Co
m
pu
t
er
Science
a
nd
I
nfo
r
m
a
t
io
n T
ec
hn
o
lo
g
ies
Vo
l.
7
,
No
.
3
,
No
v
em
b
er
20
26
,
p
p
.
3
7
7
~
3
9
3
I
SS
N:
2722
-
3
2
2
1
,
DOI
:
1
0
.
1
1
5
9
1
/csi
t
.
v
7
i
3
.
p
3
7
7
-
3
9
3
377
J
o
ur
na
l
ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
p
r
ime.
co
m/in
d
ex
.
p
h
p
/csi
t
A lit
era
ture
revi
e
w of re
curren
t
n
e
ura
l net
wo
rk
app
ro
a
ches to
ma
la
ria
outbrea
k predic
tion in
Sub
-
Sa
ha
ra
n Af
ri
ca
So
ph
ia
T
em
bu
re
,
Well
ing
t
o
n S.
M
a
njo
ro
D
e
p
a
r
t
me
n
t
o
f
S
o
f
t
w
a
r
e
E
n
g
i
n
e
e
r
i
n
g
,
S
c
h
o
o
l
o
f
I
n
f
o
r
ma
t
i
o
n
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
H
a
r
a
r
e
I
n
st
i
t
u
t
e
o
f
T
e
c
h
n
o
l
o
g
y
,
H
a
r
a
r
e
,
Zi
m
b
a
b
w
e
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
May
16
,
2
0
2
6
R
ev
is
ed
J
u
l
7
,
2
0
2
6
Acc
ep
ted
J
u
l
13
,
2
0
2
6
M
a
laria
re
m
a
in
s
a
sig
n
ifi
c
a
n
t
p
u
b
li
c
h
e
a
lt
h
issu
e
i
n
S
u
b
-
S
a
h
a
r
a
n
Afric
a
.
P
re
d
ictiv
e
m
o
d
e
ll
in
g
is
in
c
re
a
sin
g
ly
v
iew
e
d
a
s
a
wa
y
to
m
o
v
e
fro
m
c
o
u
n
ti
n
g
c
a
se
s
re
a
c
ti
v
e
ly
to
p
re
p
a
ri
n
g
p
ro
a
c
ti
v
e
ly
fo
r
o
u
tb
re
a
k
s.
T
h
is l
it
e
ra
t
u
re
re
v
iew
e
x
a
m
in
e
s
p
e
e
r
-
re
v
iew
e
d
re
se
a
rc
h
p
u
b
li
sh
e
d
fro
m
1
9
9
7
t
o
2
0
2
5
.
It
c
o
v
e
rs
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
s
(
RNN
s
)
m
e
th
o
d
s,
a
tt
e
n
ti
o
n
m
e
c
h
a
n
is
m
s,
m
u
lt
i
-
so
u
rc
e
d
a
ta
i
n
teg
ra
ti
o
n
,
a
n
d
h
e
a
lt
h
-
sy
ste
m
in
fo
rm
a
ti
c
s
re
lev
a
n
t
t
o
m
a
laria
p
re
d
ictio
n
in
lo
w
-
a
n
d
m
id
d
le
-
in
c
o
m
e
c
o
u
n
t
ries
.
A
th
o
ro
u
g
h
se
a
rc
h
th
r
o
u
g
h
G
o
o
g
le
S
c
h
o
lar,
IE
EE
Xp
l
o
re
,
S
c
ien
c
e
Dire
c
t,
P
u
b
M
e
d
,
a
n
d
S
p
rin
g
e
rLi
n
k
id
e
n
ti
fie
d
1
8
7
c
a
n
d
i
d
a
te
p
a
p
e
r
s
.
Afte
r
re
v
iew
in
g
ti
tl
e
s
a
n
d
a
b
stra
c
ts,
4
2
p
a
p
e
rs
m
e
t
th
e
i
n
c
lu
si
o
n
c
r
it
e
ria.
Th
e
re
v
iew
o
rg
a
n
ize
s
t
h
e
se
lec
ted
stu
d
ies
in
t
o
six
m
a
in
c
a
teg
o
ries
:
c
li
m
a
te
-
d
ise
a
s
e
e
c
o
lo
g
y
,
sta
ti
stica
l
fo
re
c
a
stin
g
,
m
a
c
h
in
e
lea
rn
i
n
g
(M
L)
b
a
se
li
n
e
s,
RNN
s
,
a
tt
e
n
ti
o
n
m
e
c
h
a
n
ism
s,
a
n
d
h
e
a
lt
h
-
sy
ste
m
d
a
ta
in
te
g
ra
ti
o
n
.
A
c
o
m
p
a
ra
ti
v
e
m
a
tri
x
h
ig
h
li
g
h
ts
th
e
sim
il
a
rit
ies
a
n
d
d
if
fe
re
n
c
e
s
in
m
e
th
o
d
s
a
c
ro
ss
twe
n
ty
re
p
re
se
n
tativ
e
stu
d
ies
.
Th
e
re
v
iew
p
o
i
n
ts
o
u
t
fo
u
r
p
e
rsi
ste
n
t
g
a
p
s:
t
h
e
lac
k
o
f
a
tt
e
n
ti
o
n
-
a
u
g
m
e
n
ted
RNN
s
fo
r
m
a
laria
f
o
re
c
a
stin
g
i
n
S
o
u
th
e
rn
Afric
a
,
li
m
it
e
d
in
teg
ra
ti
o
n
o
f
h
e
a
lt
h
-
fa
c
il
it
y
in
fra
stru
c
t
u
re
fe
a
tu
re
s
with
c
li
m
a
te
p
re
d
ict
o
rs,
i
n
a
d
e
q
u
a
te
h
a
n
d
l
in
g
o
f
m
issi
n
g
d
a
ta
in
Afri
c
a
n
sa
telli
te
-
d
e
riv
e
d
c
li
m
a
te
se
ri
e
s,
a
n
d
t
h
e
a
b
se
n
c
e
o
f
o
p
e
ra
ti
o
n
a
l
d
a
sh
b
o
a
rd
s
th
a
t
m
a
k
e
m
o
d
e
l
o
u
t
p
u
ts
u
sa
b
le
fo
r
d
istri
c
t
h
e
a
lt
h
o
ff
ice
rs
with
o
u
t
s
tatisti
c
a
l
train
in
g
.
F
u
t
u
re
re
se
a
rc
h
sh
o
u
l
d
fo
c
u
s o
n
c
ro
ss
-
c
o
u
n
tr
y
tran
sfe
ra
b
i
li
ty
st
u
d
ies
,
u
sin
g
g
ra
p
h
n
e
u
ra
l
n
e
two
rk
s t
o
c
a
p
tu
re
sp
a
ti
a
l
sp
il
lo
v
e
r,
a
n
d
re
a
l
-
ti
m
e
in
teg
ra
ti
o
n
with
n
a
ti
o
n
a
l
su
rv
e
il
lan
c
e
sy
ste
m
s li
k
e
d
istri
c
t
h
e
a
lt
h
in
f
o
rm
a
ti
o
n
s
o
ftwa
re
2
(
DH
IS
2
)
.
K
ey
w
o
r
d
s
:
Atten
tio
n
m
ec
h
an
is
m
E
ar
ly
war
n
in
g
s
y
s
tem
Ma
lar
ia
p
r
ed
ictio
n
Mu
lti
-
s
o
u
r
ce
d
ata
f
u
s
io
n
R
ec
u
r
r
en
t n
eu
r
al
n
etwo
r
k
s
Su
b
-
Sah
ar
an
Af
r
ica
Z
im
b
ab
we
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
:
W
ellin
g
to
n
S.
Ma
n
jo
r
o
Dep
ar
tm
en
t o
f
So
f
twar
e
E
n
g
in
ee
r
in
g
,
Sch
o
o
l o
f
I
n
f
o
r
m
atio
n
Scien
ce
an
d
T
ec
h
n
o
l
o
g
y
Har
ar
e
I
n
s
titu
te
o
f
T
ec
h
n
o
l
o
g
y
Har
ar
e,
Z
im
b
ab
we
E
m
ail: w
m
an
jo
r
o
@
h
it.a
c.
zw
1.
I
NT
RO
D
UCT
I
O
N
On
e
o
f
Su
b
-
Sah
ar
an
Af
r
ica’
s
m
o
s
t
en
d
u
r
in
g
in
f
ec
tio
u
s
-
d
is
ea
s
e
ch
allen
g
es
is
m
alar
ia.
An
esti
m
ated
249
m
illi
o
n
m
alar
ia
ca
s
es
an
d
ar
o
u
n
d
608
,
0
0
0
d
ea
th
s
we
r
e
f
o
u
n
d
b
y
th
e
W
o
r
ld
Hea
lth
O
r
g
an
izatio
n
(
W
HO)
in
th
e
y
ea
r
of
2
0
2
3
[
1
]
.
T
h
e
W
HO
Af
r
ican
R
eg
io
n
s
aw
ab
o
u
t
9
4
%
o
f
th
e
m
alar
ia
ca
s
es.
R
ef
lectin
g
a
r
is
in
g
tr
en
d
a
n
d
a
n
in
c
r
ea
s
e
o
f
1
1
m
illi
o
n
ca
s
es
co
m
p
ar
e
d
to
2
0
2
2
,
th
ese
f
ig
u
r
es
s
h
o
w
t
h
at
th
e
ca
s
e
-
co
u
n
tin
g
s
y
s
tem
s
th
at
ar
e
s
et
u
p
ca
n
n
o
t p
r
ev
en
t o
u
t
b
r
ea
k
s
f
r
o
m
escalatin
g
.
Acc
o
r
d
in
g
to
th
e
W
HO,
7
5
o
f
m
alar
ia
d
ea
th
am
o
n
g
Af
r
ica
n
s
u
n
d
er
t
h
e
a
g
e
o
f
f
iv
e.
T
h
at
is
ch
ild
r
en
l
ess
th
an
f
iv
e
y
ea
r
s
o
ld
ac
co
u
n
ts
f
o
r
7
5
%
o
f
all
m
alar
ia
d
ea
th
in
Af
r
ica.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
3
2
2
1
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
,
Vo
l.
7
,
No
.
3
,
No
v
em
b
er
20
26
:
3
7
7
-
393
378
Z
im
b
ab
we
lies
with
in
th
e
ar
ea
o
f
Su
b
-
Sah
ar
an
Af
r
ica
w
it
h
h
ig
h
m
alar
ia
tr
an
s
m
is
s
io
n
an
d
h
as
co
n
f
ir
m
e
d
ca
s
e
r
ates
o
f
1
5
-
2
9
ca
s
es
p
er
1
,
0
0
0
p
o
p
u
latio
n
at
-
r
is
k
an
n
u
ally
.
R
u
r
al
d
is
tr
icts
t
h
at
ar
e
lo
w
altitu
d
e
lik
e
Hu
r
u
n
g
we,
Mb
ir
e
an
d
C
h
ip
in
g
e
o
f
ten
h
a
v
e
ca
s
e
r
ates
s
ev
er
al
tim
es
ab
o
v
e
th
e
n
atio
n
al
av
er
ag
e.
Desp
ite
p
r
o
lo
n
g
ed
a
p
p
licatio
n
o
f
i
n
s
ec
ticid
e
-
tr
ea
ted
n
ets
(
I
T
Ns)
a
n
d
s
ea
s
o
n
al
ch
em
o
p
r
ev
e
n
tio
n
p
r
o
g
r
a
m
m
es
in
th
e
d
is
tr
icts
o
f
W
e
s
t
an
d
C
en
tr
al
Af
r
ica,
tr
an
s
m
is
s
io
n
r
em
ain
s
u
n
s
tab
le
an
d
h
ig
h
ly
clim
ate
-
d
r
iv
en
.
A
p
r
im
a
r
y
o
p
er
atio
n
al
b
o
ttlen
ec
k
is
th
e
t
im
e
g
ap
b
etwe
en
w
h
en
tr
an
s
m
is
s
io
n
b
eg
in
s
an
d
wh
en
d
is
tr
ict
h
ea
lth
team
s
ca
n
r
esp
o
n
d
in
a
co
o
r
d
in
ated
m
an
n
er
,
with
co
m
p
u
tatio
n
al
ea
r
ly
-
war
n
in
g
s
y
s
tem
s
u
n
iq
u
el
y
ab
l
e
to
r
ed
u
ce
th
is
g
ap
.
Asi
d
e
f
r
o
m
its
ep
i
d
em
io
lo
g
ic
al
s
ig
n
if
ican
ce
th
is
r
e
v
iew
f
al
ls
at
th
e
in
ter
s
ec
tio
n
o
f
two
b
u
r
g
eo
n
in
g
r
esear
ch
co
m
m
u
n
ities
,
ar
tific
i
al
in
tellig
en
ce
(
AI
)
f
o
r
s
o
cial
g
o
o
d
a
n
d
p
u
b
lic
h
ea
lth
in
f
o
r
m
atics.
T
h
e
AI
f
o
r
s
o
cial
g
o
o
d
i
n
itiativ
e
in
v
o
l
v
es
th
e
u
tili
za
tio
n
o
f
in
n
o
v
ativ
e
m
ac
h
in
e
lear
n
in
g
(
ML
)
in
th
e
cr
ea
tio
n
an
d
im
p
lem
en
tatio
n
o
f
tech
n
iq
u
es
f
o
r
s
cien
tis
ts
d
ea
lin
g
with
d
ata
an
d
attem
p
tin
g
to
s
o
lv
e
p
r
o
b
lem
s
o
f
s
o
cial
im
p
o
r
tan
ce
.
T
h
e
em
p
lo
y
m
en
t
o
f
atten
tio
n
-
au
g
m
en
ted
r
ec
u
r
r
en
t
n
e
u
r
al
n
etwo
r
k
s
(
R
NN
s
)
f
o
r
th
e
p
u
r
p
o
s
e
o
f
m
alar
ia
p
r
ed
ictio
n
in
So
u
th
e
r
n
Af
r
ica
is
a
p
ar
ad
ig
m
atic
ex
am
p
le
o
f
th
is
ag
en
d
a:
i
t
co
m
b
in
es
m
eth
o
d
o
lo
g
ical
in
n
o
v
atio
n
in
d
ee
p
lear
n
in
g
with
a
p
r
ess
in
g
h
u
m
a
n
itar
ian
ch
allen
g
e,
i
n
a
g
e
o
g
r
ap
h
ical
s
ettin
g
wh
er
e
co
m
p
u
tatio
n
al
to
o
ls
ar
e
m
o
s
t
n
ee
d
ed
b
u
t
least
u
s
ed
.
Fr
o
m
t
h
e
p
o
in
t
o
f
v
iew
o
f
p
u
b
lic
h
ea
lth
in
f
o
r
m
atics,
o
n
e
o
f
th
e
m
ain
o
p
en
p
r
o
b
lem
s
in
th
e
f
ield
is
th
e
in
teg
r
atio
n
o
f
h
eter
o
g
en
e
o
u
s
d
ata
s
tr
ea
m
s
s
atellite
clim
ate
d
ata,
d
is
tr
ict
h
ea
lth
in
f
o
r
m
atio
n
s
o
f
t
war
e
2
(
DHI
S2
)
s
u
r
v
eillan
ce
an
d
h
ea
lth
f
ac
ilit
y
in
f
r
astru
ctu
r
e
d
ata
in
to
u
n
if
ied
p
r
ed
ictio
n
f
r
am
ew
o
r
k
s
.
L
o
o
k
in
g
at
th
is
r
e
v
iew
f
r
o
m
t
h
e
p
er
s
p
ec
tiv
e
o
f
b
o
t
h
c
o
m
m
u
n
ities
b
r
o
ad
e
n
s
its
ap
p
ea
l
an
d
en
h
an
ce
s
th
e
p
o
ten
tial f
o
r
id
en
tifie
d
g
ap
s
to
s
tim
u
late
f
u
t
u
r
e
cr
o
s
s
-
d
is
cip
lin
ar
y
r
esear
ch
.
