I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
.
10
, N
o.
1
,
M
a
r
c
h
202
1
, pp.
25
7
~
264
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
10
.i
1
.pp
25
7
-
264
257
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
Plan
t
d
i
se
ase
p
r
e
d
i
c
t
i
on
u
si
n
g c
l
ass
i
f
i
c
at
i
on
al
gor
i
t
h
m
s
M
ar
ia
M
or
gan
1
, C
ar
la
B
la
n
k
2
,
R
ae
d
S
e
e
t
an
3
1
,2
Department of Mathemati
cs and Statist
ics, Slippery
Rock University,
USA
3
Department of Compu
ter Science, Slipp
ery Rock
University, US
A
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
F
e
b
8
,
20
20
R
e
vi
s
e
d
J
ul
2
1
, 20
20
A
c
c
e
pt
e
d
F
e
b 17
, 20
2
1
This
paper
investigates
the
capability
of
six
existing
classification
alg
orithms
(artificial
neural
network,
naïve
bayes,
k
-
nearest
neighbor,
support
vector
machine,
decision
tree
and
random
forest)
in
classifying
and
pre
dicting
diseases
in
soybean
and
mushroom
datasets
using
datasets
with
nume
rical
or
categorical
attribut
es.
While
many
simil
ar
studies
have
been
condu
cted
on
datasets
of
images
to
predict
plant
diseases,
the
main
objective
of
thi
s
study
is
to
suggest
classification
methods
that
can
be
used
for
diseas
e
classif
ication
and
prediction
in
datasets
th
at
contain
raw
measurements
instead
of
i
mages.
A
fungus
and
a
plant
dataset,
which
had
many
differences,
were
ch
osen
so
that
the
findings
in
this
paper
could
be
applied
to
future
research
for
disease
prediction
and
classification
in
a
variety
of
datasets
which
conta
in
raw
measureme
nts.
A
key
differe
nce
between
the
two
dat
asets,
other
th
an
one
being
a
fungus
and
one
being
a
plant,
is
that
the
mushroom
dat
aset
is
balanced
and
only
contain
ed
two
classes
while
the
soybean
dat
aset
is
imbalance
d
and
contained
eighteen
classes.
All
six
algorithms
per
formed
well
on
the
mushroom
da
taset,
while
the
artificial
neural
network
and
k
-
nearest
neighbor
algorithms
performed
best
on
the
soybean
datas
et.
The
findings
of
this
paper
can
be
applied
to
future
research
on
disease
classifi
cation
and
predicti
on
in
a
variety
of
dataset
types
such
as
fungi,
plants, humans, and animals.
K
e
y
w
o
r
d
s
:
C
la
s
s
if
ic
a
ti
on
M
us
hr
oom
P
la
nt
d
is
e
a
s
e
P
r
e
di
c
ti
on
S
oybe
a
n
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
R
a
e
d S
e
e
ta
n
D
e
pa
r
tm
e
nt
of
C
om
put
e
r
S
c
ie
nc
e
S
li
ppe
r
y R
oc
k U
ni
ve
r
s
it
y, U
S
A
E
m
a
il
:
r
a
e
d.s
e
e
ta
n@
s
r
u.e
du
1.
I
N
T
R
O
D
U
C
T
I
O
N
T
he
m
a
in
go
a
l
of
th
is
p
a
pe
r
is
to
te
s
t
th
e
a
c
c
ur
a
c
y
a
nd
c
om
pa
r
e
th
e
r
e
s
ul
t
s
of
e
xi
s
ti
ng
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
in
pr
e
di
c
ti
ng
e
di
bi
li
ty
in
m
us
hr
oom
s
a
nd
c
la
s
s
if
yi
ng
di
s
e
a
s
e
s
in
s
oybe
a
n
pl
a
nt
s
.
