
f
A
b
s
o
l
u
te
an
d
Net
w
o
r
k
E
r
r
o
r
f
o
r
b
o
th
R
o
u
te
1
a
n
d
R
o
u
te
2
.
T
ab
le
4
also
s
h
o
w
s
th
at
L
M
p
r
o
d
u
ce
s
th
e
s
m
a
lle
s
t
er
r
o
r
v
alu
e
(
R
MSE
a
n
d
M
A
P
E
)
f
o
r
b
o
th
tr
ain
i
n
g
a
n
d
v
alid
atio
n
p
ar
t
s
f
o
r
R
o
u
te
1
an
d
R
o
u
te
2
.
T
ab
le
4
.
C
o
m
p
ar
is
o
n
of
er
r
o
r
m
ea
s
u
r
es
f
o
r
tr
ain
in
g
al
g
o
r
ith
m
s
R
o
u
t
e
1
T
r
a
i
n
i
n
g
A
l
g
o
r
i
t
h
m
T
r
a
i
n
i
n
g
V
a
l
i
d
a
t
i
o
n
R
M
S
E
M
A
P
E
A
b
so
l
u
t
e
Er
r
o
r
N
e
t
w
o
r
k
Er
r
o
r
R
M
S
E
M
A
P
E
Ab
so
l
u
t
e
Er
r
o
r
N
e
t
w
o
r
k
Er
r
o
r
C
G
D
1
9
.
8
0
5
3
8
2
.
5
0
6
0
.
5
2
8
0
.
0
0
1
2
2
.
8
8
4
5
2
3
.
6
6
4
0
.
3
8
9
0
LM
4
.
3
2
3
1
8
.
2
2
2
0
.
3
4
1
0
5
.
3
8
3
2
8
.
9
8
0
.
3
2
0
QP
1
9
.
5
6
4
3
7
3
.
2
8
7
4
.
5
3
9
0
2
2
.
7
0
6
5
1
5
.
5
4
1
3
.
9
1
7
0
R
o
u
t
e
2
T
r
a
i
n
i
n
g
A
l
g
o
r
i
t
h
m
T
r
a
i
n
i
n
g
V
a
l
i
d
a
t
i
o
n
R
M
S
E
M
A
P
E
A
b
so
l
u
t
e
Er
r
o
r
N
e
t
w
o
r
k
Er
r
o
r
R
M
S
E
M
A
P
E
A
b
so
l
u
t
e
Er
r
o
r
N
e
t
w
o
r
k
Er
r
o
r
C
G
D
5
.
5
3
2
2
9
.
8
4
3
4
.
5
0
5
7
.
2
9
4
3
2
8
2
.
5
5
2
.
5
9
6
0
LM
4
.
8
0
7
2
2
.
5
3
3
0
.
6
6
9
0
2
4
.
4
3
5
9
6
.
8
3
5
0
.
5
0
5
0
QP
5
.
5
8
7
3
0
.
4
4
5
2
.
6
6
0
2
4
.
6
4
2
6
0
7
.
2
2
6
4
.
8
0
3
.
2
.
F
o
re
ca
s
t
ing
by
us
ing
t
he
bes
t
t
ra
ini
ng
a
lg
o
rit
h
m
As
p
r
ev
io
u
s
l
y
d
is
cu
s
s
ed
,
th
e
b
est
tr
ain
in
g
alg
o
r
it
h
m
w
i
ll
b
e
u
s
ed
f
o
r
p
r
ed
ictio
n
o
f
ca
r
r
ied
w
e
ig
h
t.
Hen
ce
,
t
h
e
ANN
m
o
d
el
w
it
h
L
M
a
s
t
h
e
tr
ain
i
n
g
alg
o
r
ith
m
is
u
s
ed
to
p
r
ed
ict
in
b
o
th
r
o
u
tes
,
R
o
u
te
1
an
d
R
o
u
te
2
.
T
h
e
am
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
f
o
r
ec
ast
ed
f
o
r
y
ea
r
2
0
1
9
at
R
o
u
te
1
is
ill
u
s
tr
ated
in
Fig
u
r
e
2
.
T
h
e
g
r
e
y
lin
e
r
ep
r
esen
t
s
t
h
e
f
o
r
ec
ast
v
al
u
e
a
n
d
t
h
e
d
o
tted
o
r
an
g
e
li
n
e
r
ep
r
esen
ts
t
h
e
tr
e
n
d
li
n
e
o
f
t
h
e
n
e
w
f
o
r
ec
asted
v
al
u
es
w
h
ich
n
e
g
at
iv
e
s
lo
p
e
in
d
icate
s
t
h
at
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
f
o
r
R
o
u
te
1
s
lig
h
t
l
y
d
ec
r
ea
s
e
an
d
g
o
i
n
g
to
d
ec
li
n
e
o
v
er
ti
m
e.