T
h
e
ab
ilit
y
t
o
p
r
ed
ict
m
alar
ia
o
u
tb
r
ea
k
s
is
n
o
t
n
ew.
Sin
ce
t
h
e
late
1
8
0
0
s
,
th
e
r
e
h
as
b
ee
n
e
v
id
en
ce
o
f
th
e
b
io
lo
g
ical
co
n
n
ec
tio
n
b
et
wee
n
clim
ate
d
r
iv
er
s
p
r
im
ar
il
y
p
r
ec
ip
itatio
n
,
tem
p
er
atu
r
e,
a
n
d
h
u
m
id
ity
an
d
th
e
tr
an
s
m
is
s
io
n
o
f
v
ec
to
r
-
b
o
r
n
e
d
is
ea
s
es.
I
n
ad
d
itio
n
,
a
co
n
s
id
er
ab
le
am
o
u
n
t
o
f
liter
atu
r
e
h
a
s
d
r
awn
m
ec
h
an
is
tic
m
o
d
els
o
f
m
o
s
q
u
ito
p
o
p
u
lati
o
n
d
y
n
a
m
ics
a
n
d
p
ar
asit
e
in
cu
b
atio
n
.
T
h
e
s
ea
s
o
n
al
au
to
r
eg
r
ess
iv
e
in
teg
r
ated
m
o
v
in
g
av
e
r
ag
e
(
SAR
I
MA
)
m
o
d
els
h
a
v
e
b
ee
n
p
o
p
u
lar
c
h
o
ice
f
o
r
o
p
er
atio
n
al
ea
r
ly
-
war
n
in
g
s
y
s
tem
s
i
n
lo
w
-
r
eso
u
r
ce
en
v
ir
o
n
m
en
ts
f
o
r
a
lo
n
g
tim
e.
B
u
t
SAR
I
MA
an
d
o
th
er
s
u
c
h
s
tatis
tical
m
o
d
els
r
eq
u
ir
e
lin
ea
r
ity
an
d
s
tatio
n
ar
ity
o
f
th
e
u
n
d
er
l
y
in
g
tim
e
s
er
ies;
b
o
t
h
o
f
wh
i
ch
ar
e
b
ein
g
i
n
cr
ea
s
in
g
ly
v
io
l
ated
d
u
e
t
o
clim
ate
v
ar
iab
ilit
y
,
d
r
u
g
r
esis
tan
ce
,
ex
p
an
d
in
g
co
v
e
r
ag
e
o
f
in
ter
v
e
n
tio
n
s
an
d
p
o
p
u
latio
n
m
o
v
em
en
t
.
W
ith
th
e
ad
v
en
t
o
f
d
ee
p
le
ar
n
in
g
in
r
ec
en
t
y
e
ar
s
,
p
ar
ti
cu
lar
ly
R
NN/
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
,
a
n
ew
f
r
o
n
tier
h
as
b
ee
n
o
p
e
n
ed
f
o
r
m
ala
r
ia
f
o
r
e
ca
s
tin
g
.
L
STM
n
etwo
r
k
s
s
o
l
v
e
th
e
p
r
o
b
lem
o
f
v
an
is
h
in
g
g
r
ad
ie
n
t
th
at
af
f
e
cted
ea
r
lier
r
ec
u
r
r
en
t
m
o
d
el
s
,
allo
win
g
s
u
ch
m
o
d
els
to
ca
p
tu
r
e
tem
p
o
r
al
d
ep
en
d
e
n
cies
o
v
er
s
eq
u
en
ce
s
o
f
d
o
ze
n
s
to
h
u
n
d
r
ed
s
o
f
t
im
e
s
tep
s
.
T
h
er
e
ar
e
s
til
l
la
r
g
e
g
ap
s
p
r
esen
t.
Mb
u
n
g
e
et
a
l
.
[
2
]
n
o
ted
th
at
th
e
Af
r
ican
liter
atu
r
e
o
n
AI
-
b
as
ed
m
alar
ia
p
r
ed
ictio
n
in
h
u
m
a
n
s
s
u
f
f
er
s
f
r
o
m
two
cr
itical
g
ap
s
:
i
)
th
e
ab
s
en
ce
o
f
atten
tio
n
m
ec
h
a
n
is
m
s
;
a
nd
ii
)
th
e
lack
o
f
m
u
lti
-
s
o
u
r
c
e
d
ata
co
m
b
in
atio
n
,
s
p
ec
if
ically
clim
ate
p
r
ed
icto
r
s
with
h
ea
lth
f
ac
ilit
y
in
f
r
astru
ctu
r
e
f
ea
tu
r
es.
T
h
is
r
ev
iew
h
as
th
e
n
o
v
elty
o
f
s
y
n
th
esis
in
g
th
r
ee
s
tr
an
d
s
o
f
liter
atu
r
e
th
at
a
r
e
r
ar
ely
ex
am
in
e
d
to
g
eth
er
:
clim
ate
-
d
is
ea
s
e
ec
o
lo
g
y
,
d
ee
p
lear
n
in
g
ar
ch
itectu
r
es
(
R
NNs,
at
ten
tio
n
m
ec
h
an
is
m
s
)
,
an
d
h
ea
lth
-
s
y
s
tem
in
f
o
r
m
atics.
T
o
t
h
e
b
est
o
f
o
u
r
k
n
o
wled
g
e,
n
o
p
r
e
v
io
u
s
r
ev
iews
h
av
e
ass
ess
ed
th
e
ex
ten
t
to
wh
ich
atten
tio
n
-
au
g
m
e
n
ted
R
NNs
ca
n
f
o
r
ec
ast
m
alar
ia
o
u
tb
r
ea
k
s
in
So
u
th
e
r
n
Af
r
ica
o
r
th
e
f
e
asib
ilit
y
o
f
m
er
g
in
g
DHI
S2
h
ea
lth
f
ac
ilit
y
d
ata
w
ith
s
atellite
-
d
er
iv
ed
clim
ate
p
r
ed
icto
r
s
in
a
s
in
g
le
d
ee
p
le
ar
n
in
g
f
r
am
ewo
r
k
.
T
h
e
g
a
p
h
as
clo
s
ed
with
th
is
r
ev
iew
wh
ich
o
f
f
er
s
s
y
s
tem
atic
ev
id
en
ce
b
ased
in
tellectu
al
f
o
u
n
d
atio
n
f
o
r
t
h
e
MT
ec
h
r
esear
ch
p
r
o
ject
at
th
e
Har
ar
e
I
n
s
titu
te
o
f
T
ec
h
n
o
l
o
g
y
.
Aim
an
d
o
b
jectiv
es
o
f
th
e
r
ev
iew
:
t
h
is
liter
atu
r
e
r
ev
iew
aim
s
to
cr
itically
r
ev
iew
th
e
b
o
d
y
o
f
ac
ad
em
ic
liter
atu
r
e
r
elev
an
t
to
ap
p
ly
i
n
g
R
NN
s
to
m
alar
i
a
o
u
tb
r
ea
k
p
r
ed
ictio
n
in
Su
b
-
Sah
ar
an
A
f
r
ican
s
ettin
g
s
,
an
d
to
id
en
tify
r
esea
r
ch
g
a
p
s
th
at
m
o
tiv
ate
th
e
c
u
r
r
en
t
p
r
o
ject.
I
n
p
ar
ticu
lar
,
th
e
r
ev
iew
aim
s
to
:
i
)
e
x
p
lain
th
e
th
eo
r
etica
l
lin
k
b
etwe
en
clim
ate
v
ar
iab
l
es
an
d
th
e
tr
an
s
m
is
s
io
n
d
y
n
am
ics
o
f
m
alar
ia
;
ii
)
d
escr
ib
e
th
e
d
e
v
elo
p
m
e
n
t
o
f
f
o
r
ec
asti
n
g
m
eth
o
d
s
f
r
o
m
t
r
a
d
itio
n
al
s
tatis
tical
m
o
d
els
to
co
n
tem
p
o
r
ar
y
d
ee
p
-
lear
n
in
g
ar
c
h
itectu
r
es
;
iii
)
c
o
m
p
ar
e
th
e
s
tr
en
g
t
h
s
an
d
lim
it
atio
n
s
o
f
R
NN
v
ar
i
an
ts
(
L
STM
,
g
ated
r
ec
u
r
r
en
t
u
n
it
(
GR
U)
an
d
th
e
co
n
tr
ib
u
ti
o
n
o
f
atten
tio
n
m
ec
h
an
is
m
s
to
tim
e
-
s
er
ies
p
r
ed
ictio
n
;
iv
)
e
v
alu
ate
th
e
d
eg
r
ee
to
wh
ich
cu
r
r
en
t
wo
r
k
in
c
o
r
p
o
r
a
tes
m
u
lti
-
s
o
u
r
ce
d
ata
(
clim
ate,
ep
id
em
io
lo
g
y
,
an
d
h
ea
lth
-
s
y
s
tem
in
f
r
astru
ctu
r
e)
with
in
a
s
in
g
le
p
r
ed
ictio
n
f
r
a
m
ewo
r
k
;
a
n
d
v
)
i
d
en
tify
g
ap
s
an
d
in
c
o
n
s
is
ten
cies
in
th
e
liter
atu
r
e
wh
ich
r
eq
u
ir
e
f
u
r
th
er
i
n
v
esti
g
atio
n
.
Sco
p
e
an
d
s
ig
n
if
ican
ce
:
th
e
r
e
v
iew
is
lim
ited
alo
n
g
th
r
ee
d
im
en
s
io
n
s
.
I
t is g
eo
g
r
ap
h
ically
f
o
cu
s
ed
o
n
Su
b
-
Sah
ar
an
Af
r
ica
with
a
p
ar
t
icu
lar
f
o
cu
s
o
n
So
u
th
e
r
n
Af
r
i
ca
n
co
u
n
tr
ies (
Z
im
b
a
b
we,
Mo
za
m
b
iq
u
e,
Z
am
b
ia,
B
o
ts
wan
a,
an
d
So
u
th
Af
r
ica)
b
u
t d
r
awin
g
o
n
r
ele
v
an
t c
o
m
p
ar
ativ
e
wo
r
k
f
r
o
m
E
ast an
d
W
est Af
r
ica
an
d
o
th
er
tr
o
p
ical
r
eg
i
o
n
s
s
u
ch
as
So
u
t
h
Asi
a
an
d
L
atin
Am
er
ica.
I
n
tim
e
th
e
r
ev
iew
s
p
a
n
s
p
u
b
licatio
n
s
f
r
o
m
1
9
9
7
wh
en
th
e
s
em
in
al
L
STM
p
ap
er
was
p
u
b
lis
h
ed
b
y
Ho
c
h
r
eit
er
an
d
Sch
m
id
h
u
b
e
r
[
3
]
to
s
tu
d
ies
p
u
b
lis
h
ed
in
ea
r
ly
2
0
2
5
.
T
h
is
r
ev
iew
is
p
r
i
m
ar
ily
m
eth
o
d
o
lo
g
ical,
with
e
m
p
h
asis
o
n
q
u
an
titativ
e
f
o
r
ec
asti
n
g
m
eth
o
d
s
,
an
d
ex
clu
d
es
q
u
alitativ
e
s
tu
d
ies
o
f
m
alar
ia
p
o
licy
o
r
b
eh
av
io
u
r
a
l
r
esear
ch
o
n
b
ed
-
n
et
u
s
e
an
d
tr
ea
tm
en
t
-
s
ee
k
in
g
,
u
n
less
s
u
ch
s
tu
d
ies p
r
o
v
id
e
n
e
ce
s
s
ar
y
b
ac
k
g
r
o
u
n
d
f
o
r
p
r
e
d
ictio
n
m
o
d
els.
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
I
SS
N:
2722
-
3
2
2
1
A
liter
a
tu
r
e
r
ev
ie
w
o
f
r
ec
u
r
r
e
n
t n
eu
r
a
l n
etw
o
r
k
a
p
p
r
o
a
ch
es
to
ma
la
r
ia
o
u
tb
r
ea
k
…
(
S
o
p
h
i
a
Temb
u
r
e
)
379
T
h
e
v
alu
e
o
f
th
is
r
ev
iew
lies
in
th
e
s
y
n
th
esis
o
f
f
r
ag
m
e
n
ted
s
tr
an
d
s
o
f
liter
atu
r
e
th
at
r
ar
ely
h
a
v
e
b
ee
n
co
n
s
id
er
ed
to
g
eth
e
r
u
n
til
n
o
w.
C
lim
ate
-
d
is
ea
s
e
ec
o
lo
g
y
,
ML
,
an
d
h
ea
lth
-
s
y
s
tem
in
f
o
r
m
atics
ar
e
u
s
u
ally
p
u
b
lis
h
ed
in
d
if
f
e
r
en
t
jo
u
r
n
als
an
d
h
eld
in
d
if
f
er
e
n
t
co
n
f
e
r
e
n
ce
v
en
u
es.
T
h
e
r
ev
iew
co
m
b
in
es
th
ese
v
iews
to
o
f
f
er
a
co
h
er
en
t
i
n
tellectu
al
f
r
am
ewo
r
k
f
o
r
a
m
u
lti
-
s
o
u
r
ce
p
r
ed
ictio
n
s
y
s
tem
f
o
r
Z
im
b
a
b
wea
n
co
n
d
itio
n
s
.
Nu
m
er
o
u
s
s
tu
d
ies
h
av
e
lo
o
k
ed
in
to
ML
a
p
p
licatio
n
s
in
m
alar
ia
a
n
d
in
f
ec
tio
u
s
d
is
ea
s
e
p
r
ed
ictio
n
.
Sh
r
ag
ai
et
a
l
.
[
4
]
lo
o
k
ed
in
t
o
clim
ate
-
b
ased
m
o
d
els
f
o
r
p
r
ed
ictin
g
v
ec
to
r
-
b
o
r
n
e
d
is
ea
s
e
o
u
tb
r
ea
k
s
,
wh
ile
C
ar
v
ajal
et
a
l
.
[
5
]
r
ev
iewe
d
d
ee
p
lear
n
in
g
f
o
r
d
e
n
g
u
e
e
p
id
e
m
ic
f
o
r
ec
asti
n
g
.
No
n
et
h
eless
,
th
ese
r
ev
iews
ar
e
f
u
n
d
am
e
n
tally
d
i
f
f
er
en
t
f
r
o
m
th
e
p
r
esen
t
w
o
r
k
in
th
r
ee
wa
y
s
.
T
o
b
eg
in
with
,
th
e
r
e
is
n
o
s
p
ec
if
ic
f
o
cu
s
o
n
R
NN
ar
ch
itectu
r
es
f
o
r
m
alar
ia
in
Su
b
-
Sah
ar
a
n
Af
r
ica.
Seco
n
d
,
n
o
o
n
e
ex
a
m
in
es
th
e
p
r
e
p
ar
ed
n
ess
o
f
atten
tio
n
m
ec
h
an
is
m
s
–
alr
ea
d
y
co
m
m
o
n
p
lace
in
n
atu
r
al
lan
g
u
a
g
e
p
r
o
ce
s
s
in
g
(
NL
P
)
an
d
f
in
an
cial
tim
e
-
s
er
ies
f
o
r
ec
asti
n
g
–
f
o
r
u
s
e
with
m
alar
ia
s
u
r
v
eillan
ce
d
ata
f
r
o
m
So
u
t
h
er
n
Af
r
ica.
Fu
r
th
er
m
o
r
e,
p
ast wo
r
k
h
as n
o
t u
s
ed
h
ea
lth
f
ac
ilit
y
in
f
r
astru
ct
u
r
e
d
ata
(
f
ac
ilit
y
d
en
s
ity
,
b
e
d
ca
p
a
city
,
an
d
g
eo
g
r
a
p
h
ic
ac
ce
s
s
)
,
as
m
o
d
el
co
v
a
r
iates
alo
n
g
with
clim
ate
p
r
ed
icto
r
s
.