W
hi
le
th
e
m
us
hr
oom
a
nd
s
oyb
e
a
n
d
a
ta
s
e
t
s
u
s
e
d
in
th
is
p
a
pe
r
ha
ve
m
a
n
y
di
f
f
e
r
e
nc
e
s
,
th
e
y
a
r
e
s
im
il
a
r
in
th
a
t
th
e
y
a
r
e
da
ta
s
e
ts
of
e
it
he
r
num
e
r
ic
a
l
or
c
a
te
gor
ic
a
l
a
tt
r
ib
ut
e
s
,
w
hi
le
m
a
ny
s
im
il
a
r
s
tu
di
e
s
ha
ve
b
e
e
n
c
onduc
t
e
d
on
da
ta
s
e
ts
of
i
m
a
ge
s
i
ns
te
a
d of
r
a
w
m
e
a
s
ur
e
m
e
nt
s
[
1
-
4]
. T
he
obj
e
c
ti
ve
of
t
he
a
na
ly
s
is
c
ondu
c
te
d i
n t
hi
s
pa
pe
r
i
s
to
m
a
ke
s
ug
ge
s
ti
ons
to
a
gr
ic
ul
tu
r
a
l
r
e
s
e
a
r
c
he
r
s
,
or
di
s
e
a
s
e
r
e
s
e
a
r
c
he
r
s
in
ge
ne
r
a
l,
on
c
la
s
s
if
ic
a
ti
on
m
e
th
ods
th
a
t
pe
r
f
or
m
w
e
ll
,
in
te
r
m
s
o
f
di
s
e
a
s
e
pr
e
di
c
ti
on
a
nd
c
la
s
s
if
ic
a
ti
on
a
c
c
ur
a
c
y,
on
da
ta
s
e
ts
w
it
h
r
a
w
m
e
a
s
ur
e
m
e
nt
s
.
W
hi
le
im
a
ge
da
t
a
s
e
t
s
ha
ve
be
e
n
te
s
te
d
w
it
h
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
pr
e
s
e
nt
e
d
he
r
e
,
m
a
ny
r
e
s
e
a
r
c
he
r
s
s
ti
ll
pr
e
f
e
r
to
ta
ke
a
nd
r
e
c
or
d
m
e
a
s
ur
e
m
e
nt
s
by
ha
nd
w
h
e
n
s
tu
dyi
ng
pl
a
nt
s
or
f
ungi
.
S
oybe
a
ns
a
nd
m
us
hr
oom
s
a
r
e
ve
r
y
im
por
ta
nt
to
hum
a
ns
;
th
us
,
it
is
im
po
r
ta
nt
to
ha
ve
a
c
c
ur
a
te
m
e
th
ods
to
pr
e
di
c
t
w
he
th
e
r
or
not
di
f
f
e
r
e
nt
va
r
ia
ti
ons
a
r
e
s
a
f
e
f
o
r
hum
a
n
c
ons
um
pt
io
n
a
nd
pr
e
di
c
t
th
e
pr
e
s
e
nc
e
of
a
ny
di
s
e
a
s
e
s
t
ha
t
c
a
n
a
f
f
e
c
t
th
e
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
10
, N
o.
1
,
M
a
r
c
h
20
2
1
:
25
7
–
26
4
258
T
he
r
e
a
r
e
bot
h pois
onous
a
nd e
di
bl
e
m
u
s
hr
oom
s
. A
c
c
or
di
ng t
o T
he
A
udubon S
oc
ie
ty
F
ie
ld
G
ui
de
of
N
or
th
A
m
e
r
ic
a
n
M
us
hr
oom
s
,
th
e
r
e
i
s
no
s
in
gl
e
c
h
a
r
a
c
te
r
is
ti
c
t
o
di
s
ti
ngui
s
h
be
twe
e
n
e
di
bl
e
m
us
hr
oom
s
a
nd
poi
s
onous
m
us
hr
oom
s
[
5
-
6]
.