T
h
e
tr
en
d
li
n
e
o
f
t
h
e
n
e
w
f
o
r
ec
asted
ca
r
r
ied
w
ei
g
h
t
v
a
lu
e
s
is
co
n
s
tr
u
cted
a
n
d
it
ca
n
b
e
co
n
cl
u
d
ed
th
at
th
er
e
is
a
d
ec
r
ea
s
e
i
n
a
m
o
u
n
t
o
f
to
tal
to
n
n
ag
e
ca
r
r
ied
ea
ch
d
a
y
b
y
0
.
0
4
8
9
th
is
d
u
e
t
o
n
eg
at
iv
e
r
elatio
n
s
h
ip
.
T
h
e
eq
u
ati
o
n
is
y
=
-
0
.
0
4
8
9
x
+
3
2
4
9
.
8
.
T
h
e
f
o
r
ec
ast
v
al
u
e
s
h
o
w
s
t
h
at
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
f
lu
c
tu
ate
s
o
v
er
ti
m
e
a
n
d
d
ec
r
ea
s
e
b
y
4
8
.
9
k
g
p
er
d
ay
.
T
h
en
,
th
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
f
o
r
ec
ast
f
o
r
y
ea
r
2
0
1
9
at
R
o
u
te
2
w
a
s
ill
u
s
tr
ated
i
n
Fig
u
r
e
3
.
T
h
e
g
r
e
y
lin
e
r
ep
r
esen
t
s
t
h
e
f
o
r
ec
ast
v
al
u
e
a
n
d
t
h
e
d
o
tted
o
r
an
g
e
li
n
e
r
ep
r
esen
ts
t
h
e
tr
e
n
d
li
n
e
o
f
t
h
e
n
e
w
f
o
r
ec
asted
v
al
u
e
.
T
h
e
f
o
r
ec
as
ted
lin
e
f
o
r
R
o
u
te
2
al
s
o
is
h
av
in
g
a
n
e
g
ati
v
e
s
lo
p
e
in
d
icat
in
g
th
a
t
a
m
o
u
n
t
o
f
ca
r
r
ied
w
e
ig
h
t
f
o
r
R
o
u
te
2
s
l
ig
h
tl
y
d
ec
r
ea
s
e
s
o
v
er
ti
m
e.
T
h
e
tr
en
d
eq
u
atio
n
is
ca
lc
u
late
d
,
y
=
-
0
.
1
1
8
6
x
+
6
0
7
9
.
3
w
h
ich
s
h
o
w
s
a
d
ec
r
ea
s
e
b
y
1
1
8
.
6
k
g
p
er
d
a
y
.
C
o
m
p
ar
in
g
th
e
tr
en
d
lin
e
o
f
th
e
n
e
w
f
o
r
ec
asted
ca
r
r
ied
w
ei
g
h
t
w
i
th
ea
r
lier
tr
e
n
d
in
Sectio
n
4
.
5
.
2
,
it
ca
n
b
e
s
ee
n
th
at
t
h
e
d
ec
r
ea
s
e
o
f
a
v
er
ag
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
f
o
r
R
o
u
te
1
s
lig
h
tl
y
c
h
an
g
es
f
r
o
m
6
9
.
1
k
g
p
er
d
a
y
to
o
n
l
y
4
8
.
9
k
g
p
er
d
a
y
.
T
h
er
ef
o
r
e,
it
ca
n
b
e
co
n
clu
d
ed
t
h
at
t
h
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
o
f
ca
r
g
o
i
s
i
n
cr
ea
s
i
n
g
an
d
ca
r
g
o
b
u
s
i
n
es
s
i
s
i
m
p
r
o
v
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
p
p
lica
ti
o
n
o
f a
r
tifi
cia
l n
eu
r
a
l n
etw
o
r
k
to
p
r
ed
ict
a
mo
u
n
t
o
f…
(
S
iti N
a
s
u
h
a
Zu
b
ir
)
485
On
t
h
e
o
t
h
er
h
an
d
,
f
o
r
R
o
u
te
2
,
th
e
d
ec
r
ea
s
e
o
f
a
v
er
ag
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
ei
g
h
t
c
h
an
g
e
s
f
r
o
m
1
1
7
.