T
h
is
r
ev
iew
s
y
n
th
esizes
R
NN
ar
ch
itectu
r
es,
atten
tio
n
m
ec
h
an
is
m
s
an
d
h
ea
lth
-
s
y
s
tem
in
f
o
r
m
atics
in
a
So
u
th
er
n
Af
r
ican
m
alar
ia
p
r
e
d
ictio
n
co
n
tex
t
f
o
r
th
e
f
ir
s
t
tim
e.
I
n
d
o
in
g
s
o
,
it
p
r
o
v
i
d
es
a
s
tr
u
ctu
r
ed
ev
id
en
ce
b
ase
wh
ich
m
o
tiv
ates
an
d
ju
s
tifie
s
th
e
p
r
o
p
o
s
ed
M
.
T
ec
h
.
r
esear
ch
d
esig
n
at
th
e
Har
ar
e
I
n
s
titu
te
o
f
T
ec
h
n
o
lo
g
y
.
2.
M
E
T
H
O
D
2
.
1
.
Sea
rc
h
s
t
ra
t
eg
y
W
e
p
e
r
f
o
r
m
e
d
a
s
y
s
t
e
m
at
i
c
s
ea
r
c
h
i
n
f
i
v
e
a
c
a
d
e
m
i
c
d
at
a
b
a
s
e
s
f
r
o
m
J
a
n
u
a
r
y
t
o
M
a
r
c
h
2
0
2
6
.
P
r
i
m
a
r
y
s
e
a
r
c
h
t
e
r
m
s
w
e
r
e
c
o
m
b
i
n
e
d
u
s
i
n
g
B
o
o
l
ea
n
o
p
e
r
a
t
o
r
s
w
it
h
i
n
a
c
o
n
t
r
o
l
l
e
d
v
o
c
a
b
u
l
a
r
y
.
T
h
e
b
a
s
e
q
u
e
r
y
t
e
m
p
l
a
te
u
s
e
d
w
a
s
:
(
“
m
a
l
a
r
i
a
”
OR
“
v
ec
t
o
r
b
o
r
n
e
d
i
s
e
a
s
e
”
)
A
ND
(
“
p
r
e
d
i
c
t
i
o
n
”
OR
“
f
o
r
e
c
a
s
t
i
n
g
”
OR
“
e
a
r
l
y
w
a
r
n
i
n
g
”
)
A
N
D
(
“
n
e
u
r
a
l
n
e
t
w
o
r
k
”
O
R
“
L
S
T
M
”
OR
“
GR
U
”
OR
“
d
e
e
p
l
e
a
r
n
i
n
g
”
O
R
“
m
a
c
h
i
n
e
l
e
a
r
n
in
g
”
O
R
“
a
t
t
e
n
ti
o
n
”
)
A
N
D
(
“
A
f
r
i
c
a
”
O
R
“
S
u
b
-
S
a
h
ar
a
n
”
O
R
“
t
r
o
p
i
c
a
l
”
OR
“
l
o
w
an
d
m
i
d
d
l
e
i
n
c
o
m
e
”
)
.
T
h
e
s
e
m
in
a
l
p
a
p
e
r
s
i
d
e
n
ti
f
i
e
d
i
n
t
h
e
f
i
r
s
t
p
a
s
s
w
e
r
e
c
i
t
at
i
o
n
-
t
r
a
c
e
d
b
a
c
k
w
a
r
d
s
a
n
d
f
o
r
w
a
r
d
s
.
T
h
e
r
e
t
a
i
n
e
d
a
r
t
i
c
le
s
w
e
r
e
s
c
r
e
e
n
e
d
f
o
r
f
u
r
t
h
e
r
r
e
l
e
v
a
n
c
e
b
y
s
c
r
e
e
n
i
n
g
p
a
p
e
r
s
c
i
t
e
d
f
r
e
q
u
e
n
t
l
y
b
y
t
h
e
r
e
t
ai
n
e
d
a
r
t
i
c
l
es
a
n
d
p
a
p
e
r
s
ci
t
i
n
g
t
h
e
r
e
t
a
i
n
e
d
a
r
t
ic
l
es
.
2
.2
.
Da
t
a
ba
s
es
c
o
ns
ulte
d
T
h
e
s
h
o
r
tco
m
i
n
g
s
ass
o
ciate
d
with
th
is
r
ev
iew
co
n
s
tr
ain
t
h
e
r
ep
r
esen
tativ
en
ess
o
f
t
h
e
s
tu
d
ies
it
en
co
m
p
ass
es.
T
o
b
eg
in
with
,
th
e
s
ea
r
ch
was
lim
ited
to
E
n
g
lis
h
-
lan
g
u
a
g
e
p
u
b
licatio
n
s
.
T
h
is
m
ay
h
av
e
ex
clu
d
ed
r
elev
an
t
s
tu
d
ies
p
u
b
lis
h
ed
in
Fre
n
ch
,
Po
r
tu
g
u
ese
o
r
Swah
ili.
T
h
ese
lan
g
u
ag
es
ar
e
co
m
m
o
n
i
n
h
ig
h
-
b
u
r
d
e
n
m
alar
ia
co
u
n
tr
ies
lik
e
th
e
Dem
o
cr
atic
R
ep
u
b
lic
o
f
C
o
n
g
o
M
o
za
m
b
iq
u
e
an
d
T
an
za
n
ia.
W
r
iter
s
o
f
Fre
n
ch
-
lan
g
u
ag
e
liter
atu
r
e
f
r
o
m
Fra
n
co
p
h
o
n
e
W
est
Af
r
ica
wo
u
ld
th
er
e
f
o
r
e
n
o
t
b
e
w
id
ely
r
ep
r
esen
ted
.
Sy
s
tem
s
b
ased
o
n
SA
R
I
MA
h
av
e
b
ee
n
wid
ely
d
ep
l
o
y
ed
.
An
o
th
er
lim
itatio
n
is
th
at
th
e
f
iv
e
d
atab
ases
s
ea
r
ch
ed
Go
o
g
le
Sch
o
lar
,
I
E
E
E
Xp
lo
r
e,
Scien
ce
Dir
ec
t,
Pu
b
Me
d
,
an
d
Sp
r
in
g
er
L
in
k
.
T
h
e
d
atab
ases
co
n
s
u
lted
,
th
eir
co
r
r
esp
o
n
d
in
g
UR
L
s
,
an
d
th
e
s
ea
r
ch
p
e
r
io
d
ar
e
s
u
m
m
ar
is
ed
in
T
ab
le
1
.
W
h
ile
th
ese
s
o
u
r
ce
s
p
r
o
v
id
e
b
r
o
ad
co
v
e
r
ag
e
o
f
th
e
a
v
ailab
l
e
liter
atu
r
e,
t
h
ey
d
o
n
o
t
in
clu
d
e
ev
er
y
Af
r
ica
n
r
e
g
io
n
al
jo
u
r
n
al.
So
m
e
r
eg
io
n
al
jo
u
r
n
als
in
clu
d
e
p
r
ac
titi
o
n
er
-
l
ed
m
alar
ia
s
u
r
v
eillan
ce
s
tu
d
ie
s
wh
ich
d
o
n
o
t
m
ee
t
t
h
e
t
h
r
esh
o
ld
f
o
r
in
d
ex
i
n
g
i
n
an
y
o
f
th
e
m
ajo
r
d
atab
ases
.
Fu
r
th
er
m
o
r
e
,
p
u
b
licatio
n
b
ia
s
ag
ain
s
t
f
in
d
in
g
s
th
at
ar
e
n
o
t
p
o
s
itiv
e
o
r
s
tatis
tically
in
s
ig
n
if
ican
t
m
ea
n
s
th
at
f
ailed
f
o
r
ec
asti
n
g
ex
p
er
im
en
ts
wh
ic
h
ar
e
lik
ely
to
o
f
f
er
eq
u
ally
u
s
ef
u
l
m
et
h
o
d
o
lo
g
ical
in
s
ig
h
ts
will
n
o
t
co
n
tr
ib
u
te
to
t
h
e
ev
i
d
en
ce
th
at
s
tu
d
ies
s
y
n
th
esis
e.
R
ec
o
g
n
is
in
g
t
h
ese
lim
itatio
n
s
a
s
an
in
teg
r
al
p
a
r
t
o
f
s
y
s
tem
atic
liter
atu
r
e
r
ev
iews
m
ea
n
s
th
ey
d
o
n
o
t
in
v
alid
ate
th
e
f
in
d
in
g
s
.
Ho
wev
e
r
,
th
ey
d
o
s
u
g
g
est
th
e
f
o
u
r
r
esear
ch
g
ap
s
we
id
e
n
tifie
d
m
ay
b
e
lar
g
e
r
th
an
wh
at
th
e
in
cl
u
d
ed
liter
atu
r
e
in
d
icate
s
.
T
ab
le
1
.
Data
b
ases
co
n
s
u
lted
d
u
r
in
g
th
e
liter
atu
r
e
s
ea
r
c
h
D
a
t
a
b
a
s
e
U
R
L
D
a
t
e
r
a
n
g
e
se
a
r
c
h
e
d
G
o
o
g
l
e
S
c
h
o
l
a
r
sch
o
l
a
r
.
g
o
o
g
l
e
.
c
o
m
1
9
9
7
-
2
0
2
5
I
EEE
X
p
l
o
r
e
i
e
e
e
x
p
l
o
r
e
.
i
e
e
e
.
o
r
g
1
9
9
7
-
2
0
2
5
S
c
i
e
n
c
e
D
i
r
e
c
t
sci
e
n
c
e
d
i
r
e
c
t
.
c
o
m
1
9
9
7
-
2
0
2
5
P
u
b
M
e
d
p
u
b
m
e
d
.
n
c
b
i
.
n
l
m
.
n
i
h
.
g
o
v
1
9
9
7
-
2
0
2
5
S
p
r
i
n
g
e
r
Li
n
k
l
i
n
k
.
sp
r
i
n
g
e
r
.
c
o
m
1
9
9
7
-
2
0
2
5
2
.
3
.
I
nclus
io
n a
nd
ex
clus
io
n
cr
it
er
ia
2
.
3
.
1
.
I
nclus
io
n c
rit
er
ia
T
o
en
s
u
r
e
m
eth
o
d
o
lo
g
ical
co
n
s
is
ten
cy
an
d
r
elev
an
ce
,
e
x
p
l
icit
in
clu
s
io
n
cr
iter
ia
wer
e
estab
lis
h
ed
p
r
io
r
to
th
e
s
cr
ee
n
in
g
p
r
o
ce
s
s
.
T
h
ese
cr
iter
ia
wer
e
d
esig
n
ed
to
id
en
tif
y
s
tu
d
ies
th
at
d
ir
ec
tl
y
ad
d
r
ess
ed
m
alar
ia
p
r
ed
ictio
n
u
s
in
g
q
u
a
n
titativ
e
m
o
d
ellin
g
ap
p
r
o
a
c
h
es
wh
ile
m
ain
tain
in
g
s
u
f
f
icien
t
m
eth
o
d
o
lo
g
ical
q
u
ality
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
3
2
2
1
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
,
Vo
l.
7
,
No
.
3
,
No
v
em
b
er
20
26
:
3
7
7
-
393
380
On
ly
p
u
b
licatio
n
s
m
ee
tin
g
al
l
o
f
th
e
f
o
llo
win
g
c
r
iter
ia
wer
e
r
etain
ed
f
o
r
f
u
r
th
e
r
r
ev
iew
:
i
)
p
eer
-
r
e
v
iewe
d
jo
u
r
n
al
ar
ticles,
co
n
f
er
en
ce
p
ap
er
s
an
d
b
o
o
k
ch
ap
ter
s
p
u
b
l
is
h
ed
in
th
e
p
e
r
io
d
1
9
9
7
to
2
0
2
5
;
ii)
s
tu
d
ies
th
at
q
u
an
titativ
ely
p
r
e
d
ict
o
r
f
o
r
ec
ast
m
alar
ia
in
cid
en
ce
o
r
o
th
er
r
elate
d
m
etr
ics
f
o
r
v
e
cto
r
-
b
o
r
n
e
d
is
ea
s
es
;
iii)
p
ap
er
s
th
at
p
r
esen
t
ML
,
s
tatis
t
ical
o
r
h
y
b
r
id
m
o
d
ellin
g
ap
p
r
o
ac
h
es
th
at
ca
n
b
e
ap
p
lied
to
m
o
n
th
l
y
o
r
wee
k
ly
tim
e
-
s
er
ies
d
ata
;
i
v
)
s
tu
d
ies
with
d
atasets
f
r
o
m
Su
b
-
Sah
ar
an
Af
r
ica,
Asi
a
o
r
o
t
h
er
tr
o
p
ical
r
eg
i
o
n
s
with
s
im
ilar
ep
id
em
io
lo
g
ical
m
alar
ia
d
y
n
am
ics
;
an
d
v
)
m
eth
o
d
o
lo
g
ical
p
ap
er
s
o
n
R
NN
ar
ch
itectu
r
es
an
d
atten
tio
n
m
ec
h
an
is
m
s
,
n
o
t lim
ited
to
a
p
ar
ticu
lar
ap
p
licatio
n
d
o
m
ain
.
2
.
3
.
2
.
E
x
clus
io
n c
rit
er
ia
C
o
m
p
lem
en
tar
y
ex
clu
s
io
n
c
r
iter
ia
wer
e
ap
p
lied
to
r
em
o
v
e
s
tu
d
ies
th
at
d
id
n
o
t
alig
n
with
th
e
o
b
jectiv
es
o
f
th
is
r
e
v
iew.
T
h
ese
cr
iter
ia
h
elp
ed
e
n
s
u
r
e
th
at
th
e
f
in
al
ev
i
d
en
ce
b
ase
r
e
m
ain
ed
f
o
cu
s
ed
o
n
p
r
ed
ictiv
e
m
o
d
ellin
g
m
et
h
o
d
o
lo
g
ies
ap
p
lica
b
le
to
m
alar
ia
f
o
r
ec
asti
n
g
.
Pu
b
licatio
n
s
th
a
t
m
ee
t
an
y
o
f
t
h
e
f
o
llo
win
g
co
n
d
itio
n
s
wer
e
ex
clu
d
ed
f
r
o
m
th
e
r
ev
iew
:
i)
g
r
ey
liter
atu
r
e,
co
n
f
er
en
ce
p
r
o
ce
ed
in
g
s
ab
s
tr
ac
ts
with
o
u
t
f
u
ll
tex
t
an
d
e
d
ito
r
ial
co
m
m
en
tar
ies
;
ii)
s
tu
d
ies
r
estricte
d
to
m
o
lecu
lar
b
io
l
o
g
y
,
d
r
u
g
d
ev
el
o
p
m
en
t
o
r
v
ac
cin
e
tr
ials
,
with
o
u
t
a
co
m
p
o
n
en
t
o
f
p
r
ed
ictiv
e
m
o
d
ellin
g
;
iii)
p
ap
er
s
with
q
u
alitativ
e
f
in
d
in
g
s
o
n
l
y
(
f
o
c
u
s
g
r
o
u
p
d
is
cu
s
s
io
n
s
,
p
o
licy
r
ev
iews)
with
o
u
t
q
u
an
titativ
e
ev
alu
atio
n
;
iv
)
p
u
b
licatio
n
s
n
o
t
av
ailab
le
in
E
n
g
lis
h
;
an
d
v
)
d
u
p
licate
p
u
b
licatio
n
s
a
n
d
p
r
e
p
r
in
t
v
er
s
io
n
s
wh
en
th
e
p
ee
r
-
r
ev
iewe
d
v
er
s
io
n
was a
ls
o
av
ailab
le.
2
.
3
.
3
.
J
us
t
if
ica
t
io
n o
f
ex
clus
io
n c
rit
er
ia
E
x
clu
s
io
n
b
y
d
esig
n
o
f
q
u
alitativ
e
an
d
eth
n
o
g
r
ap
h
ic
s
tu
d
ies
s
h
o
u
ld
b
e
clea
r
ly
ju
s
tifie
d
.