O
ne
m
us
t
be
c
e
r
ta
in
a
m
us
hr
oo
m
is
one
of
th
e
e
di
bl
e
va
r
ie
ti
e
s
,
ot
he
r
w
is
e
,
th
e
m
us
hr
oom
s
houl
d be
c
ons
id
e
r
e
d pois
onou
s
. S
in
c
e
va
r
io
us
t
ype
s
of
m
us
hr
oom
s
a
r
e
c
ons
um
e
d by humans
, i
t
is
im
por
ta
nt
to
e
s
ta
bl
is
h
s
om
e
gui
de
li
ne
s
to
d
e
te
r
m
in
e
if
a
m
us
hr
oom
is
e
di
bl
e
or
not
.
I
n
th
is
pa
pe
r
w
e
w
il
l
a
tt
e
m
pt
to
tr
a
in
e
xi
s
ti
ng
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
th
a
t
c
a
n
be
u
s
e
d
to
c
la
s
s
if
y
m
us
hr
oom
s
,
gi
ve
n
a
da
ta
s
e
t
of
r
a
w
m
e
a
s
ur
e
m
e
nt
s
, a
s
e
it
he
r
e
di
bl
e
or
poi
s
onou
s
.
S
oybe
a
ns
a
r
e
pr
oc
e
s
s
e
d
f
or
th
e
ir
oi
l
a
nd
m
e
a
l
[
7
]
.
S
oybe
a
n
oi
l
is
us
e
d
in
m
a
ny
f
oods
th
a
t
hum
a
ns
c
ons
um
e
da
il
y
s
uc
h
a
s
m
a
r
ga
r
in
e
,
ba
ke
d
br
e
a
ds
,
c
a
nne
d
tu
na
,
a
nd
f
r
ie
d
f
ood.
S
oybe
a
n
m
e
a
l
is
u
s
e
d
in
f
ood
f
or
m
a
ny
f
a
r
m
a
n
im
a
ls
s
uc
h
a
s
poul
tr
y,
por
k,
a
nd
c
a
tt
le
.
S
oyb
e
a
ns
a
r
e
a
n
im
por
ta
nt
c
r
op
be
c
a
us
e
th
e
oi
l
is
di
r
e
c
tl
y
put
in
to
f
ood
th
a
t
hum
a
ns
c
ons
um
e
,
a
nd
th
e
m
e
a
l
i
s
f
e
d
to
th
e
a
ni
m
a
ls
th
a
t
a
r
e
w
id
e
ly
c
ons
um
e
d
by
hum
a
ns
.
T
h
e
r
e
a
r
e
v
a
r
io
us
di
s
e
a
s
e
s
th
a
t
a
f
f
e
c
t
s
oybe
a
n
c
r
o
ps
,
in
th
is
pa
pe
r
w
e
w
il
l
a
tt
e
m
pt
to
tr
a
in
e
xi
s
ti
ng
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
th
a
t
c
a
n
be
us
e
d
to
c
la
s
s
if
y
s
oybe
a
ns
pl
a
nt
s
a
s
ha
vi
ng
a
pa
r
ti
c
ul
a
r
di
s
e
a
s
e
,
ba
s
e
d
on
a
da
ta
s
e
t
of
r
a
w
m
e
a
s
ur
e
m
e
nt
s
pe
r
ta
in
in
g t
o t
he
s
oyb
e
a
n pl
a
nt
s
.
D
is
c
ove
r
in
g
a
ppl
ic
a
ti
ons
a
nd
t
e
c
hni
que
s
f
or
pr
e
di
c
ti
ng
di
s
e
a
s
e
pr
e
s
e
nc
e
a
nd
c
la
s
s
if
yi
ng
di
s
e
a
s
e
s
is
ve
r
y
im
por
ta
nt
w
he
n
it
c
om
e
s
to
a
gr
ic
ul
tu
r
e
.