8
k
g
p
er
d
ay
to
1
1
8
.
6
k
g
p
er
d
ay
.
Ho
wev
er
,
th
e
ch
a
n
g
e
is
o
n
l
y
0
.
6
8
% a
n
d
v
er
y
m
i
n
i
m
al.
Fig
u
r
e
2
.
Am
o
u
n
t o
f
ca
r
r
ied
w
eig
h
t f
o
r
ec
ast
f
o
r
r
o
u
te
1
Fig
u
r
e
3
.
Am
o
u
n
t o
f
ca
r
r
ied
w
eig
h
t f
o
r
ec
ast
f
o
r
r
o
u
te
2
4.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
p
r
esen
ts
th
e
ap
p
licatio
n
o
f
A
r
ti
f
icia
l
Neu
r
al
Net
w
o
r
k
(
ANN)
to
p
r
ed
ict
th
e
a
m
o
u
n
t
o
f
ca
r
r
ied
o
f
ca
r
g
o
tr
ain
,
u
s
i
n
g
t
h
r
ee
tr
ai
n
in
g
a
lg
o
r
it
h
m
s
:
L
e
v
e
n
b
er
g
-
Ma
r
q
u
ar
d
t
alg
o
r
ith
m
(
L
M)
as
a
w
ell
p
er
f
o
r
m
ed
al
g
o
r
ith
m
to
p
r
ed
ict
d
if
f
er
en
t
s
et
o
f
ca
r
r
ied
w
ei
g
h
t
d
ata,
C
o
n
j
u
g
at
e
Gr
ad
ien
t
Descen
t
(
GC
D)
as
a
w
ell
p
er
f
o
r
m
a
lg
o
r
ith
m
f
o
r
p
r
ed
ictio
n
o
f
o
th
er
s
et
s
o
f
d
ata
a
n
d
Q
u
ic
k
P
r
o
p
ag
atio
n
(
QP)
as
a
n
e
w
al
g
o
r
ith
m
u
s
ed
to
p
r
ed
ict
ca
r
r
ied
w
eig
h
t.
T
h
e
a
ch
iev
ed
r
e
s
u
l
ts
s
h
o
w
t
h
e
ap
p
r
o
p
r
iaten
ess
o
f
t
h
e
A
r
ti
f
icial
Ne
u
r
al
Net
w
o
r
k
in
p
r
ed
ictin
g
t
h
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
e
ig
h
t
b
ased
o
n
th
e
co
r
r
elatio
n
,
f
it
n
es
s
an
d
test
er
r
o
r
v
alu
es.
B
ased
o
n
t
h
e
R
MSE
an
d
M
A
P
E
,
L
M
s
h
o
w
s
th
e
s
m
alle
s
t
v
alu
e
s
f
o
r
R
o
u
te
1
an
d
R
o
u
te
2
t
h
at
ca
r
r
y
ce
m
en
t c
ar
g
o
s
f
o
r
KT
M
B
cu
s
to
m
er
s
.
Fu
r
t
h
er
m
o
r
e,
t
h
e
A
NN
m
o
d
el
b
ased
o
n
t
h
e
b
est
tr
ain
in
g
al
g
o
r
ith
m
f
o
u
n
d
i
n
t
h
e
f
ir
s
t
p
h
ase
o
f
th
e
s
tu
d
y
is
u
s
ed
to
f
o
r
ec
ast
v
al
u
e
o
f
ca
r
r
ied
w
e
ig
h
t
o
f
ca
r
g
o
t
r
ain
f
o
r
b
o
th
r
o
u
tes
i
n
r
ail
tr
an
s
p
o
r
tatio
n
s
y
s
te
m
.
R
es
u
lts
s
h
o
w
t
h
at
th
e
v
alu
e
s
o
f
ca
r
r
ied
w
eig
h
t
f
l
u
ctu
a
te
an
d
d
ec
lin
e
o
v
er
ti
m
e
f
o
r
y
ea
r
2
0
1
9
(
3
6
5
d
ay
s
ah
ea
d
)
.
I
t
is
h
o
p
e
th
at
th
e
r
es
u
lt
s
ca
n
h
elp
KT
MB
to
p
lan
th
e
r
ig
h
t
a
m
o
u
n
t
to
b
e
ca
r
r
ied
b
y
it
s
c
ar
g
o
p
er
tr
ip
in
its
ef
f
o
r
t
to
p
r
ev
en
t
f
o
r
m
m
o
r
e
d
er
ail
m
en
t
s
o
cc
u
r
r
en
ce
.