Su
ch
s
tu
d
ies
c
an
p
r
o
v
i
d
e
v
alu
ab
le
co
n
tex
t
u
al
in
f
o
r
m
atio
n
o
n
h
ea
lth
-
s
ee
k
in
g
b
eh
a
v
io
u
r
a
n
d
co
m
m
u
n
i
ty
-
lev
el
d
r
iv
er
s
o
f
m
alar
ia
b
u
r
d
en
.
T
h
is
r
ev
iew
e
x
clu
d
ed
q
u
alitativ
e
r
esear
ch
f
o
r
two
p
r
in
cip
led
r
ea
s
o
n
s
.
Fir
s
t,
th
e
p
r
im
ar
y
aim
o
f
th
e
r
ev
iew
is
to
ev
al
u
ate
th
e
tech
n
ical
p
e
r
f
o
r
m
a
n
ce
a
n
d
ar
ch
itectu
r
al
c
h
ar
ac
ter
is
tics
o
f
co
m
p
u
te
r
ized
p
r
ed
ictio
n
m
o
d
els.
T
h
is
o
b
jec
tiv
e
r
eq
u
ir
es
q
u
an
titativ
e
o
u
tco
m
e
m
ea
s
u
r
es,
s
u
ch
as
r
o
o
t
m
ea
n
s
q
u
ar
e
er
r
o
r
(
R
MSE
)
,
m
ea
n
ab
s
o
lu
te
er
r
o
r
(
MA
E
)
,
an
d
th
e
ar
ea
u
n
d
e
r
th
e
r
ec
eiv
er
o
p
er
atin
g
ch
a
r
ac
ter
is
tic
(
AU
R
OC
)
,
wh
ich
q
u
alitativ
e
s
tu
d
ies,
b
y
d
esig
n
,
d
o
n
o
t
r
ep
o
r
t.
I
n
th
e
li
ter
atu
r
e
r
ev
iewe
d
,
m
alar
ia
p
r
e
d
ictio
n
m
o
d
els
ar
e
v
alid
ated
u
s
in
g
n
u
m
er
ically
q
u
an
tifie
d
s
u
r
v
eillan
ce
r
ec
o
r
d
s
,
wh
ile
co
n
tex
tu
al
f
ac
to
r
s
d
o
cu
m
e
n
ted
in
q
u
alitativ
e
s
tu
d
ies
h
av
e
y
et
to
b
e
in
co
r
p
o
r
ated
in
t
o
th
e
f
e
atu
r
e
en
g
in
ee
r
i
n
g
o
f
t
h
e
d
ee
p
lear
n
in
g
m
o
d
els
r
ev
iewe
d
.
Ne
v
er
th
eless
,
it
is
r
ec
o
g
n
ized
th
at
q
u
alitativ
e
e
v
id
en
ce
o
n
h
ea
lth
-
s
ee
k
in
g
b
e
h
av
io
u
r
,
h
ea
lth
ca
r
e
ac
ce
s
s
ib
ilit
y
,
an
d
co
m
m
u
n
ity
-
lev
el
p
r
ac
tices
co
u
ld
im
p
r
o
v
e
th
e
ec
o
lo
g
ical
v
alid
ity
o
f
f
u
tu
r
e
p
r
e
d
ictiv
e
m
o
d
els,
p
ar
ticu
lar
ly
in
u
n
d
er
-
r
eso
u
r
ce
d
s
ettin
g
s
wh
er
e
s
u
r
v
eillan
ce
d
ata
m
ay
s
y
s
tem
ati
ca
lly
u
n
d
er
esti
m
ate
th
e
tr
u
e
b
u
r
d
e
n
o
f
m
alar
ia
tr
an
s
m
is
s
io
n
.
2
.
4
.
Scre
ening
o
utc
o
m
es
T
h
e
f
ir
s
t
s
ea
r
ch
r
esu
lted
in
1
8
7
p
a
p
er
s
as
ca
n
d
id
ates
f
r
o
m
th
e
f
iv
e
d
atab
ases
.
T
h
e
2
3
d
u
p
licate
p
ap
er
s
an
d
1
8
p
ap
e
r
s
th
at
d
id
n
o
t
m
ee
t
th
e
in
clu
s
io
n
cr
iter
ia
o
n
titl
e
an
d
a
b
s
tr
ac
t
r
ev
iew
wer
e
ex
clu
d
ed
,
an
d
1
4
6
p
ap
e
r
s
p
r
o
ce
ed
ed
to
f
u
ll
-
tex
t
ev
alu
atio
n
.
At
t
h
is
s
tag
e
a
f
u
r
th
er
1
0
4
p
ap
er
s
wer
e
ex
clu
d
ed
f
o
r
r
ea
s
o
n
s
r
ec
o
r
d
e
d
b
el
o
w
leav
in
g
4
2
p
a
p
er
s
r
etain
ed
f
o
r
s
y
n
th
esis
.
Of
th
ese,
2
0
a
r
e
m
e
n
tio
n
e
d
as
c
o
r
e
liter
atu
r
e
in
th
e
co
m
p
ar
ativ
e
m
at
r
ix
in
s
ec
tio
n
4
.
2
.
5
.
Repro
du
cibili
t
y
s
t
a
t
em
e
nt
a
nd
s
ea
rc
h str
ing
s
I
n
t
h
e
in
ter
est
o
f
r
e
p
r
o
d
u
cib
ili
ty
,
th
e
ex
ac
t
B
o
o
lean
s
ea
r
ch
q
u
er
ies
u
s
ed
ac
r
o
s
s
th
e
f
i
v
e
d
at
ab
ases
ar
e
d
o
cu
m
e
n
ted
b
elo
w.
All
s
ea
r
ch
es
wer
e
co
n
d
u
cted
b
etwe
en
J
a
n
u
ar
y
a
n
d
Ma
r
c
h
2
0
2
6
.
T
h
e
b
ase
q
u
er
y
tem
p
late
ap
p
lied
to
all
d
atab
ases
wa
s
:
Q
u
er
y
1
(
R
NN/L
STM
f
o
c
u
s
)
:
(
m
alar
ia
OR
“
v
ec
to
r
-
b
o
r
n
e
d
is
ea
s
e
”
)
AND
(
p
r
ed
ictio
n
OR
f
o
r
ec
asti
n
g
O
R
“
ea
r
ly
war
n
in
g
”
)
AND
(
“
n
eu
r
al
n
etwo
r
k
”
OR
L
STM
OR
GR
U
O
R
“
d
ee
p
lear
n
in
g
”
OR
“
r
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
”
)
AND
(
Af
r
ica
O
R
“
Su
b
-
Sah
ar
an
”
OR
tr
o
p
ical
OR
“
lo
w
-
in
co
m
e
”
)
.
Qu
er
y
2
(
a
tten
tio
n
m
e
ch
a
n
is
m
s
)
:
(
m
alar
ia
OR
d
en
g
u
e
OR
“
i
n
f
ec
tio
u
s
d
is
ea
s
e
”
)
AND
(
“
att
en
tio
n
m
ec
h
an
is
m
”
OR
“
atten
tio
n
-
b
ased
”
OR
“
s
elf
-
atten
tio
n
”
OR
tr
an
s
f
o
r
m
er
)
AND
(
f
o
r
ec
asti
n
g
O
R
p
r
ed
ictio
n
OR
s
u
r
v
eillan
ce
)
.
Qu
er
y
3
(
h
ea
lth
s
y
s
tem
in
teg
r
atio
n
)
:
(
m
alar
ia
OR
“
d
i
s
ea
s
e
s
u
r
v
eillan
ce
”
)
AND
(
DHI
S2
O
R
“
h
ea
lth
in
f
o
r
m
atio
n
s
y
s
tem
”
OR
“
h
ea
lth
f
ac
ilit
y
”
OR
“
r
o
u
t
in
e
s
u
r
v
eillan
ce
”
)
AND
(
“
m
a
ch
in
e
lear
n
in
g
”
OR
“
d
ee
p
lear
n
in
g
”
OR
p
r
ed
ictio
n
)
.
Qu
er
y
4
(
c
lim
ate
-
d
is
ea
s
e
ec
o
lo
g
y
)
:
(
m
alar
ia
OR
Plas
m
o
d
iu
m
)
AND
(
r
ain
f
all
OR
tem
p
er
atu
r
e
OR
c
l
im
ate
OR
ND
VI
OR
“
lan
d
s
u
r
f
ac
e
tem
p
er
atu
r
e
”
)
AND
(
p
r
ed
icti
o
n
OR
tr
an
s
m
is
s
io
n
OR
in
cid
en
ce
)
AND
Af
r
ica
.
Data
b
ase
-
s
p
ec
i
f
i
c
f
i
lte
r
s
a
p
p
lie
d
:
p
u
b
li
ca
t
i
o
n
d
at
e
r
a
n
g
e
1
9
9
7
–
2
0
2
5
;
E
n
g
lis
h
la
n
g
u
a
g
e;
p
ee
r
-
r
e
v
iew
e
d
jo
u
r
n
als
a
n
d
c
o
n
f
e
r
e
n
ce
p
r
o
ce
e
d
i
n
g
s
o
n
l
y
.
Gr
e
y
lit
e
r
at
u
r
e,
p
r
ep
r
i
n
ts
w
it
h
o
u
t
a
p
u
b
lis
h
e
d
p
ee
r
-
r
ev
iew
e
d
co
u
n
te
r
p
a
r
t
,
a
n
d
ed
it
o
r
ials
w
er
e
ex
cl
u
d
e
d
at
t
h
e
d
a
ta
b
as
e
f
ilt
er
s
t
a
g
e
b
e
f
o
r
e
t
itl
e
a
n
d
a
b
s
t
r
a
ct
s
c
r
e
en
in
g
.
T
h
e
PR
I
S
MA
f
l
o
w
d
i
ag
r
a
m
in
F
ig
u
r
e
1
,
d
o
c
u
m
en
ts
t
h
e
f
u
ll
s
cr
ee
n
i
n
g
p
r
o
c
ess
.
R
es
ea
r
c
h
e
r
s
wis
h
i
n
g
t
o
r
e
p
li
ca
t
e
t
h
is
r
e
v
ie
w
a
r
e
a
d
v
is
ed
t
o
r
e
-
r
u
n
t
h
e
q
u
e
r
ies
ab
o
v
e
,
as
n
ew
p
u
b
lic
ati
o
n
s
e
n
te
r
i
n
g
t
h
e
d
a
ta
b
as
es
a
f
te
r
Ma
r
c
h
2
0
2
6
m
ay
y
i
el
d
ad
d
i
ti
o
n
al
r
el
e
v
a
n
t
s
t
u
d
ies
.
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
I
SS
N:
2722
-
3
2
2
1
A
liter
a
tu
r
e
r
ev
ie
w
o
f
r
ec
u
r
r
e
n
t n
eu
r
a
l n
etw
o
r
k
a
p
p
r
o
a
ch
es
to
ma
la
r
ia
o
u
tb
r
ea
k
…
(
S
o
p
h
i
a
Temb
u
r
e
)
381
Fig
u
r
e
1
.
PR
I
SMA
f
lo
w
d
iag
r
am
o
f
liter
atu
r
e
s
cr
ee
n
in
g
p
r
o
c
ess
3.
T
H
E
M
AT
I
C
L
I
T
E
RA
T
UR
E
RE
V
I
EW
T
h
e
r
etain
ed
liter
atu
r
e
f
alls
in
to
s
ix
b
r
o
ad
th
em
atic
s
tr
an
d
s
.
E
ac
h
s
tr
an
d
is
r
ev
iewe
d
in
tu
r
n
,
with
atten
tio
n
g
iv
en
to
ch
r
o
n
o
l
o
g
ic
al
d
ev
elo
p
m
e
n
t,
m
eth
o
d
o
lo
g
ic
al
ev
o
lu
tio
n
,
an
d
r
ec
u
r
r
i
n
g
d
e
b
ates.
3
.
1
.
Clima
t
e
-
dis
ea
s
e
ec
o
lo
g
y
a
nd
v
ec
t
o
r
bio
lo
g
y
R
esear
ch
co
n
ce
r
n
i
n
g
th
e
r
elatio
n
s
h
ip
b
etwe
en
clim
atic
f
ac
to
r
s
an
d
th
e
tr
an
s
m
is
s
io
n
o
f
m
alar
i
a
b
io
lo
g
ically
b
eg
an
i
n
th
e
late
n
in
etee
n
th
ce
n
tu
r
y
.
Ho
wev
er
,
f
r
o
m
a
m
o
d
e
r
n
-
d
ay
p
er
s
p
ec
tiv
e,
th
e
f
ir
s
t
attem
p
t
at
m
o
d
ellin
g
was
m
ad
e
b
y
C
r
aig
et
a
l
.
[
6
]
.
T
h
e
clim
at
e
s
u
itab
ilit
y
m
o
d
el
p
r
o
p
o
s
ed
b
y
C
r
aig
et
a
l
.
[
6
]
id
en
tifie
d
tem
p
er
at
u
r
e
an
d
p
r
ec
ip
itatio
n
as
th
e
two
m
ajo
r
clim
atic
f
ac
to
r
s
th
at
d
eter
m
i
n
e
th
e
in
ten
s
ity
o
f
tr
an
s
m
is
s
io
n
o
f
f
alcip
ar
u
m
m
a
lar
ia
in
Su
b
-
Sah
ar
an
Af
r
ica
b
a
s
ed
o
n
m
o
n
th
l
y
en
v
elo
p
es
o
f
clim
atic
co
n
d
itio
n
s
wh
er
e
th
e
An
o
p
h
eles
v
ec
to
r
wo
u
ld
co
m
p
lete
its
life
cy
cle
an
d
wh
er
e
th
e
Plas
m
o
d
iu
m
f
alcip
ar
u
m
p
a
r
asit
e
wo
u
ld
m
atu
r
e
in
th
e
An
o
p
h
ele
s
m
o
s
q
u
ito
.
Patz
et
a
l
.
[
7
]
m
eth
o
d
o
lo
g
y
was
ex
ten
d
ed
to
en
c
o
m
p
ass
tim
e
-
d
ep
en
d
en
cy
.
T
h
e
m
o
d
ellin
g
r
esu
lts
d
em
o
n
s
tr
ated
a
c
o
n
s
is
ten
t
lag
in
th
e
r
ain
f
all
p
ea
k
-
to
-
m
alar
i
a
ca
s
e
in
cid
en
ce
p
e
r
io
d
o
f
f
o
u
r
to
eig
h
t
wee
k
s
.
T
h
is
f
in
d
in
g
is
p
r
ac
tically
r
elev
an
t
as
it
im
p
lies
th
at
o
n
e
-
m
o
n
th
-
ah
ea
d
f
o
r
ec
asts
co
u
ld
in
co
r
p
o
r
ate
o
b
s
er
v
ed
r
ain
f
all
d
ata.
I
n
co
n
t
r
ast,
th
r
e
e
-
m
o
n
th
-
ah
ea
d
f
o
r
ec
as
ts
wo
u
ld
h
av
e
to
d
ep
e
n
d
s
o
lely
o
n
th
e
s
ea
s
o
n
al
clim
ate
in
d
ices
an
d
au
to
c
o
r
r
elatio
n
ef
f
ec
ts
.
T
h
is
d
elay
o
cc
u
r
s
b
ec
au
s
e
r
ain
f
all
p
r
o
v
id
es
th
e
n
ec
ess
ar
y
b
r
ee
d
in
g
g
r
o
u
n
d
s
,
lar
v
ae
n
ee
d
a
b
o
u
t
tw
o
wee
k
s
to
d
e
v
elo
p
in
to
ad
u
lt
s
,
p
ar
asit
es
n
ee
d
10
to
12
d
ay
s
to
in
cu
b
ate
with
in
th
e
m
o
s
q
u
ito
,
an
d
in
f
ec
tio
n
o
f
h
u
m
an
s
an
d
th
e
d
ev
elo
p
m
en
t
o
f
s
y
m
p
to
m
s
ad
d
an
o
t
h
er
o
n
e
to
two
wee
k
s
.