D
is
e
a
s
e
s
in
c
r
ops
c
a
n
ha
ve
a
s
e
r
io
us
im
pa
c
t
on
th
e
c
r
op
yi
e
ld
[
8]
.
B
e
c
a
us
e
di
s
e
a
s
e
s
w
il
l
m
or
e
th
a
n
li
ke
ly
d
a
m
a
ge
a
la
r
ge
nu
m
be
r
of
c
r
ops
in
a
gr
ow
in
g
c
yc
le
,
f
a
r
m
e
r
s
c
a
n
be
ne
f
it
f
r
om
c
la
s
s
if
ic
a
ti
on
of
c
r
op
di
s
e
a
s
e
s
a
nd
r
is
k
f
a
c
to
r
s
th
a
t
m
a
y
le
a
d
to
th
e
s
e
di
s
e
a
s
e
s
.
A
f
or
e
c
a
s
ti
ng
s
ys
te
m
ha
s
be
e
n
de
ve
lo
p
e
d
to
pr
e
di
c
t
di
s
e
a
s
e
out
br
e
a
k
in
s
tr
a
w
be
r
r
y
pl
a
nt
s
in
F
lo
r
id
a
,
w
he
r
e
15%
o
f
U
S
be
r
r
ie
s
a
r
e
pr
oduc
e
d
a
nd
a
ll
be
r
r
ie
s
gr
ow
n
in
th
e
w
in
te
r
[
9]
.
T
he
f
or
e
c
a
s
ti
ng
s
ys
t
e
m
,
c
a
ll
e
d
th
e
S
tr
a
w
be
r
r
y
A
dvi
s
or
y
S
ys
te
m
(
S
A
S
)
,
he
lp
s
f
a
r
m
e
r
s
by
pr
e
di
c
ti
ng
th
e
di
s
e
a
s
e
in
c
id
e
nc
e
r
e
c
om
m
e
ndi
ng
f
ungi
c
id
e
a
ppl
ic
a
ti
ons
[
9]
. T
hi
s
s
y
s
te
m
ha
s
r
e
du
c
e
d pr
oduc
ti
on c
os
ts
by
e
li
m
in
a
ti
ng unne
c
e
s
s
a
r
y f
ungi
c
id
e
a
ppl
ic
a
ti
on
s
,
w
hi
le
not
r
is
ki
ng
th
e
c
r
op
yi
e
ld
.
A
s
R
ic
ha
r
d
S
t
r
a
nge
not
e
d,
a
lm
os
t
10%
of
gl
oba
l
f
ood
pr
oduc
ti
on
is
lo
s
t
due
to
pl
a
nt
di
s
e
a
s
e
[
10]
.
T
he
s
e
lo
s
s
e
s
c
a
n
b
e
m
in
im
iz
e
d
if
a
c
c
ur
a
te
m
e
th
ods
a
r
e
de
ve
lo
pe
d
f
or
pr
e
di
c
ti
ng
a
nd
c
la
s
s
if
yi
ng dis
e
a
s
e
.
T
he
r
e
m
a
in
de
r
of
th
is
pa
pe
r
is
s
tr
uc
tu
r
e
d
a
s
:
s
e
c
ti
on
2
di
s
c
us
s
e
s
th
e
li
te
r
a
tu
r
e
r
e
vi
e
w
w
or
ks
.
S
e
c
ti
on
3
pr
e
s
e
nt
s
our
r
e
s
e
a
r
c
h
m
e
th
od.
S
e
c
ti
on
4
di
s
c
us
s
e
s
th
e
r
e
s
ul
t
s
of
our
p
a
pe
r
.
S
e
c
ti
on
5
pr
ovi
de
s
c
onc
lu
s
io
n a
nd r
e
c
om
m
e
nda
ti
on
s
f
or
f
ur
th
e
r
s
tu
di
e
s
.
2.