A
t
t
h
e
s
a
m
e
ti
m
e,
a
s
t
h
e
a
m
o
u
n
t
o
f
ca
r
r
ied
w
eig
h
t
is
p
r
ed
icted
t
o
d
ec
lin
e
o
v
er
ti
m
e,
KT
M
B
ca
n
p
lan
a
s
tr
ateg
ic
in
it
iati
v
e
in
g
et
tin
g
m
o
r
e
cu
s
to
m
er
s
w
h
ile
m
o
n
ito
r
i
n
g
t
h
e
r
ig
h
t a
m
o
u
n
t t
o
ca
r
r
y
ea
ch
tr
ip
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8938
I
n
t J
A
r
ti
f
I
n
tell
,
Vo
l.
9
,
No
.
3
,
Sep
te
m
b
er
20
20
:
4
8
0
–
487
486
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
e
au
th
o
r
s
w
o
u
ld
li
k
e
to
th
a
n
k
t
h
e
Mi
n
is
tr
y
o
f
E
d
u
ca
tio
n
Ma
la
y
s
ia
(
MO
E
)
a
n
d
Ma
la
y
s
i
a
I
n
s
tit
u
te
o
f
T
r
an
s
p
o
r
t
(
MI
T
R
A
NS)
,
U
n
iv
er
s
itiT
ek
n
o
lo
g
i
M
AR
A
,
Ma
la
y
s
ia
f
o
r
s
u
p
p
o
r
tin
g
t
h
i
s
p
r
o
ject.
Sp
ec
ial
th
a
n
k
s
to
KT
M
B
C
ar
g
o
tea
m
led
b
y
E
n
Mo
h
d
Din
f
o
r
p
r
o
v
id
in
g
i
n
f
o
r
m
atio
n
an
d
s
u
p
p
o
r
t th
r
o
u
g
h
o
u
t th
i
s
s
t
u
d
y
.
RE
F
E
R
E
NC
E
S
[1
]
G
o
o
d
a
ll
,
C.
Ho
w
to
re
d
u
c
e
y
o
u
r
c
a
rb
o
n
f
o
o
tp
rin
t
.
T
h
e
G
u
a
rd
ian
,
Re
tri
e
v
e
d
h
tt
p
s:/
/www
.
th
e
g
u
a
rd
ian
.
c
o
m
/en
v
iro
n
m
e
n
t/
2
0
1
7
/j
a
n
/
1
9
/
h
o
w
to
re
d
u
c
e
-
c
a
rb
o
n
-
f
o
o
tp
r
in
t.
[2
]
No
o
r,
H.
M
.
S
e
m
u
a
‘terg
e
li
n
c
ir’
a
p
a
b
il
a
k
e
re
ta
a
p
i
k
e
lu
a
r
lan
d
a
sa
n
.
Utu
sa
n
On
li
n
e
,
Re
tri
e
v
e
d
h
tt
p
s:/
/www
.
u
tu
sa
n
.
c
o
m
.
m
y
/ren
c
a
n
a
/u
tam
a
/s
e
m
u
a
-
terg
e
li
n
c
ir
-
a
p
a
b
il
a
-
k
e
re
ta
-
a
p
i
-
k
e
lu
a
r
-
lan
d
a
sa
n
-
1
.
5
1
9
8
7
2
,
2
0
1
7
.
[3
]
Ku
m
a
r
a
p
p
a
,
K.
M
.
Ke
re
ta
a
p
i
Ku
a
n
g
-
Ra
wa
n
g
terje
ja
s.
M
y
m
e
tro
,
Re
tri
e
v
e
d
h
tt
p
s:/
/www
.
h
m
e
tro
.
c
o
m
.
m
y
/
m
u
t
a
k
h
ir/
2
0
1
7
/
0
9
/
2
6
6
2
8
2
/k
e
re
tap
ik
u
a
n
g
-
ra
w
a
n
g
-
terje
ja
s.
[4
]
Be
rn
a
m
a
.
(2
0
1
7
,
No
v
e
m
b
e
r
2
3
).
T
h
e
d
e
ra
il
m
e
n
t
d
isru
p
ts
KT
M
ra
il
se
rv
ic
e
.