T
h
is
d
elay
h
as
im
p
o
r
ta
n
t
r
am
if
icati
o
n
s
f
o
r
f
o
r
ec
asti
n
g
,
in
th
e
last
o
n
e
m
o
n
th
ah
ea
d
p
r
ed
ictio
n
s
will
u
tili
ze
r
ain
f
all
d
ata
f
r
o
m
th
e
p
r
ev
io
u
s
m
o
n
th
,
wh
er
ea
s
th
r
ee
m
o
n
th
s
ah
ea
d
p
r
ed
ictio
n
s
will
h
av
e
to
r
eso
r
t
to
m
o
r
e
g
en
e
r
al
s
ea
s
o
n
al
f
ac
to
r
s
an
d
au
t
o
co
r
r
e
latio
n
ef
f
ec
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
3
2
2
1
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
,
Vo
l.
7
,
No
.
3
,
No
v
em
b
er
20
26
:
3
7
7
-
393
382
R
ain
f
all
v
ar
iatio
n
is
th
e
m
o
s
t
im
p
o
r
tan
t
b
eliev
ed
e
x
ter
n
al
d
r
iv
er
o
f
m
alar
ia
tr
an
s
m
is
s
io
n
,
ac
co
r
d
in
g
to
Ma
b
aso
et
a
l
.
[
8
]
u
s
in
g
13
y
ea
r
s
o
f
n
atio
n
al
Z
im
b
a
b
wea
n
d
ata.
T
h
o
u
g
h
th
is
f
in
d
in
g
ca
n
b
e
a
p
p
lied
t
o
f
o
r
ec
asti
n
g
,
th
e
f
in
d
in
g
s
m
ay
n
o
t
b
e
ap
p
licab
le
f
o
r
th
e
m
o
d
er
n
-
d
ay
d
is
tr
ict
-
b
ased
p
r
ed
ic
tio
n
m
o
d
els
d
u
e
to
th
eir
co
ar
s
e
an
n
u
al
tim
e
s
ca
l
e
an
d
p
r
e
-
DHI
S2
d
ata
co
llec
tio
n
m
eth
o
d
o
lo
g
y
,
alth
o
u
g
h
d
r
u
g
r
esis
tan
ce
an
d
h
u
m
an
im
m
u
n
o
d
e
f
icien
cy
v
ir
u
s
(
HI
V)
co
-
in
f
ec
tio
n
also
r
e
s
u
lted
in
a
n
in
c
r
ea
s
e
in
ca
s
e
n
u
m
b
er
s
in
t
h
e
late
1
9
9
0
s
.
I
n
th
eir
s
tu
d
y
p
u
b
lis
h
ed
in
2
0
1
8
,
Gwitir
a
et
a
l
.
[
9
]
em
p
lo
y
ed
g
e
o
g
r
a
p
h
ic
in
f
o
r
m
atio
n
s
y
s
tem
(
GI
S)
ap
p
r
o
ac
h
es
f
o
r
m
ap
p
in
g
th
e
p
r
ev
alen
ce
o
f
m
alar
ia
b
ased
o
n
altitu
d
e,
lan
d
u
s
e/co
v
e
r
ty
p
es,
an
d
p
o
p
u
latio
n
d
en
s
ity
at
th
e
war
d
le
v
el,
f
o
c
u
s
in
g
o
n
s
p
atial
h
eter
o
g
en
eity
b
u
t n
o
t o
n
tem
p
o
r
al
p
r
e
d
ictio
n
.
C
am
in
ad
e
et
a
l
.
[
1
0
]
r
elied
o
n
d
y
n
am
ical
clim
ate
m
o
d
els
to
in
v
esti
g
ate
th
e
im
p
ac
t
o
f
f
u
tu
r
e
clim
ate
s
ce
n
ar
io
s
o
n
th
e
s
p
r
ea
d
o
f
m
alar
ia,
wh
er
ea
s
T
h
o
m
s
o
n
et
a
l
.
[
11
]
illu
s
tr
ated
th
at
clim
atic
in
d
icato
r
s
d
er
i
v
e
d
f
r
o
m
th
e
eq
u
ato
r
ial
Pacif
ic,
s
u
ch
as
E
l
Niñ
o
-
So
u
th
er
n
Oscill
atio
n
(
E
NSO)
,
co
u
ld
b
e
u
s
ed
to
f
o
r
ec
ast
ab
n
o
r
m
al
m
ala
r
ia
s
ea
s
o
n
s
i
n
n
eig
h
b
o
u
r
in
g
B
o
ts
wan
a
u
p
to
f
o
u
r
m
o
n
th
s
ah
ea
d
.
T
h
e
ab
o
v
e
wo
r
k
s
clea
r
ly
s
h
o
w
th
at
clim
atic
f
ac
to
r
s
ar
e
n
ec
ess
ar
y
b
u
t
n
o
t su
f
f
icie
n
t d
eter
m
in
an
ts
f
o
r
p
r
e
d
ictin
g
th
e
d
y
n
am
i
cs o
f
m
alar
ia.
3
.
2
.
St
a
t
is
t
ica
l a
pp
ro
a
ches t
o
m
a
l
a
ria
f
o
re
ca
s
t
ing
E
ar
l
y
m
alar
ia
ea
r
ly
-
war
n
i
n
g
s
y
s
tem
s
wer
e
b
ased
o
n
tr
ad
itio
n
al
tim
e
-
s
er
ies
an
aly
s
is
te
ch
n
iq
u
es.
Acc
o
r
d
in
g
to
Z
in
s
ze
r
et
a
l
.
[
12
]
,
th
e
m
ai
n
r
esu
lt
o
f
SAR
I
MA
m
o
d
ellin
g
is
its
av
ailab
ilit
y
to
lo
w
-
r
eso
u
r
ce
s
ettin
g
s
.
T
h
e
d
o
w
n
s
id
e
is
SAR
I
MA
is
s
p
ec
if
ic
as
to
th
e
ty
p
e
o
f
d
ata
it
ca
n
h
an
d
le.
I
t
as
s
u
m
es
lin
ea
r
ity
an
d
s
tatio
n
ar
ity
.
Du
e
to
th
e
c
o
m
b
in
ed
ef
f
ec
ts
o
f
clim
ate
v
ar
ian
ce
,
d
r
u
g
r
esis
tan
ce
an
d
in
c
r
e
asin
g
co
v
er
a
g
es
o
f
in
ter
v
en
tio
n
s
,
th
is
ass
u
m
p
tio
n
d
o
es
n
o
t
h
o
ld
a
n
y
m
o
r
e.
T
h
e
SAR
I
MA
m
o
d
el
co
n
s
id
er
s
th
e
s
ea
s
o
n
al
n
atu
r
e
o
f
m
alar
ia
in
f
ec
tio
n
s
a
n
d
th
e
a
u
t
o
co
r
r
elatio
n
o
f
ca
s
e
s
er
ies,
an
d
b
ec
a
u
s
e
o
f
its
in
ter
p
r
etab
ilit
y
,
it
is
th
e
p
r
e
f
er
r
ed
ap
p
r
o
ac
h
in
lo
w
-
r
eso
u
r
ce
s
ettin
g
s
wh
er
e
b
lack
b
o
x
m
o
d
els ar
e
s
u
b
ject
to
d
o
u
b
t
b
y
p
o
licy
m
ak
er
s
.
Gen
er
alize
d
lin
ea
r
m
o
d
els
(
G
L
Ms)
an
d
g
en
e
r
alize
d
ad
d
itiv
e
m
o
d
els
(
GAM
s
)
h
av
e
also
b
ee
n
u
s
ed
ex
ten
s
iv
ely
.
L
o
ev
in
s
o
h
n
[
13
]
was
am
o
n
g
th
e
f
ir
s
t
to
q
u
an
tif
y
th
e
r
elatio
n
s
h
ip
b
etwe
en
cli
m
ate
v
ar
iab
les
an
d
m
alar
ia
tr
an
s
m
is
s
io
n
,
u
s
in
g
a
GL
M
to
p
r
ed
ict
m
alar
ia
in
f
ec
tio
n
s
in
h
ig
h
lan
d
s
o
f
R
wan
d
a
u
s
in
g
tem
p
er
atu
r
e
an
o
m
alies.
B
ay
esian
g
eo
s
tati
s
tical
m
ap
p
in
g
f
o
r
esti
m
atin
g
co
n
tin
en
tal
Plas
m
o
d
iu
m
f
alcip
ar
u
m
p
r
e
v
alen
ce
was e
m
p
lo
y
ed
b
y
B
h
att
et
a
l
. [
14
]
b
u
t th
e
s
tu
d
y
em
p
h
asized
s
p
atial
m
ap
p
in
g
r
at
h
er
th
a
n
f
u
t
u
r
e
p
r
e
d
ictio
n
.
I
t
is
clea
r
th
at
c
lass
ical
s
tati
s
tics
h
av
e
b
ec
o
m
e
in
cr
ea
s
in
g
ly
d
if
f
icu
lt
to
u
s
e
ef
f
ec
tiv
e
ly
.
L
in
ea
r
d
ep
en
d
e
n
ce
an
d
s
tatio
n
ar
y
p
r
o
ce
s
s
es
ar
e
ass
u
m
ed
b
y
SA
R
I
MA
an
d
GL
M
m
eth
o
d
o
lo
g
ies,
two
co
n
d
itio
n
s
th
at
ar
e
n
o
t
alwa
y
s
s
ati
s
f
ied
b
y
th
e
m
alar
ia
tim
e
s
er
ies.
Dr
u
g
r
esis
tan
ce
,
in
ter
v
en
tio
n
co
v
er
ag
e,
p
o
p
u
latio
n
m
ig
r
atio
n
,
a
n
d
clim
atic
ch
a
n
g
e
r
ep
r
esen
t
s
o
m
e
o
f
th
e
n
o
n
-
s
tatio
n
ar
ities
in
m
alar
ia
tim
e
s
er
ies
d
ata.
ML
g
ain
ed
p
o
p
u
la
r
ity
in
th
e
e
ar
ly
2
0
1
0
s
.
3
.
3
.
ML
a
pp
ro
a
ches
T
r
ad
itio
n
al
a
p
p
licatio
n
s
o
f
ML
in
m
ala
r
ia
p
r
e
d
ictio
n
wer
e
f
o
cu
s
ed
o
n
th
e
u
s
e
o
f
en
s
e
m
b
les
lik
e
r
an
d
o
m
f
o
r
ests
an
d
b
o
o
s
ted
d
e
cisi
o
n
tr
ee
s
.
I
n
th
eir
g
eo
s
tatis
ti
ca
l m
o
d
el,
B
h
att
et
a
l
. [
14
]
in
c
o
r
p
o
r
ate
d
b
o
o
s
ted
r
eg
r
ess
io
n
tr
ee
s
,
s
h
o
win
g
th
a
t
en
s
em
b
les
ca
n
d
etec
t
n
o
n
li
n
ea
r
r
elatio
n
s
h
ip
s
b
etwe
en
cl
im
atic
f
ac
to
r
s
an
d
d
is
ea
s
e
p
r
ev
alen
ce
.
Su
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
a
n
d
n
eu
r
al
n
etwo
r
k
s
as
class
if
ier
s
h
av
e
also
b
ee
n
in
v
esti
g
ated
in
m
alar
ia
r
esear
ch
.
I
n
p
a
r
ticu
lar
,
Kian
g
et
a
l
.
[
15
]
im
p
lem
e
n
ted
an
o
u
tb
r
ea
k
v
e
r
s
u
s
non
-
o
u
tb
r
ea
k
SVM
in
T
h
ailan
d
with
h
ig
h
s
p
ec
if
icity
b
u
t lo
w
s
en
s
iti
v
ity
an
im
p
o
r
tan
t im
b
alan
ce
as th
e
co
s
t o
f
er
r
o
n
e
o
u
s
ly
m
is
s
in
g
an
o
u
tb
r
e
ak
(
a
f
alse
n
eg
ativ
e)
is
f
ar
g
r
e
ater
th
an
th
at
o
f
a
f
alse
alar
m
.
T
h
ey
s
u
g
g
est
th
at
m
o
d
els
f
r
am
ed
in
ter
m
s
o
f
cl
ass
if
icatio
n
ar
e
in
h
er
en
tly
m
is
alig
n
ed
with
th
e
r
eq
u
ir
em
en
ts
o
f
m
alar
ia
e
ar
ly
war
n
in
g
;
class
if
icatio
n
p
r
o
v
e
d
in
f
er
io
r
to
r
eg
r
ess
io
n
b
ec
au
s
e
it
lo
s
es
in
f
o
r
m
atio
n
o
n
ca
s
e
n
u
m
b
er
s
.
Ho
we
v
er
,
d
esp
ite
all
th
ese
ad
v
an
tag
es,
th
e
en
s
em
b
le
an
d
SVM
m
o
d
e
ls
s
u
f
f
er
f
r
o
m
th
e
s
am
e
d
r
aw
b
ac
k
,
f
o
r
ex
a
m
p
le
,
th
e
d
ata
is
co
n
s
id
er
ed
as
a
v
e
cto
r
,
n
o
t
as
a
s
eq
u
en
ce
.
Fo
r
in
s
tan
ce
,
f
o
r
a
twelv
e
-
m
o
n
th
p
er
io
d
,
if
th
e
r
e
ar
e
twen
ty
f
ea
tu
r
es
p
er
m
o
n
th
,
th
e
in
p
u
t
will
b
e
a
2
4
0
-
v
ec
to
r
in
wh
ich
th
er
e
is
n
o
s
eq
u
e
n
ce
o
r
d
er
b
etwe
e
n
d
if
f
er
en
t
o
b
s
er
v
atio
n
s
.
3
.
4
.
RNNs
On
e
s
u
ch
im
p
o
r
ta
n
t
m
iles
to
n
e
was
th
e
L
STM
n
etwo
r
k
d
es
ig
n
o
f
Ho
c
h
r
eiter
an
d
Sch
m
id
h
u
b
er
[
3
]
.
B
y
in
tr
o
d
u
cin
g
a
n
ar
ch
itectu
r
e
in
co
r
p
o
r
atin
g
in
p
u
t,
f
o
r
g
et,
an
d
o
u
tp
u
t
g
ates
to
co
n
tr
o
l
th
e
in
f
o
r
m
atio
n
f
l
o
w
ac
r
o
s
s
d
if
f
er
en
t
tim
e
s
tep
s
,
th
e
L
STM
ad
d
r
ess
ed
th
e
v
an
is
h
in
g
g
r
ad
ien
t
is
s
u
e
o
f
p
r
ev
io
u
s
r
ec
u
r
r
en
t
n
etw
o
r
k
s
.
T
h
is
allo
wed
tr
ain
in
g
o
n
s
eq
u
en
ce
s
o
f
te
n
s
to
h
u
n
d
r
ed
s
o
f
tim
estep
s
,
wh
ich
is
th
e
v
er
y
r
an
g
e
o
f
s
eq
u
en
ce
len
g
th
n
ee
d
ed
f
o
r
m
o
n
th
ly
e
p
i
d
em
io
lo
g
y
f
o
r
ec
asti
n
g
.
T
h
e
GR
U
was
p
r
esen
ted
b
y
C
h
o
et
a
l
.
[
16
]
a
s
an
ap
p
r
o
ac
h
th
at
is
s
o
m
ewh
at
co
m
p
u
tatio
n
ally
m
o
r
e
ef
f
icien
t th
an
th
e
L
STM
.