L
I
T
E
R
A
T
U
R
E
R
E
V
I
E
W
T
o
da
te
,
m
os
t
s
tu
di
e
s
of
th
is
ty
pe
ha
ve
us
e
d
im
a
ge
s
of
pl
a
nt
s
or
f
ungi
a
s
th
e
da
ta
s
e
ts
w
hi
c
h
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
a
r
e
te
s
te
d
on.
P
r
e
vi
ous
s
tu
di
e
s
ha
v
e
f
ound
th
a
t
de
c
is
io
n
tr
e
e
s
a
r
e
w
id
e
ly
us
e
d
be
c
a
us
e
of
th
e
ir
e
a
s
e
of
in
te
r
pr
e
ta
ti
on,
s
uppor
t
ve
c
to
r
m
a
c
hi
ne
s
(
S
V
M
)
a
nd
a
r
t
if
ic
ia
l
ne
ur
a
l
ne
twor
ks
(
A
N
N
)
a
r
e
ty
pi
c
a
ll
y
th
e
m
os
t
a
c
c
ur
a
te
,
a
nd
k
-
ne
a
r
e
s
t
ne
ig
hbor
(
K
N
N
)
a
nd
na
ïv
e
ba
ye
s
a
r
e
not
th
e
be
s
t
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
f
or
a
g
r
ic
ul
tu
r
e
but
th
e
y
a
r
e
e
a
s
y
to
tr
a
in
a
nd
th
us
ha
ve
be
e
n
us
ed
in
m
a
ny
pl
a
nt
a
nd
f
ungi
di
s
e
a
s
e
c
la
s
s
if
ic
a
ti
on s
tu
di
e
s
[
1
1
].
T
he
s
ix
c
la
s
s
if
ic
a
ti
on
m
e
th
ods
c
ho
s
e
n
f
or
c
om
pa
r
is
on
in
th
i
s
s
tu
dy
w
e
r
e
b
a
s
e
d
on
th
e
li
te
r
a
tu
r
e
r
e
vi
e
w
e
d
pr
io
r
to
be
gi
nni
ng
th
e
e
xpe
r
im
e
nt
.
I
n
N
ove
m
be
r
2018,
a
s
tu
dy
w
a
s
pu
bl
is
he
d
in
w
hi
c
h
3
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
w
e
r
e
te
s
te
d
to
c
om
pa
r
e
th
e
ir
a
c
c
ur
a
c
y
in
pr
e
di
c
ti
ng
di
s
e
a
s
e
s
in
pl
a
nt
s
,
ba
s
e
d
on
a
da
ta
s
e
t
of
pl
a
nt
l
e
a
f
i
m
a
ge
s
.
T
hi
s
s
tu
dy f
ound tha
t
th
e
de
c
is
io
n t
r
e
e
a
lg
or
it
hm
pe
r
f
or
m
e
d be
tt
e
r
t
ha
n A
N
N
a
nd
na
ïv
e
ba
ye
s
[
1]
.
A
not
he
r
s
tu
dy,
publ
is
he
d
in
M
a
r
c
h
2018,
c
om
pa
r
e
d
th
e
c
la
s
s
if
ic
a
ti
on
a
c
c
ur
a
c
y
of
pr
e
di
c
ti
ng
lo
s
s
c
a
us
e
d
by
gr
a
s
s
gr
ub
in
s
e
c
t
us
in
g
th
e
f
ol
lo
w
in
g
te
c
hni
que
s
:
de
c
is
io
n
tr
e
e
,
r
a
ndom
f
or
e
s
t,
ne
ur
a
l
ne
twor
ks
,
ga
us
s
ia
n
na
ïv
e
ba
y
e
s
,
S
V
M
s
,
a
nd
K
N
N
[
12]
.
T
he
da
ta
s
e
t
us
e
d
in
[
12]
w
a
s
c
om
pa
r
a
bl
e
to
th
e
da
ta
us
e
d
in
th
is
s
tu
dy
be
c
a
u
s
e
it
w
a
s
a
da
t
a
s
e
t
of
r
e
a
l
r
e
c
or
de
d
v
a
lu
e
s
,
in
s
te
a
d
of
im
a
ge
s
.