M
a
la
y
m
a
il
,
Re
tri
e
v
e
d
h
tt
p
s:/
/www
.
m
a
la
y
m
a
il
.
c
o
m
/n
e
ws
/ma
la
y
si
a
/2
0
1
7
/1
1
/2
3
/t
ra
i
n
d
e
ra
il
m
e
n
-
d
isru
p
ts
-
k
tm
-
ra
il
-
se
r
v
ice
s/1
5
1
6
9
1
9
.
[5
]
Hu
ss
in
,
M
.
H.
(2
0
1
9
,
Ju
l
y
1
8
).
T
re
n
d
ij
a
n
g
k
a
b
e
ro
p
e
ra
si
se
m
u
la
Isn
in
.
M
y
m
e
t
ro
,
Re
tri
e
v
e
d
h
tt
p
s:/
/www
.
h
m
e
tro
.
c
o
m
.
m
y
/
m
u
t
a
k
h
ir/
2
0
1
9
/
0
7
/
4
7
6
8
3
5
/
tren
-
d
ij
a
n
g
k
a
-
b
e
ro
p
e
ra
si
-
se
m
u
la
-
isn
in
.
[6
]
S
.
V
e
rm
a
e
t
a
l.
“
AN
N
b
a
se
d
m
e
th
o
d
f
o
r
im
p
ro
v
in
g
g
o
ld
p
rice
f
o
re
c
a
stin
g
a
c
c
u
ra
c
y
th
ro
u
g
h
m
o
d
if
ied
G
r
a
d
ien
t
De
sc
e
n
t
M
e
th
o
d
s”
,
I
AE
S
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
Arti
f
icia
l
I
n
telli
g
e
n
c
e
,
v
o
l.
9
(1
),
2
0
2
0
.
[7
]
He
sa
m
K
a
ri
m
e
t
a
l.
,
Co
m
p
a
riso
n
o
f
Ne
u
ra
l
Ne
tw
o
rk
T
ra
in
in
g
A
l
g
o
rit
h
m
s
f
o
r
Clas
sif
ic
a
ti
o
n
o
f
He
a
rt
Dise
a
s
e
s,
IAE
S
I
n
ter
n
a
t
io
n
a
l
J
o
u
r
n
a
l
o
f
Art
if
icia
l
I
n
telli
g
e
n
c
e
,
v
o
l.
/i
ss
u
e
7
(
4
),
p
p
.
1
8
5
-
1
8
9
,
2
0
1
8
.
[8
]
L
in
a
w
a
ti
,
L
.
,
S
u
d
a
rm
a
,
M
&
W
is
n
a
w
a
,
I.
P
.
O.
F
o
re
c
a
stin
g
r
u
p
iah
e
x
c
h
a
n
g
e
ra
te
w
it
h
lea
rn
in
g
v
e
c
t
o
r
q
u
a
n
ti
z
a
ti
o
n
n
e
u
ra
l
n
e
tw
o
rk
,
In
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
,
V
o
l
1
8
(1
),
2
4
-
31
,
2
0
2
0
.
[9
]
Be
rh
il
,
S
.
,
Be
n
lah
m
a
r,
H.
&
L
a
b
a
n
i,
N.
A
re
v
ie
w
p
a
p
e
r
o
n
a
rti
f
icia
l
in
telli
g
e
n
c
e
a
t
th
e
se
rv
ice
o
f
h
u
m
a
n
re
so
u
rc
e
s
m
a
n
a
g
e
m
e
n
t,
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
i
n
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
,
Vo
l
1
8
(
1
),
p
p
3
2
-
40
,
2
0
2
0
.
[1
0
]
A
b
.
Wah
a
b
,
N.,
M
a
t
Ya
sin
,
Z.
,
S
a
li
m
,
N.
A
.
&
A
z
iz,
N.
F
.
A
.
A
rti
f
icia
l
n
e
u
ra
l
n
e
t
w
o
rk
-
b
a
se
d
tec
h
n
iq
u
e
f
o
r
e
n
e
rg
y
m
a
n
a
g
e
m
e
n
t
p
re
d
icti
o
n
,
In
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
Vo
l
1
7
(1
)
,
p
p
9
4
-
101
,
2
0
1
9
.
[1
1
]
Am
a
n
i,
J.,
&
M
o
e
in
i,
R.
P
re
d
ict
io
n
o
f
sh
e
a
r
stre
n
g
th
o
f
re
in
f
o
rc
e
d
c
o
n
c
re
te
b
e
a
m
s
u
sin
g
a
d
a
p
ti
v
e
n
e
u
ro
-
f
u
z
z
y
in
f
e
re
n
c
e
s
y
ste
m
a
n
d
a
rti
f
icia
l
n
e
u
ra
l
n
e
tw
o
rk
.