I
n
th
e
GR
U
m
o
d
el
th
e
f
o
r
g
et
an
d
in
p
u
t g
ates a
r
e
co
m
b
in
ed
in
t
o
a
s
in
g
le
u
p
d
ate
g
ate,
an
d
th
e
ce
ll
s
tate
i
s
in
co
r
p
o
r
at
ed
in
to
th
e
h
id
d
en
s
tate,
h
alv
i
n
g
th
e
n
u
m
b
er
o
f
p
ar
a
m
eter
s
u
s
ed
in
th
e
L
STM
.
W
h
en
co
m
p
a
r
in
g
th
e
L
STM
a
n
d
GR
U,
C
h
u
n
g
et
a
l
.
[
17
]
o
b
s
er
v
ed
n
o
clea
r
s
u
p
er
i
o
r
a
r
ch
it
ec
tu
r
e
in
m
o
d
ellin
g
m
u
s
ic
an
d
s
p
ee
c
h
.
On
th
e
o
th
er
h
an
d
,
in
Gr
ef
f
et
a
l
.
[
18
]
s
t
u
d
y
c
o
m
p
ar
i
n
g
t
h
e
im
p
ac
t
o
f
d
if
f
er
en
t
p
ar
ts
o
f
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
I
SS
N:
2722
-
3
2
2
1
A
liter
a
tu
r
e
r
ev
ie
w
o
f
r
ec
u
r
r
e
n
t n
eu
r
a
l n
etw
o
r
k
a
p
p
r
o
a
ch
es
to
ma
la
r
ia
o
u
tb
r
ea
k
…
(
S
o
p
h
i
a
Temb
u
r
e
)
383
L
STM
ar
ch
itectu
r
e,
th
e
f
o
r
g
et
g
ate
was
id
e
n
tifie
d
as
t
h
e
m
o
s
t
cr
u
cial
co
m
p
o
n
e
n
t,
f
o
llo
w
ed
b
y
in
p
u
t/o
u
tp
u
t
g
ates.
Use
o
f
R
NNs
f
o
r
d
is
ea
s
e
f
o
r
ec
asti
n
g
g
ain
e
d
m
o
m
en
tu
m
in
th
e
late
2
0
1
0
s
.
T
h
is
g
ain
is
d
u
e
to
L
STM
’
s
ab
ilit
y
to
ca
p
tu
r
e
n
o
n
lin
ea
r
r
elatio
n
s
h
ip
s
am
o
n
g
clim
ate
v
ar
iab
les
a
n
d
ca
s
e
co
u
n
ts
o
v
er
tim
e
f
r
am
es
o
f
m
u
ltip
le
wee
k
s
.
T
h
e
m
o
d
el
r
e
lied
s
o
lely
o
n
clim
ate
in
p
u
ts
;
it
h
ad
n
o
atten
tio
n
m
ec
h
an
is
m
an
d
in
c
o
r
p
o
r
ated
n
o
h
ea
lth
f
ac
ilit
y
co
v
ar
iates,
s
o
its
ce
ilin
g
p
er
f
o
r
m
a
n
ce
is
b
e
lo
w
wh
at
a
m
u
lti
-
s
o
u
r
ce
a
r
ch
it
ec
tu
r
e
co
u
ld
r
ea
ch
.
Acc
o
r
d
in
g
t
o
Mb
u
n
g
e
et
a
l
.
[
19
]
th
er
e
ar
e
two
m
ain
lim
it
at
io
n
s
ass
o
ciate
d
with
th
e
u
s
e
o
f
AI
tech
n
iq
u
es
in
co
n
tr
o
llin
g
m
alar
ia
in
Su
b
-
Sah
ar
an
Af
r
ica:
lack
o
f
u
s
e
o
f
at
ten
tio
n
-
b
ased
ap
p
r
o
ac
h
es
an
d
f
ailu
r
e
to
in
teg
r
ate
m
u
ltip
le
s
o
u
r
ce
s
o
f
d
ata
s
u
ch
as c
lim
ate
an
d
h
ea
lth
s
y
s
tem
d
ata.
3
.
4
.
1
.
M
a
t
hem
a
t
ic
a
l f
o
r
m
ula
t
io
n o
f
L
ST
M
a
nd
G
RU
a
rc
hite
ct
ures
T
h
e
f
o
u
n
d
atio
n
al
m
o
tiv
atio
n
f
o
r
L
STM
n
etwo
r
k
s
lies
in
t
h
eir
s
o
lu
tio
n
to
th
e
v
a
n
is
h
in
g
g
r
ad
ien
t
p
r
o
b
lem
th
at
p
r
ev
en
ts
s
tan
d
ar
d
R
NNs
f
r
o
m
lear
n
in
g
lo
n
g
-
r
an
g
e
tem
p
o
r
al
d
ep
en
d
en
cies.
I
n
a
s
tan
d
ar
d
R
NN,
th
e
h
id
d
e
n
s
tate
ℎ
at
tim
e
s
tep
is
co
m
p
u
ted
as:
ℎ
=
ℎ
(
ℎ
ℎ
−
1
+
+
)
w
h
er
e
ℎ
,
ar
e
weig
h
t
m
atr
ices,
is
th
e
i
n
p
u
t
at
tim
e,
an
d
is
a
b
i
as
v
ec
to
r
.
Du
r
in
g
b
ac
k
p
r
o
p
a
g
atio
n
th
r
o
u
g
h
tim
e
(
B
PTT
)
,
g
r
a
d
ien
ts
ar
e
m
u
ltip
lied
r
e
p
ea
ted
ly
b
y
ℎ
,
ca
u
s
in
g
th
em
to
v
an
is
h
ex
p
o
n
en
tially
o
v
er
lo
n
g
s
eq
u
en
ce
s
a
cr
itical
lim
itatio
n
wh
en
m
o
d
ellin
g
m
alar
ia
tr
an
s
m
is
s
io
n
,
wh
er
e
clim
ate
-
ca
s
e
lag
s
s
p
an
f
o
u
r
to
eig
h
t w
ee
k
s
.
Ho
ch
r
eiter
an
d
Sch
m
id
h
u
b
er
[
3
]
ad
d
r
ess
ed
th
is
b
y
in
t
r
o
d
u
cin
g
t
h
r
ee
g
atin
g
m
ec
h
a
n
is
m
s
th
at
r
eg
u
late
in
f
o
r
m
atio
n
f
lo
w
th
r
o
u
g
h
a
ce
ll
s
tate
.
T
h
e
f
o
r
g
et
g
ate
d
ec
id
es
wh
at
p
r
o
p
o
r
tio
n
o
f
th
e
p
r
ev
io
u
s
ce
ll st
ate
to
d
is
ca
r
d
:
=
(
⋅
[
ℎ
−
1
,
]
+
)
T
h
e
in
p
u
t
g
ate
an
d
ca
n
d
i
d
ate
ce
ll st
ate
̃
d
eter
m
in
e
wh
at
n
ew
in
f
o
r
m
atio
n
to
s
to
r
e:
̃
=
(
⋅
[
ℎ
−
1
,
]
+
)
T
h
e
ce
ll st
ate
is
th
en
u
p
d
ated
as:
=
⊙
−
1
+
⊙
̃
w
h
er
e
⊙
d
en
o
tes
elem
e
n
t
-
wis
e
(
Had
am
ar
d
)
m
u
ltip
licatio
n
.
Fin
ally
,
th
e
o
u
tp
u
t
g
ate
f
ilter
s
th
e
ce
ll
s
tate
to
p
r
o
d
u
ce
th
e
h
i
d
d
en
s
tate
p
ass
ed
to
th
e
n
ex
t tim
e
s
tep
:
=
(
⋅
[
ℎ
−
1
]
+
)
ℎ
=
⊙
ta
n
h
(
)
T
h
e
GR
U,
in
tr
o
d
u
ce
d
b
y
C
h
o
et
a
l
.
[
16
]
s
im
p
lifie
s
th
e
L
S
T
M
b
y
m
er
g
in
g
th
e
f
o
r
g
et
an
d
in
p
u
t
g
ates
in
to
a
s
in
g
le
u
p
d
ate
g
ate
an
d
a
d
d
in
g
a
r
eset g
ate
.
=
(
⋅
[
ℎ
−
1
,
]
)
=
(
⋅
[
ℎ
−
1
,
]
)
ℎ
̃
=
ta
n
h
(
⋅
[
⊙
ℎ
−
1
,
]
)
ℎ
=
(
1
−
)
⊙
ℎ
−
1
+
⊙
ℎ
̃
T
h
e
GR
U
is
co
m
p
u
tatio
n
ally
lig
h
ter
th
an
t
h
e
L
STM
an
d
h
as
b
ee
n
s
h
o
wn
t
o
ac
h
iev
e
c
o
m
p
ar
ab
le
p
er
f
o
r
m
an
ce
o
n
s
h
o
r
ter
,
n
o
is
ier
s
eq
u
en
ce
s
a
p
r
o
p
er
ty
r
el
ev
an
t
to
Af
r
ican
d
is
tr
ict
-
lev
el
s
u
r
v
eillan
ce
d
ata
wh
er
e
wee
k
ly
ca
s
e
r
ec
o
r
d
s
ar
e
o
f
ten
in
co
m
p
lete.
Fo
r
tr
ain
i
n
g
,
b
o
th
a
r
ch
itectu
r
es
ar
e
o
p
ti
m
is
ed
u
s
in
g
B
PTT
with
Ad
am
o
r
R
MSp
r
o
p
o
p
tim
is
er
s
,
m
in
im
is
in
g
th
e
m
ea
n
s
q
u
ar
ed
er
r
o
r
(
MSE
)
o
r
MA
E
b
etwe
en
p
r
ed
icted
an
d
ac
tu
al
ca
s
e
c
o
u
n
ts
.
As
illu
s
tr
ated
in
Fig
u
r
e
2
,
t
h
e
p
r
o
g
r
ess
io
n
f
r
o
m
co
n
v
en
tio
n
al
R
NNs
to
L
STM
an
d
atten
tio
n
-
b
ased
ar
ch
itectu
r
es
r
ef
lects
s
u
cc
ess
iv
e
im
p
r
o
v
em
en
ts
in
m
o
d
ellin
g
lo
n
g
-
te
r
m
d
ep
en
d
e
n
cies
an
d
s
elec
tiv
ely
weig
h
tin
g
r
elev
an
t
h
is
to
r
ical
in
f
o
r
m
atio
n
,
th
er
e
b
y
en
h
an
cin
g
s
eq
u
en
ce
p
r
e
d
ictio
n
p
er
f
o
r
m
an
ce
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
3
2
2
1
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
,
Vo
l.
7
,
No
.
3
,
No
v
em
b
er
20
26
:
3
7
7
-
393
384
Fig
u
r
e
2
.
R
NN
v
s
L
STM
v
s
atten
tio
n
s
ch
em
atic
3
.
5
.
At
t
ent
io
n m
ec
ha
nis
m
s
in s
eque
ntia
l m
o
dels
Atten
tio
n
was
o
r
ig
in
ally
d
ev
e
lo
p
ed
f
o
r
m
ac
h
in
e
tr
an
s
latio
n
ap
p
licatio
n
s
.
B
ah
d
an
au
et
a
l
.
[
20
]
ca
m
e
u
p
with
s
o
f
t
alig
n
m
en
t
atten
tio
n
in
o
r
d
er
to
en
a
b
le
a
s
eq
u
en
ce
-
to
-
s
eq
u
e
n
ce
d
ec
o
d
e
r
to
s
elec
t
wh
ich
p
ar
ts
o
f
th
e
s
o
u
r
ce
s
en
ten
ce
ar
e
r
elev
an
t
d
u
r
in
g
ea
ch
s
tep
o
f
d
ec
o
d
in
g
.
T
h
is
is
ac
h
iev
ed
th
r
o
u
g
h
ca
lcu
latin
g
th
e
weig
h
ted
s
u
m
o
f
th
e
en
c
o
d
er
s
tates,
with
th
e
weig
h
ts
b
ein
g
c
alcu
lated
u
s
in
g
an
ali
g
n
m
en
t
f
u
n
ctio
n
.
Ap
p
licatio
n
o
f
atten
tio
n
m
e
ch
an
is
m
s
in
tim
e
-
s
er
ies
p
r
e
d
ictio
n
h
as
y
ield
ed
p
r
o
m
is
in
g
r
esu
lts
.
Fo
r
ex
am
p
le,
Qin
et
a
l
.
[
21
]
p
r
esen
ted
d
u
al
-
s
tag
e
atten
tio
n
-
b
ased
R
NNs
th
at
u
tili
ze
atte
n
tio
n
b
o
th
f
o
r
in
p
u
t
f
ea
tu
r
es
an
d
tim
e
s
tep
s
.
Su
c
h
m
o
d
els
h
av
e
b
ee
n
s
h
o
w
n
to
i
m
p
r
o
v
e
u
p
o
n
L
STM
m
o
d
els
wh
en
wo
r
k
in
g
with
s
to
ck
p
r
ice
an
d
air
q
u
ality
d
ata.
A
p
r
ac
tical
is
s
u
e
in
ap
p
ly
in
g
atten
tio
n
m
ec
h
an
is
m
s
t
o
Af
r
ican
m
alar
ia
s
u
r
v
eillan
ce
d
ata
is
th
at
it
i
s
s
m
all
an
d
s
p
ar
s
e.
T
h
e
n
atio
n
al
lev
el
DHI
S2
r
ec
o
r
d
s
f
o
r
a
s
in
g
le
co
u
n
tr
y
wo
u
ld
af
ter
m
is
s
in
g
v
alu
e
im
p
u
tatio
n
y
ield
o
n
ly
a
f
ew
h
u
n
d
r
ed
d
i
s
tr
ict
-
m
o
n
th
s
wo
r
th
o
f
u
s
ab
le
wee
k
ly
d
ata.
T
h
is
is
co
n
s
id
er
ab
ly
f
ewe
r
tr
ain
in
g
ex
am
p
les th
an
ar
e
ty
p
ically
u
s
ed
in
th
e
NL
P o
r
f
in
an
cial
ap
p
licatio
n
s
o
f
atten
tio
n
.
Ma
n
y
ap
p
r
o
ac
h
es
h
av
e
b
ee
n
tr
ied
in
s
im
ilar
d
is
ea
s
e
p
r
e
d
ictio
n
s
itu
atio
n
s
.
T
r
a
n
s
f
er
l
ea
r
n
in
g
is
a
tech
n
iq
u
e
f
o
r
p
r
e
-
tr
ain
i
n
g
an
atten
tio
n
-
L
STM
o
n
a
d
ata
-
r
ich
co
u
n
tr
y
lik
e
Ken
y
a
o
r
T
an
za
n
ia,
an
d
f
in
e
-
tu
n
in
g
it
o
n
a
c
o
m
p
a
r
ativ
ely
s
m
aller
Z
im
b
a
b
wea
n
d
ataset.
T
h
is
was
s
h
o
wn
to
s
ig
n
i
f
ican
tly
r
ed
u
ce
th
e
m
in
im
u
m
tr
ai
n
in
g
d
ata
r
eq
u
ir
em
en
t
in
s
im
ilar
p
u
b
lic
h
ea
lth
ap
p
licatio
n
s
.
I
n
ad
d
itio
n
,
au
g
m
en
tatio
n
tech
n
iq
u
es,
win
d
o
w
s
licin
g
,
m
ag
n
itu
d
e
war
p
in
g
,
tim
e
-
s
er
i
es
b
o
o
ts
tr
ap
p
in
g
,
e
x
p
lo
it
d
at
asets
to
ar
tific
ially
ex
p
an
d
s
m
all
tr
ain
in
g
s
ets
wi
th
a
r
ea
lis
tic
f
o
o
tp
r
in
t.
An
o
th
er
ap
p
r
o
ac
h
is
u
tili
zin
g
s
p
ar
s
e
atten
tio
n
v
ar
ian
ts
lik
e
L
o
n
g
f
o
r
m
er
,
I
n
f
o
r
m
er
.