H
o
w
e
ve
r
,
th
e
m
a
in
di
f
f
e
r
e
nc
e
be
twe
e
n
our
pr
opos
e
d
s
tu
dy
a
nd
[
12]
is
th
a
t
our
s
t
udy
is
f
oc
us
e
d
on
pr
e
di
c
ti
ng
th
e
pr
e
s
e
nc
e
of
di
s
e
a
s
e
a
nd
c
l
a
s
s
if
yi
ng
th
e
ty
pe
s
of
di
s
e
a
s
e
s
;
w
hi
le
th
e
m
a
in
go
a
l
of
[
12]
is
to
pr
e
di
c
t
th
e
lo
s
s
of
c
r
ops
due
to
di
s
e
a
s
e
.
T
he
M
a
r
c
h
2018
s
tu
dy
f
ound
th
a
t
ne
ur
a
l
ne
twor
ks
,
r
a
n
dom
f
or
e
s
t,
a
nd
ga
us
s
ia
n
na
iv
e
b
a
ye
s
w
e
r
e
th
e
m
os
t
a
c
c
ur
a
te
in
pr
e
di
c
ti
ng
di
s
e
a
s
e
s
in
c
r
ops
.
F
in
a
ll
y,
a
s
tu
d
y
publ
is
he
d
in
F
e
br
ua
r
y
2019
c
om
pa
r
e
d
th
e
a
c
c
ur
a
c
y
of
S
V
M
a
nd
A
N
N
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
in
pr
e
di
c
ti
ng
di
s
e
a
s
e
s
in
pl
a
nt
s
u
s
in
g
a
da
ta
s
e
t
of
im
a
ge
s
, t
hi
s
s
tu
dy f
ound tha
t
A
N
N
w
a
s
t
he
m
o
s
t
a
c
c
ur
a
te
a
lg
or
i
th
m
[
2]
.
I
n
th
is
s
tu
dy,
w
e
w
il
l
c
om
pa
r
e
th
e
a
c
c
ur
a
c
y
of
s
ix
di
f
f
e
r
e
nt
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
in
pr
e
di
c
ti
ng
di
s
e
a
s
e
s
in
s
oybe
a
n
pl
a
nt
s
a
nd
e
di
bi
li
ty
in
m
us
hr
oom
s
:
a
r
ti
f
i
c
ia
l
ne
ur
a
l
ne
twor
k
(
A
N
N
)
,
na
ïv
e
ba
ye
s
,
k
-
ne
a
r
e
s
t
ne
ig
hbor
(
K
N
N
)
,
s
uppor
t
ve
c
to
r
m
a
c
hi
ne
(
S
V
M
)
,
de
c
i
s
io
n
tr
e
e
,
a
n
d
r
a
ndom
f
or
e
s
t.
T
he
r
e
s
ul
ts
of
th
is
s
tu
dy w
il
l
be
c
om
pa
r
e
d t
o t
hos
e
m
e
nt
io
ne
d i
n t
he
l
it
e
r
a
tu
r
e
r
e
vi
e
w
of
s
im
il
a
r
s
tu
di
e
s
t
ha
t
ha
ve
b
e
e
n done
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
P
la
nt
di
s
e
as
e
pr
e
di
c
ti
on u
s
in
g c
la
s
s
if
ic
at
io
n al
gor
it
hm
s
(
M
ar
ia
M
or
gan
)
259
3.