S
c
ien
ti
a
Ira
n
ica
,
1
9
(2
),
2
4
2
-
2
4
8
,
2
0
1
2
.
[1
2
]
Ba
rh
m
i,
S
.
&
El
F
a
tn
i
,
O.
Ho
u
rl
y
w
in
d
sp
e
e
d
f
o
re
c
a
stin
g
b
a
se
d
o
n
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
in
e
a
n
d
A
rti
f
icia
l
Ne
u
ra
l
Ne
tw
o
rk
s
,
IAE
S
I
n
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
Arti
fi
c
i
a
l
I
n
te
ll
ig
e
n
c
e
,
Vo
l
8
,
No
3
,
p
p
2
8
6
-
2
9
1
,
2
0
1
9
.
[1
3
]
Ef
e
n
d
ig
il
,
T
.
,
&
On
u
t,
S
.
A
n
in
teg
ra
ti
o
n
m
e
th
o
d
o
l
o
g
y
b
a
se
d
o
n
f
u
z
z
y
in
f
e
re
n
c
e
s
y
ste
m
s
a
n
d
n
e
u
ra
l
a
p
p
ro
a
c
h
e
s
f
o
r
m
u
lt
i
-
sta
g
e
su
p
p
ly
-
c
h
a
in
s.
Co
mp
u
ter
s
&
In
d
u
stria
l
En
g
in
e
e
rin
g
,
6
2
(2
),
5
5
4
-
5
6
9
,
2
0
1
2
.
[1
4
]
Ra
h
m
a
n
,
M
.
N.
A
.
,
Ja
fa
rz
a
d
e
h
-
Gh
o
u
s
h
c
h
i,
S
.
,
W
a
h
a
b
,
D.
A
.
,
&
Ja
f
a
r
z
a
d
e
h
-
G
h
o
u
sh
ji
,
M
.
A
rti
f
icia
l
Ne
u
ra
l
Ne
tw
o
rk
M
o
d
e
li
n
g
S
tu
d
ies
to
P
re
d
ict
th
e
Am
o
u
n
t
o
f
Ca
rried
Weig
h
t
B
y
Ir
a
n
Kh
o
d
r
o
T
ra
n
sp
o
rtati
o
n
S
y
ste
m
.
L
if
e
S
c
ien
c
e
J
o
u
rn
a
l
,
1
1
(S
P
EC.
I
S
S
.
2
),
1
4
6
-
1
5
4
,
2
0
1
4
.
[1
5
]
G
h
o
u
sh
c
h
i,
S
.
J.,
&
Ra
h
m
a
n
,
M
.
N.
A
.
P
e
rf
o
rm
a
n
c
e
stu
d
y
o
f
a
rti
f
i
c
ial
n
e
u
ra
l
n
e
tw
o
rk
m
o
d
e
ll
in
g
to
p
re
d
ict
c
a
rried
w
e
i
g
h
t
in
t
h
e
tran
sp
o
rtati
o
n
sy
ste
m
.
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
L
o
g
is
ti
c
s S
y
ste
ms
a
n
d
M
a
n
a
g
e
me
n
t
,
2
4
(2
),
2
0
0
,
2
0
1
6
.
[1
6
]
P
o
k
ra
jac
,
D.,
&
L
a
z
a
re
v
ic,
A
.
Ap
p
li
c
a
ti
o
n
s
o
f
u
n
s
u
p
e
rv
ise
d
n
e
u
r
a
l
n
e
tw
o
rk
s
in
d
a
ta
m
in
in
g
.
In
7
th
S
e
min
a
r
o
n
Ne
u
ra
l
Ne
two
rk
A
p
p
li
c
a
ti
o
n
a
n
d
El
e
c
c
trica
l
En
g
in
e
e
rin
g
NEURE
L
,
1
7
-
2
0
,
2
0
0
4
.
[1
7
]
S
a
y
e
d
A
,
S
a
rd
e
sh
m
u
k
h
M
,
&
L
i
m
k
a
r,
S
.
Op
ti
m
isa
ti
o
n
Us
in
g
L
e
v
e
n
b
e
rg
-
M
a
rq
u
a
rd
t
A
lg
o
rit
h
m
o
f
Ne
u
ra
l
Ne
tw
o
rk
s
f
o
r
Iris
.