T
h
ese
ar
ch
itectu
r
al
m
o
d
els
ar
e
s
p
ec
if
ically
d
esig
n
ed
to
lim
it
th
e
s
tan
d
ar
d
atten
tio
n
’
s
q
u
ad
r
atic
co
m
p
u
tatio
n
al
co
s
t
o
n
s
h
o
r
t
s
eq
u
en
ce
s
.
T
h
is
m
ak
es
th
em
s
u
itab
le
f
o
r
s
m
all
-
N,
lo
n
g
-
lag
p
r
ed
ictio
n
p
r
o
b
lem
s
f
o
r
d
is
tr
ict
-
lev
el
m
alar
ia
f
o
r
ec
asti
n
g
.
An
y
f
u
tu
r
e
s
tu
d
y
ap
p
ly
i
n
g
a
tten
tio
n
-
L
STM
to
So
u
th
e
r
n
Af
r
ican
m
alar
ia
d
ata
s
h
o
u
ld
ex
p
licitly
ev
alu
ate
th
ese
ad
a
p
tatio
n
s
tr
a
teg
ies.
Als
o
,
Vaswan
i
et
a
l
.
[
22
]
,
d
esp
ite
d
e
v
elo
p
in
g
th
e
tr
an
s
f
o
r
m
er
m
o
d
el
r
ath
er
th
an
R
NN
s
,
s
h
o
wed
th
at
atten
tio
n
alo
n
e
m
ay
b
e
s
u
p
er
io
r
to
th
e
latter
o
n
s
p
ec
if
ic
d
atasets
.
T
h
e
tr
an
s
f
o
r
m
er
b
o
o
m
h
as
n
o
t
f
o
u
n
d
wid
esp
r
ea
d
u
s
e
in
d
is
ea
s
e
f
o
r
ec
asti
n
g
s
in
ce
th
e
s
ize
o
f
d
atasets
u
s
ed
in
s
u
ch
f
o
r
ec
asti
n
g
r
em
ain
s
s
m
all
co
m
p
ar
ed
to
th
e
m
illi
o
n
s
r
eq
u
ir
ed
f
o
r
tr
ain
in
g
t
r
an
s
f
o
r
m
er
s
.
I
n
th
e
co
n
tex
t
o
f
d
is
ea
s
e
f
o
r
ec
asti
n
g
,
th
e
s
tu
d
y
co
n
d
u
cted
b
y
Z
h
u
et
a
l
.
[
23
]
u
s
ed
tem
p
o
r
al
atten
tio
n
in
p
r
ed
ictin
g
ca
s
es
o
f
d
en
g
u
e
f
ev
er
in
C
h
in
a.
T
h
e
tr
a
in
ed
atten
tio
n
weig
h
ts
co
r
r
esp
o
n
d
ed
p
r
ec
is
ely
with
th
e
d
en
g
u
e
v
ir
u
s
in
cu
b
atio
n
tim
e
win
d
o
w,
wh
ic
h
s
h
o
wed
th
at
th
e
m
o
d
el
h
a
d
in
d
e
p
en
d
e
n
tly
lear
n
ed
th
is
f
ea
tu
r
e
with
o
u
t
b
ein
g
ex
p
licitly
p
r
o
g
r
a
m
m
ed
to
d
o
s
o
.
Su
ch
a
f
in
d
in
g
wo
u
ld
b
e
v
er
y
r
elev
a
n
t
to
m
alar
ia
f
o
r
ec
asti
n
g
s
in
c
e
it
wo
u
ld
s
u
g
g
est
Evaluation Warning : The document was created with Spire.PDF for Python.
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
I
SS
N:
2722
-
3
2
2
1
A
liter
a
tu
r
e
r
ev
ie
w
o
f
r
ec
u
r
r
e
n
t n
eu
r
a
l n
etw
o
r
k
a
p
p
r
o
a
ch
es
to
ma
la
r
ia
o
u
tb
r
ea
k
…
(
S
o
p
h
i
a
Temb
u
r
e
)
385
th
at
an
atten
tio
n
-
L
STM
tr
ain
ed
o
n
Af
r
ican
m
alar
ia
d
ata
co
u
ld
b
e
ab
le
to
lear
n
to
g
i
v
e
weig
h
ts
to
t
h
e
f
o
u
r
-
to
-
eig
h
t
-
wee
k
clim
ate
-
ca
s
e
lag
id
en
tifie
d
b
y
Patz
et
a
l
.
[
7
]
th
er
e
b
y
ac
h
ie
v
in
g
im
p
r
o
v
ed
ac
cu
r
ac
y
an
d
a
d
eg
r
ee
o
f
e
p
id
em
io
lo
g
ical
i
n
ter
p
r
etab
ilit
y
t
h
at
v
a
n
illa
L
STM
m
o
d
els
lack
.
T
h
er
e
w
er
e
two
im
p
o
r
tan
t
im
p
licatio
n
s
o
f
th
is
o
b
s
er
v
at
io
n
.
Firstl
y
,
it
p
r
o
v
id
ed
a
p
r
o
o
f
th
at
atten
tio
n
m
o
d
els
ca
n
d
etec
t
b
io
lo
g
ical
s
ig
n
if
ican
ce
in
tim
e
s
tep
s
.
Se
co
n
d
ly
,
th
e
r
e
was
an
in
ter
p
r
e
tab
ilit
y
f
ac
to
r
in
atten
tio
n
m
o
d
els
th
at
d
o
es
n
o
t
ex
is
t in
p
u
r
e
R
NN
m
o
d
els.
Atten
tio
n
-
au
g
m
e
n
ted
L
STM
ar
ch
itectu
r
es
ca
n
b
e
ap
p
lie
d
to
o
th
er
e
p
id
em
ics
b
e
y
o
n
d
m
alar
ia.
Z
h
u
et
a
l
.
[
23
]
s
h
o
wed
th
at
a
n
atten
tio
n
-
L
STM
m
o
d
el
r
esu
lted
in
a
4
-
8
%
im
p
r
o
v
em
en
t
i
n
MA
E
f
o
r
d
en
g
u
e
p
r
ed
ictio
n
in
Ma
lay
s
ia
as
co
m
p
ar
ed
to
a
v
an
illa
L
STM
.
Mo
r
eo
v
er
,
th
e
atten
tio
n
weig
h
ts
o
f
f
er
ed
in
ter
p
r
etab
l
e
ev
id
en
ce
as to
wh
ich
h
is
to
r
ical
clim
ate
m
o
n
th
s
in
f
lu
en
ce
d
c
u
r
r
en
t tr
a
n
s
m
is
s
io
n
a
tr
an
s
p
ar
en
cy
p
r
o
p
er
ty
th
at
is
o
f
d
ir
ec
t
v
alu
e
to
p
u
b
lic
h
ea
lt
h
d
ec
is
io
n
-
m
a
k
er
s
.
I
n
a
C
OVI
D
-
1
9
f
o
r
ec
asti
n
g
s
tu
d
y
,
Dair
i
et
a
l
.
[
24
]
p
r
o
p
o
s
ed
a
s
tack
ed
atten
tio
n
-
L
STM
m
o
d
el
d
esig
n
ed
t
o
f
o
r
ec
ast
d
aily
ca
s
e
co
u
n
ts
f
o
r
te
n
co
u
n
tr
ies.
R
esu
lts
s
h
o
wed
th
is
m
o
d
el
ac
h
iev
e
d
a
m
ea
n
ab
s
o
l
u
te
p
er
ce
n
tag
e
e
r
r
o
r
(
MA
PE
)
o
f
3
.
2
%,
v
e
r
s
u
s
7
.
8
%
f
o
r
SA
R
I
MA
.
Fu
r
th
er
m
o
r
e,
th
e
au
th
o
r
s
d
em
o
n
s
tr
ate
th
e
atten
tio
n
m
ec
h
an
is
m
ap
p
r
o
p
r
ia
tely
r
ev
ea
led
th
e
7
-
14
-
d
ay
in
cu
b
atio
n
p
er
i
o
d
lag
win
d
o
w
as th
e
m
o
s
t p
r
e
d
ictiv
e
.
3
.
6
.
M
ulti
-
s
o
urce
da
t
a
f
us
io
n a
nd
hea
lt
h
-
s
y
s
t
em
info
rm
a
t
ics
T
h
e
v
ast
m
ajo
r
ity
o
f
ML
m
alar
ia
p
r
ed
ictio
n
s
tu
d
ies
th
at
u
s
e
an
R
NN
m
eth
o
d
o
lo
g
y
ex
clu
s
iv
ely
in
clu
d
e
clim
ate
in
f
o
r
m
atio
n
an
d
h
is
to
r
ical
ca
s
e
r
ec
o
r
d
s
wh
ile
ig
n
o
r
in
g
o
th
e
r
f
ac
to
r
s
th
at
in
f
lu
en
ce
th
e
co
n
n
ec
tio
n
b
etwe
en
d
is
ea
s
e
tr
an
s
m
is
s
io
n
an
d
in
cid
en
ce
,
s
u
ch
as
h
ea
lth
s
y
s
tem
ch
ar
ac
t
er
is
tics
.
T
h
er
e
ar
e
s
ev
er
al
ex
p
lan
atio
n
s
f
o
r
th
is
p
r
ac
tice,
o
n
e
o
f
wh
ich
is
p
u
r
ely
tech
n
ical,
as
h
ea
lth
f
ac
ilit
y
in
f
o
r
m
atio
n
ten
d
s
t
o
b
e
u
n
d
er
-
d
ig
itized
in
m
an
y
Af
r
ican
co
u
n
tr
ies
an
d
n
ee
d
s
t
o
b
e
g
ath
er
e
d
m
an
u
ally
f
r
o
m
r
eg
i
s
tr
y
r
ec
o
r
d
s
.
At
th
e
s
am
e
tim
e,
th
e
lack
o
f
in
f
o
r
m
atio
n
ab
o
u
t
h
ea
lth
f
ac
ilit
ies
cr
ea
tes
b
ias
b
ec
au
s
e
p
r
ev
io
u
s
r
esear
ch
f
o
u
n
d
a
clea
r
ass
o
ciatio
n
b
etwe
en
d
is
tan
ce
to
h
ea
lth
ca
r
e
ce
n
tr
es
an
d
m
ed
ical
v
is
its
.
Acc
o
r
d
in
g
to
Gith
in
ji
et
a
l
.
[
25
]
,
th
e
lik
elih
o
o
d
o
f
s
ee
k
in
g
h
ea
lth
ca
r
e
s
er
v
ices
d
ec
lin
es
b
y
r
o
u
g
h
ly
5
%
with
ea
c
h
ad
d
itio
n
al
k
ilo
m
etr
e.
Fo
r
m
alar
ia
in
p
ar
ticu
lar
,
th
e
r
ep
o
r
ted
in
cid
en
ce
d
ep
en
d
s
o
n
b
o
th
ac
tu
al
tr
an
s
m
is
s
io
n
lev
els
an
d
r
ep
o
r
tin
g
q
u
ality
.
I
n
an
a
r
ea
with
f
ew
h
ea
lt
h
f
ac
ilit
ies,
th
er
e
will
b
e
f
ewe
r
ca
s
es
r
ep
o
r
ted
,
ev
en
if
ac
tu
al
tr
a
n
s
m
is
s
io
n
r
ates
ar
e
eq
u
iv
alen
t
to
th
o
s
e
in
an
ar
ea
with
n
u
m
er
o
u
s
h
ea
lth
f
ac
ilit
ies.
T
h
e
o
m
is
s
io
n
o
f
h
ea
lth
f
ac
il
ity
in
f
o
r
m
atio
n
in
a
p
r
ed
ictiv
e
m
o
d
el
wo
u
ld
m
ea
n
eq
u
atin
g
r
ep
o
r
tin
g
q
u
ality
wit
h
r
is
k
lev
els.
Mb
u
n
g
e
et
a
l
.
[
19
]
cited
t
h
is
as
o
n
e
o
f
two
k
e
y
s
h
o
r
tco
m
in
g
s
in
th
e
Af
r
ican
AI
liter
atu
r
e
o
n
m
al
ar
ia.
Asi
d
e
f
r
o
m
s
tatic
f
ac
ilit
y
in
f
o
r
m
atio
n
,
in
teg
r
atin
g
p
r
ed
ic
tiv
e
m
o
d
els
with
n
atio
n
al
s
u
r
v
eillan
ce
s
y
s
tem
s
in
r
ea
l
tim
e
o
p
en
s
u
p
a
n
ew
r
ea
lm
th
at
r
em
ai
n
s
lar
g
ely
u
n
ex
p
lo
r
ed
.
DHI
S2
,
th
e
s
e
co
n
d
v
e
r
s
io
n
o
f
th
e
d
is
tr
ict
h
ea
lth
in
f
o
r
m
atio
n
s
y
s
tem
,
is
th
e
g
o
-
to
s
y
s
tem
f
o
r
m
an
y
Af
r
ican
m
in
is
tr
ies
o
f
h
ea
lth
an
d
co
n
tain
s
h
ig
h
ly
d
etailed
wee
k
ly
ca
s
e
n
u
m
b
er
s
.
W
h
ile
d
ir
ec
t
ap
p
licatio
n
p
r
o
g
r
a
m
m
in
g
i
n
ter
f
ac
e
(
A
PI
)
in
teg
r
atio
n
with
DHI
S2
ca
n
en
a
b
le
p
r
ed
ictiv
e
m
o
d
ellin
g
u
s
in
g
r
ea
l
-
tim
e
d
ata,
s
ec
u
r
ity
an
d
g
o
v
er
n
a
n
ce
co
n
ce
r
n
s
h
a
v
e
p
r
ev
en
ted
m
an
y
ex
am
p
les f
r
o
m
b
ein
g
p
u
b
lis
h
ed
t
o
d
ate.
T
h
is
r
ev
iew’
s
s
y
n
th
esis
h
as
c
lear
o
p
e
r
atio
n
al
im
p
licatio
n
s
f
o
r
n
atio
n
al
h
ea
lth
i
n
f
o
r
m
atio
n
s
y
s
tem
s
b
ased
o
n
DHI
S2
.
DHI
S2
cu
r
r
en
tly
d
ep
l
o
y
e
d
in
7
3
c
o
u
n
tr
ies
,
in
clu
d
in
g
Z
im
b
ab
we,
Z
am
b
i
a,
Mo
za
m
b
iq
u
e
a
n
d
B
o
ts
wan
a
all
h
ig
h
-
b
u
r
d
e
n
m
a
lar
ia
co
u
n
t
r
ies
in
So
u
t
h
er
n
Af
r
ica.
T
h
e
r
ea
l
-
tim
e
API
(
DHI
S2
w
eb
API
v
2
.
3
9
)
allo
ws
p
r
o
g
r
am
m
atic
ex
tr
ac
ti
o
n
o
f
wee
k
ly
f
ac
ilit
y
-
lev
el
ca
s
e
co
u
n
ts
th
at
ca
n
b
e
u
s
ed
as
b
o
th
tr
ain
in
g
tar
g
ets
an
d
h
ea
lth
-
in
f
r
astru
ctu
r
e
c
o
v
ar
iates
f
o
r
atten
tio
n
-
L
STM
m
o
d
els.
B
u
r
k
in
a
Fas
o
’
s
in
teg
r
atio
n
o
f
DHI
S2
p
r
o
v
e
d
th
at
th
e
ad
o
p
tio
n
o
f
th
e
p
latf
o
r
m
e
n
h
an
ce
d
th
e
tim
elin
ess
o
f
m
alar
ia
d
ata
f
r
o
m
6
2
%
to
8
0
%,
an
d
co
m
p
leten
ess
f
r
o
m
4
3
%
to
8
3
%.
T
h
ese
d
ata
q
u
ality
im
p
r
o
v
e
m
en
ts
u
ltima
tely
r
eso
lv
e
a
lead
in
g
is
s
u
e
d
eter
r
in
g
d
ee
p
lear
n
in
g
m
o
d
el
p
er
f
o
r
m
an
ce
in
Af
r
ica
in
co
m
p
lete
an
d
in
c
o
n
s
is
ten
t
s
u
r
v
eillan
ce
d
ata.