R
E
S
E
A
R
C
H
M
E
T
H
O
D
T
he
pur
pos
e
of
th
is
s
tu
dy
is
to
a
s
s
e
s
s
th
e
c
a
pa
bi
li
ty
of
s
i
x
e
xi
s
ti
ng
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
(
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k
,
na
ïv
e
ba
ye
s
,
k
-
ne
a
r
e
s
t
ne
ig
hbor
,
s
up
por
t
ve
c
to
r
m
a
c
hi
ne
,
de
c
is
io
n
tr
e
e
a
nd
r
a
ndom
f
or
e
s
t)
in
c
la
s
s
if
yi
ng
a
nd
p
r
e
di
c
ti
ng
di
s
e
a
s
e
s
in
s
oybe
a
n
a
nd
m
us
hr
oom
da
ta
s
e
ts
.
I
n
th
is
s
e
c
ti
on,
w
e
w
il
l
di
s
c
us
s
our
m
e
th
odol
ogy
s
ta
r
ti
ng
w
it
h
da
ta
pr
e
pa
r
a
ti
on,
th
e
n
in
tr
oduc
in
g
th
e
c
la
s
s
if
ic
a
ti
on
m
e
th
od
s
,
a
nd
f
in
a
ll
y
e
va
lu
a
ti
on me
tr
ic
s
. I
n t
he
ne
xt
s
e
c
ti
on, we
w
il
l
di
s
c
us
s
t
he
e
xpe
r
im
e
nt
s
r
e
s
ul
ts
.
3.1.
D
at
a
p
r
e
p
ar
at
io
n
T
he
m
us
hr
oom
da
ta
s
e
t,
obt
a
in
e
d f
r
om
U
C
I
m
a
c
hi
ne
l
e
a
r
ni
ng r
e
pos
it
or
y
, c
ont
a
in
s
8,124 hypothetica
l
s
a
m
pl
e
s
of
23 s
pe
c
ie
s
of
gi
ll
e
d m
us
hr
oom
s
i
n t
he
A
ga
r
ic
u
s
a
nd
L
e
pi
ot
a
f
a
m
il
ie
s
w
it
h 22 c
a
te
gor
ic
a
l
a
tt
r
ib
ut
e
s
[6
, 13]
. T
he
s
pe
c
ie
s
a
r
e
c
la
s
s
if
ie
d a
s
e
di
bl
e
or
poi
s
onous
. A
ny mus
hr
oom
t
ha
t
c
a
nnot
be
c
a
te
gor
iz
e
d a
s
e
di
bl
e
is
c
ons
id
e
r
e
d
poi
s
onous
,
r
e
ga
r
dl
e
s
s
of
w
he
th
e
r
it
is
poi
s
onous
.
F
or
th
e
pu
r
pos
e
of
our
c
om
pa
r
is
on
in
th
is
s
tu
dy
be
twe
e
n
th
e
m
u
s
hr
oom
a
nd
s
oybe
a
n
da
ta
s
e
ts
,
th
e
poi
s
on
ous
c
l
a
s
s
if
ic
a
ti
on
w
il
l
be
tr
e
a
te
d
a
s
th
e
di
s
e
a
s
e
be
in
g
pr
e
s
e
nt
a
nd
th
e
e
di
bl
e
c
la
s
s
if
ic
a
ti
on
w
il
l
be
tr
e
a
te
d
a
s
th
e
di
s
e
a
s
e
not
be
in
g
pr
e
s
e
nt
.
T
he
a
tt
r
ib
ut
e
s
of
th
e
m
us
hr
oom
da
ta
s
e
t
a
r
e
:
c
a
p
-
s
ha
p
e
,
c
a
p
-
s
ur
f
a
c
e
,
c
a
p
-
c
ol
or
,
b
r
ui
s
e
s
,
odor
,
gi
ll
-
a
tt
a
c
hm
e
nt
,
gi
ll
-
sp
a
c
in
g,
gi
ll
-
s
iz
e
,
gi
ll
-
c
ol
or
,
s
ta
lk
-
s
ha
p
e
,
s
t
a
lk
-
r
oot
,
s
ta
lk
-
s
ur
f
a
c
e
-
a
bove
-
r
in
g,
s
ta
lk
-
s
ur
f
a
c
e
-
be
lo
w
-
r
in
g,
s
ta
lk
-
c
ol
or
-
a
bove
-
r
in
g,
s
ta
lk
-
c
ol
or
-
be
lo
w
-
r
in
g,
ve
il
-
ty
pe
,
ve
il
-
c
ol
or
,
r
in
g
-
num
be
r
,
r
in
g
-
ty
pe
,
s
por
e
-
pr
in
t
-
c
ol
or
,
popula
ti
on,
a
nd
ha
bi
ta
t.