In
:
S
a
tap
a
th
y
S
,
Ud
g
a
ta
S
,
Bisw
a
l
B
(
e
d
s)
P
ro
c
e
e
d
i
n
g
s
o
f
th
e
In
tern
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
F
r
o
n
ti
e
rs
o
f
In
telli
g
e
n
t
C
o
m
p
u
ti
n
g
:
T
h
e
o
ry
a
n
d
A
p
p
l
ica
ti
o
n
s
(F
ICT
A
)
2
0
1
3
.
Ad
v
a
n
c
e
s
i
n
I
n
telli
g
e
n
t
S
y
ste
ms
a
n
d
Co
m
p
u
t
in
g
,
v
o
l.
2
4
7
.
S
p
r
in
g
e
r,
Ch
a
m
.
[1
8
]
X
iao
,
Y.,
&
Z
h
u
,
H
.
A
c
o
n
j
u
g
a
te
g
ra
d
ien
t
m
e
th
o
d
t
o
so
lv
e
c
o
n
v
e
x
c
o
n
stra
in
e
d
m
o
n
o
to
n
e
e
q
u
a
ti
o
n
s
w
it
h
a
p
p
li
c
a
ti
o
n
s in
c
o
m
p
re
ss
iv
e
se
n
si
n
g
.
J
o
u
r
n
a
l
o
f
M
a
t
h
e
ma
ti
c
a
l
A
n
a
l
y
sis a
n
d
Ap
p
li
c
a
t
io
n
s
,
4
0
5
(
1
),
3
1
0
-
3
1
9
,
2
0
1
3
.
[1
9
]
Ch
o
u
h
a
n
,
S
.
S
.
,
Ka
u
l,
A
.
&
S
in
g
h
,
U.P
.
Im
a
g
e
se
g
m
e
n
tatio
n
u
sin
g
f
u
z
z
y
c
o
m
p
e
ti
ti
v
e
le
a
rn
in
g
b
a
se
d
c
o
u
n
te
r
p
ro
p
a
g
a
ti
o
n
n
e
tw
o
rk
.
M
u
lt
ime
d
T
o
o
ls A
p
p
l
7
8
,
3
5
2
6
3
-
3
5
2
8
7
,
2
0
1
9
.
[2
0
]
F
il
d
e
s,
R.
,
Hi
b
o
n
,
M
.
,
M
a
k
rid
a
k
is,
S
.
,
&
M
e
a
d
e
,
N.
G
e
n
e
ra
li
sin
g
a
b
o
u
t
u
n
iv
a
riate
f
o
re
c
a
stin
g
m
e
th
o
d
s:
f
u
rth
e
r
e
m
p
iri
c
a
l
e
v
id
e
n
c
e
.
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
F
o
re
c
a
stin
g
,
1
4
(3
)
,
3
3
9
-
3
5
8
,
1
9
9
8
.
[2
1
]
Ya
p
,
S
ZZ
.
,
Zah
a
ri
,
SM
.
,
De
ra
sit
,
Z
.
&
S
h
a
rif
f
,
S
S
R.
A
n
it
e
ra
ti
v
e
Ne
w
to
n
-
Ra
p
h
s
o
n
(NR)
m
e
th
o
d
o
n
L
e
e
-
Ca
rter
p
a
ra
m
e
ter’s
e
sti
m
a
ti
o
n
f
o
r
p
re
d
ictin
g
h
o
sp
it
a
l
a
d
m
issio
n
ra
tes
,
AIP
Co
n
fer
e
n
c
e
Pro
c
e
e
d
in
g
s
,
1
9
7
4
(
1
),
P
a
g
e
s
0
2
0
0
4
9
,
2
0
1
8
.
[2
2
]
Ya
p
,
S
ZZ
.
,
Zah
a
ri
,
SM
.
,
De
ra
sit
,
Z
.
&
S
h
a
rif
f
,
SSR
.
Co
m
p
a
rin
g
M
e
th
o
d
s
f
o
r
L
e
e
–
Ca
rter
P
a
ra
m
e
t
e
r’s
Esti
m
a
ti
o
n
f
o
r
P
re
d
ictin
g
Ho
sp
it
a
l
A
d
m
issi
o
n
Ra
tes
,
Pro
c
e
e
d
in
g
s
o
f
th
e
S
e
c
o
n
d
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
th
e
Fu
tu
re
o
f
AS
EA
N
(
ICo
FA
)
,
V
o
l
u
m
e
2
,
3
6
1
-
3
7
2
,
2
0
1
8
.
[2
3
]
S
h
a
rif
f
,
SSR
.