T
h
e
s
y
n
th
esis
p
r
esen
ted
in
th
is
r
ev
iew,
th
er
ef
o
r
e
h
as
n
o
t
ju
s
t
a
th
eo
r
etica
l
p
ath
way
f
o
r
in
teg
r
atio
n
with
ex
is
tin
g
n
atio
n
al
h
ea
lth
in
f
r
ast
r
u
ctu
r
e,
b
u
t r
ath
e
r
an
o
p
er
atio
n
al
o
n
e
th
r
o
u
g
h
th
e
DHI
S2
ec
o
s
y
s
tem
.
3
.
6
.
1
.
Clo
ud
co
m
pu
t
ing
a
nd
big
da
t
a
infr
a
s
t
ruct
ure
f
o
r
m
a
la
ria
s
urv
eilla
nce
T
h
e
ac
tu
al
d
ep
l
o
y
m
en
t
o
f
R
NN
-
b
ased
m
alar
ia
p
r
e
d
ictio
n
m
o
d
els
in
Su
b
-
Sah
ar
a
n
Af
r
ica
i
s
n
o
t
ju
s
t
a
m
o
d
ellin
g
p
r
o
b
lem
it
i
s
a
d
ata
in
f
r
astru
ctu
r
e
p
r
o
b
lem
.
Po
p
u
lar
h
ea
lth
d
ata
p
latf
o
r
m
DHI
S
2
is
wid
ely
u
s
ed
in
Af
r
ica,
g
en
er
atin
g
ten
s
o
f
th
o
u
s
an
d
s
o
f
wee
k
ly
ca
s
e
r
ec
o
r
d
s
f
r
o
m
h
ea
lt
h
f
ac
ilit
ies,
wh
ile
a
n
ea
r
r
ea
l
-
tim
e
API
allo
ws
f
o
r
p
r
o
g
r
am
m
e
d
ata
e
x
tr
ac
tio
n
.
I
n
g
esti
n
g
,
p
r
ep
r
o
ce
s
s
in
g
,
an
d
s
to
r
in
g
m
u
lti
-
y
ea
r
an
d
m
u
lti
-
d
is
tr
ict
clim
ate
an
d
s
u
r
v
eil
lan
ce
tim
e
s
er
ies,
at
th
e
s
ca
le
th
at
is
n
ee
d
ed
f
o
r
d
ee
p
lear
n
in
g
tr
ai
n
in
g
,
is
a
b
i
g
d
ata
p
r
o
b
lem
t
h
at
r
eq
u
ir
es c
lo
u
d
.
Mo
r
e
r
ec
e
n
t
e
x
am
p
les
f
r
o
m
s
im
ilar
d
o
m
ai
n
s
s
h
o
w
th
at
a
cl
o
u
d
-
h
o
s
ted
e
p
id
em
io
lo
g
ical
f
o
r
ec
asti
n
g
p
ip
elin
e
is
f
ea
s
ib
le.
Dep
lo
y
in
g
L
STM
m
o
d
els
o
n
Am
a
zo
n
web
s
er
v
ices
(
AW
S
)
r
ed
u
ce
d
en
d
-
to
-
en
d
p
r
ed
ictio
n
laten
cy
b
y
6
7
%
c
o
m
p
ar
ed
to
o
n
-
p
r
em
is
e
co
m
p
u
tatio
n
lead
in
g
to
n
ea
r
r
ea
l
-
ti
m
e
o
u
tb
r
ea
k
aler
ts
B
an
s
al
et
a
l
.
[
26
]
.
Dis
tr
ib
u
ted
f
r
am
ewo
r
k
s
lik
e
Ap
ac
h
e
Sp
a
r
k
an
d
Go
o
g
le
B
ig
Qu
er
y
h
av
e
also
b
ee
n
u
tili
ze
d
to
p
r
ep
r
o
ce
s
s
m
u
lti
-
s
o
u
r
ce
cli
m
ate
d
atasets
lik
e
clim
ate
h
az
ar
d
s
g
r
o
u
p
in
f
r
ar
ed
p
r
ec
ip
itati
o
n
with
s
tatio
n
d
ata
(
C
HI
R
P
S
)
r
ain
f
all,
m
o
d
er
ate
r
eso
lu
tio
n
im
ag
in
g
s
p
ec
tr
o
r
ad
io
m
eter
(
MO
DI
S
)
lan
d
s
u
r
f
ac
e
tem
p
er
atu
r
e,
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
3
2
2
1
C
o
m
p
u
t Sci
I
n
f
T
ec
h
n
o
l
,
Vo
l.
7
,
No
.
3
,
No
v
em
b
er
20
26
:
3
7
7
-
393
386
E
R
A5
r
ea
n
aly
s
is
d
ata
at
a
c
o
n
tin
en
tal
s
c
ale,
m
a
k
in
g
in
teg
r
atio
n
in
th
e
cl
o
u
d
n
o
t
ju
s
t
c
o
n
v
en
ien
t
b
u
t
r
ath
er
co
m
p
u
tatio
n
ally
n
ec
ess
ar
y
f
o
r
o
p
er
atio
n
al
m
ala
r
ia
f
o
r
ec
as
tin
g
s
y
s
tem
.
T
h
e
im
p
lem
en
t
atio
n
o
f
atten
tio
n
-
L
STM
p
ip
elin
es
in
th
e
clo
u
d
is
th
e
m
o
s
t
p
lau
s
ib
le
r
o
u
te
t
o
s
ca
lab
le
o
p
er
atio
n
f
o
r
So
u
t
h
er
n
Af
r
ica
wh
er
e
n
atio
n
al
h
ea
lth
m
i
n
is
tr
ies h
av
e
lo
w
-
p
er
f
o
r
m
in
g
lo
ca
l c
o
m
p
u
t
in
g
in
f
r
astru
ct
u
r
e.
T
h
e
k
ey
ar
ch
itectu
r
al
in
n
o
v
atio
n
o
f
L
STM
an
d
GR
U
n
etwo
r
k
s
,
f
r
o
m
a
c
o
m
p
u
ter
s
cien
tis
t’
s
p
er
s
p
ec
tiv
e,
is
n
o
t
s
o
m
u
ch
th
eir
p
r
ed
ictiv
e
ac
cu
r
ac
y
b
u
t
th
eir
co
m
p
u
ta
tio
n
al
s
ca
lab
ilit
y
.
B
o
th
o
f
o
u
r
ar
ch
itectu
r
es
ca
n
b
e
im
p
lem
e
n
ted
o
n
s
tan
d
ar
d
g
r
ap
h
ics
p
r
o
ce
s
s
in
g
u
n
it
(
GPU
)
-
ac
ce
ler
ated
h
ar
d
war
e,
a
n
d
h
av
e
n
ativ
e
s
u
p
p
o
r
t
in
T
en
s
o
r
Flo
w
an
d
Py
T
o
r
ch
f
r
am
ewo
r
k
s
,
en
ab
lin
g
tr
ain
in
g
o
n
m
u
lti
-
y
ea
r
,
m
u
lti
-
d
is
tr
ict
s
u
r
v
eillan
ce
d
at
asets
an
d
wit
h
in
p
r
ac
tical
tim
e
b
u
d
g
ets.
T
o
en
s
u
r
e
th
at
L
STM
m
o
d
els
ca
n
b
e
d
ep
lo
y
e
d
in
r
eso
u
r
ce
-
co
n
s
tr
ain
ed
Af
r
ican
h
ea
lth
s
y
s
tem
s
,
we
lev
er
ag
e
m
o
d
el
co
m
p
r
ess
io
n
tech
n
i
q
u
es
(
p
r
u
n
i
n
g
an
d
q
u
an
tizatio
n
)
.
W
e
f
in
d
th
at
th
ese
tech
n
iq
u
es
ca
n
r
e
d
u
ce
m
o
d
el
s
izes
b
y
alm
o
s
t
9
0
%
with
o
u
t
a
s
ig
n
if
ican
t
l
o
s
s
o
f
ac
cu
r
ac
y
.
T
h
is
m
ea
n
s
th
at
d
ep
lo
y
m
e
n
ts
ca
n
b
e
m
a
d
e
at
th
e
lev
el
o
f
d
is
tr
ict
h
ea
lth
o
f
f
ices,
eith
er
o
n
th
e
clo
u
d
o
r
ev
en
o
n
ed
g
e
d
ev
ic
es.
Dee
p
lear
n
i
n
g
m
eth
o
d
s
c
an
s
ca
le,
u
n
lik
e
o
th
er
co
m
p
u
tatio
n
ally
in
ten
s
iv
e
en
s
em
b
le
m
eth
o
d
s
,
w
h
ich
n
ee
d
a
lar
g
e
am
o
u
n
t
o
f
m
e
m
o
r
y
d
u
r
in
g
e
v
er
y
i
n
f
er
en
ce
ca
ll.
T
h
e
h
is
to
r
ical
ev
o
lu
tio
n
o
f
th
e
v
ar
io
u
s
m
eth
o
d
o
l
o
g
ies o
n
ef
f
o
r
ts
to
p
r
ed
ict
m
alar
ia
o
u
tb
r
ea
k
f
r
o
m
th
e
y
ea
r
1
9
9
8
to
2
0
2
5
is
s
h
o
w
n
in
F
ig
u
r
e
3
.
T
h
e
im
p
o
r
tan
t
s
tag
es
th
at
h
av
e
b
ee
n
p
ass
ed
t
o
attain
p
r
ed
ictiv
e
s
o
p
h
is
ticatio
n
ar
e
th
e
clim
ate
en
v
elo
p
e
m
o
d
els
an
d
th
e
n
t
h
e
SAR
I
MA
b
ased
s
tatis
tica
l
s
y
s
tem
s
an
d
th
en
m
o
s
tly
th
e
ML
b
aselin
es.
T
h
en
th
e
L
STM
an
d
T
r
an
s
f
o
r
m
er
m
o
d
els
an
d
m
o
s
t
r
ec
e
n
tly
th
e
atten
tio
n
-
au
g
m
en
ted
m
u
lti
-
s
o
u
r
ce
f
u
s
io
n
p
r
o
p
o
s
ed
b
y
th
is
MT
ec
h
p
r
o
ject.
Fig
u
r
e
3.
Me
th
o
d
o
lo
g
ical
ev
o
l
u
tio
n
tim
elin
e
4.
CO
M
P
ARA
T
I
V
E
ANA
L
YS
I
S O
F
RE
V
I
E
WE
D
S
T
UD
I
E
S
T
ab
le
2
s
u
m
m
ar
is
es
20
r
ep
r
e
s
en
tativ
e
s
tu
d
ies
alo
n
g
f
iv
e
d
im
en
s
io
n
s
:
m
eth
o
d
o
lo
g
y
,
g
e
o
g
r
ap
h
ical
s
ettin
g
,
d
ata
s
o
u
r
ce
s
,
k
ey
f
i
n
d
in
g
s
,
a
n
d
s
tated
lim
itatio
n
s
.
T
h
e
m
atr
ix
s
u
p
p
o
r
ts
r
ap
i
d
co
m
p
ar
is
o
n
a
n
d
id
en
tifie
s
wh
ich
s
tu
d
y
attr
ib
u
tes r
ec
u
r
m
o
s
t o
f
te
n
an
d
wh
ich
ar
e
co
n
s
p
icu
o
u
s
ly
ab
s
en
t.
4
.
1
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
o
f
re
v
iewe
d studies
T
ab
le
2
s
u
m
m
ar
izes
twen
ty
r
ep
r
esen
tativ
e
s
tu
d
ies
ac
r
o
s
s
t
h
e
f
o
llo
win
g
d
im
en
s
io
n
s
m
eth
o
d
o
lo
g
y
,
g
eo
g
r
a
p
h
ical
s
ettin
g
,
d
ata
s
o
u
r
ce
s
,
k
ey
f
in
d
in
g
s
a
n
d
lim
itatio
n
s
s
tated
.
T
h
is
m
atr
ix
is
in
te
n
d
ed
to
f
ac
ilit
ate
an
ef
f
icien
t
s
tr
u
ctu
r
ed
co
m
p
a
r
is
o
n
;
it
d
o
es
n
o
t
r
eh
ash
th
e
th
em
atic
r
ev
iew
in
s
ec
tio
n
3
,
r
ath
e
r
it
is
an
an
aly
tical
cr
o
s
s
-
s
ec
tio
n
th
at
illu
s
tr
ates
wh
er
e
p
r
ev
ale
n
ce
an
d
ab
s
en
c
e
ex
is
t
in
m
eth
o
d
o
lo
g
ical
ch
a
r
ac
ter
is
tic
ac
r
o
s
s
th
e
liter
atu
r
e.
On
e
ca
n
n
o
tice
m
u
ltip
le
p
atter
n
s
f
r
o
m
in
s
p
ec
tio
n
o
f
T
a
b
le
2
.
On
e
ca
n
o
b
s
er
v
e
a
d
ef
i
n
itiv
e
an
d
s
tead
y
tem
p
o
r
al
ev
o
lu
tio
n
in
m
eth
o
d
o
l
o
g
y
:
Stu
d
ies
th
at
we
r
e
p
u
b
lis
h
ed
b
ef
o
r
e
2
0
1
5
,
p
r
e
d
o
m
in
an
tly
em
p
l
o
y
s
tatis
t
ical
ap
p
r
o
ac
h
es
(
SAR
I
MA
,
GL
M,
an
d
clim
ate
en
v
elo
p
e
m
o
d
els)
wh
er
ea
s
s
tu
d
ies
th
at
wer
e
p
u
b
lis
h
ed
f
r
o
m
2
0
1
8
.
T
h
e
s
h
if
t
s
ee
n
h
er
e
is
ak
in
to
th
e
wid
er
m
atu
r
at
io
n
o
f
d
ee
p
lear
n
i
n
g
in
ap
p
lied
s
ettin
g
s
,
an
d
it
is
co
n
s
is
ten
t
with
Z
in
s
ze
r
et
a
l
.
[
12
]
f
i
n
d
in
g
th
at
o
p
e
r
atio
n
al
m
alar
ia
ea
r
ly
war
n
in
g
s
y
s
tem
s
wer
e
d
e
v
elo
p
ed
alm
o
s
t
ex
clu
s
iv
ely
o
n
SAR
I
MA
f
r
am
ewo
r
k
s
as
r
ec
en
tly
as
2
0
1
2
.
T
h
e
im
p
licatio
n
f
o
r
p
r
a
ctitio
n
er
s
is
th
at
th
e
f
ield
h
as a
d
v
a
n
ce
d
at
a
q
u
ick
e
r
p
ac
e
th
an
th
e
o
p
er
atio
n
al
h
ea
lth
s
y
s
tem
s
th
at
it is
aim
ed
at
b
en
ef
itti
n
g
f
r
o
m
.
Nex
t,
atten
tio
n
m
ec
h
a
n
is
m
s
,
wh
ich
ar
e
u
b
iq
u
ito
u
s
in
N
L
P
an
d
f
in
an
cial
tim
e
-
s
er
ies
p
r
ed
ictio
n
,
f
ea
tu
r
e
in
o
n
ly
o
n
e
d
is
ea
s
e
-
f
o
r
ec
asti
n
g
s
tu
d
y
in
t
h
e
m
atr
i
x
,
an
d
th
at
ap
p
licatio
n
was
d
e
n
g
u
e
in
E
ast
Asi
a,
n
o
t
m
alar
ia
in
A
f
r
ica.
T
h
er
e
i
s
an
im
p
o
r
tan
t
m
eth
o
d
o
lo
g
ical
asy
m
m
etr
y
h
er
e:
t
h
e
m
et
h
o
d
th
at
h
as
m
o
s
t
im
p
r
o
v
e
d
Evaluation Warning : The document was created with Spire.PDF for Python.