,
M
a
a
d
,
HA
.
,
H
a
li
m
,
N
N
A
.
&
D
e
ra
sit
,
Z
.
De
ter
m
in
in
g
h
o
tsp
o
ts
o
f
ro
a
d
a
c
c
id
e
n
ts
u
sin
g
sp
a
ti
a
l
a
n
a
ly
sis
,
In
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
E
lec
trica
l
En
g
in
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
,
Vo
lu
m
e
9
(
1
)
,
P
a
g
e
s
146
-
1
5
1
,
2
0
1
8
.
[2
4
]
Is
m
a
e
e
l,
S
.
,
A
l
-
k
h
a
z
ra
ji
,
A
.
,
&
A
l
-
d
e
li
m
i,
K.
F
u
z
z
y
In
f
o
rm
a
ti
o
n
M
o
d
e
li
n
g
in
a
Da
tab
a
se
S
y
st
e
m
.
IAE
S
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
Art
if
icia
l
In
telli
g
e
n
c
e
(
IJ
-
AI)
,
V
o
l
.
6
,
No
.
1
,
p
p
.
1
-
7
,
2
0
1
7
.
[2
5
]
Be
n
n
e
tt
,
D.
A
.
Ho
w
c
a
n
I
d
e
a
l
w
it
h
m
issin
g
d
a
ta
in
m
y
stu
d
y
?
Au
stra
li
a
n
a
n
d
Ne
w
Z
e
a
l
a
n
d
J
o
u
r
n
a
l
o
f
Pu
b
li
c
He
a
lt
h
,
2
5
(
5
),
4
6
4
-
4
6
9
,
2
0
0
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
A
r
ti
f
I
n
tell
I
SS
N:
2252
-
8938
A
p
p
lica
ti
o
n
o
f a
r
tifi
cia
l n
eu
r
a
l n
etw
o
r
k
to
p
r
ed
ict
a
mo
u
n
t
o
f…
(
S
iti N
a
s
u
h
a
Zu
b
ir
)
487
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
S
it
i
Na
su
h
a
Zu
b
ir
,
s
h
e
is
a
p
o
st
g
r
a
d
u
a
te
stu
d
e
n
t
i
n
M
a
ste
r
o
f
Qu
a
n
t
it
a
ti
v
e
S
c
ien
c
e
s
in
F
a
c
u
lt
y
o
f
Co
m
p
u
ter an
d
M
a
th
e
m
a
ti
c
a
l
S
c
ie
n
c
e
s,
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
A
RA
,
S
h
a
h
A
la
m
.
T
s
Dr
S
.
S
a
ri
f
a
h
Ra
d
iah
S
h
a
rif
f
,
s
h
e
is
a
S
e
n
io
r
L
e
c
tu
re
r
in
Ce
n
t
re
o
f
S
tatisti
c
s
a
n
d
De
c
isio
n
S
c
ien
c
e
S
tu
d
ies
in
F
a
c
u
l
ty
o
f
Co
m
p
u
ter
a
n
d
M
a
t
h
e
m
a
ti
c
a
l
S
c
ien
c
e
s,
Un
iv
e
rsiti
T
e
k
n
lo
g
i
M
A
R
A
,
S
h
a
h
A
la
m
.
He
r
e
x
p
e
rti
s
e
is
in
S
u
p
p
ly
Ch
a
in
a
n
d
L
o
g
isti
c
s
M
o
d
e
ll
i
n
g
.
Cu
rre
n
tl
y
sh
e
is
a
tt
a
c
h
e
d
to
M
a
lay
si
a
In
stit
u
t
e
o
f
T
ra
n
sp
o
rt,
UiT
M
.
Dr
S
it
i
M
e
riam
Z
a
h
a
ri
,
s
h
e
is
a
S
e
n
io
r
L
e
c
tu
re
r
in
Ce
n
tre
o
f
S
tatisti
c
s
a
n
d
De
c
isio
n
S
c
ien
c
e
S
tu
d
ies
in
F
a
c
u
lt
y
o
f
Co
m
p
u
ter
a
n
d
M
a
th
e
m
a
ti
c
a
l
S
c
ien
c
e
s,
Un
iv
e
rsiti
Tek
n
lo
g
i
M
AR
A
,
S
h
a
h
A
la
m
.
H
e
r
e
x
p
e
rti
se
is i
n
Eco
n
o
m
e
tri
c
s a
n
d
T
ime
S
e
ries
.
Evaluation Warning : The document was created with Spire.PDF for Python.