T
he
s
oyb
e
a
n
da
ta
s
e
t,
a
ls
o
obt
a
in
e
d
f
r
om
U
C
I
m
a
c
hi
ne
le
a
r
ni
ng
r
e
pos
it
or
y,
c
ont
a
in
s
307
obs
e
r
va
ti
ons
f
r
om
s
oybe
a
n
pl
a
nt
s
in
f
e
c
te
d
w
it
h
19
di
f
f
e
r
e
nt
di
s
e
a
s
e
s
a
nd
35
c
a
te
gor
ic
a
l
a
tt
r
ib
ut
e
s
[
14]
a
nd
[
13]
.
T
he
di
s
e
a
s
e
s
pr
e
s
e
nt
in
th
e
s
oybe
a
n
da
ta
s
e
t
a
r
e
di
a
por
th
e
-
s
te
m
-
can
ke
r
,
c
ha
r
c
oa
l
-
r
ot
,
r
hi
z
oc
to
ni
a
-
r
oot
-
r
ot
,
phyt
opht
hor
a
-
r
ot
,
br
ow
n
-
s
te
m
-
r
ot
,
pow
de
r
y
-
m
i
ld
e
w
,
dow
ny
-
m
il
de
w
,
br
ow
n
-
s
pot
,
ba
c
te
r
ia
l
-
bl
ig
ht
,
ba
c
te
r
ia
l
-
pus
tu
le
,
pur
pl
e
-
s
e
e
d
-
s
ta
in
,
a
nt
hr
a
c
nos
e
,
phyl
lo
s
ti
c
ta
-
le
a
f
-
s
pot
,
a
lt
e
r
na
r
ia
le
a
f
-
s
pot
,
f
r
og
-
e
ye
-
le
a
f
-
s
pot
,
di
a
po
r
th
e
-
pod
-
&
-
s
te
m
-
bl
ig
ht
,
c
ys
t
-
ne
m
a
to
de
,
2
-
4
-
d
-
in
ju
r
y,
a
nd
he
r
bi
c
id
e
-
in
ju
r
y.
T
he
35
c
a
te
gor
ic
a
l
a
tt
r
ib
ut
e
s
pr
e
s
e
nt
in
th
e
s
oybe
a
n
da
ta
s
e
t
a
r
e
da
t
e
,
pl
a
nt
-
s
ta
nd
,
pr
e
c
ip
,
te
m
p,
ha
il
,
c
r
op
-
hi
s
t,
a
r
e
a
-
da
m
a
ge
d,
s
e
ve
r
it
y,
s
e
e
d
-
tm
t,
ge
r
m
in
a
ti
on,
pl
a
nt
-
gr
ow
th
,
le
a
ve
s
,
le
a
f
s
p
ot
s
-
ha
lo
,
le
a
f
s
pot
s
-
m
a
r
g,
le
a
f
s
pot
-
s
iz
e
,
le
a
f
-
s
hr
e
a
d,
le
a
f
-
m
a
lf
,
le
a
f
-
m
il
d,
s
te
m
,
lo
dgi
ng,
s
te
m
-
c
a
nke
r
s
,
c
a
nke
r
-
le
s
io
n,
f
r
ui
ti
ng
-
bodi
e
s
,
e
xt
e
r
na
l
de
c
a
y,
m
yc
e
li
um
,
in
t
-
di
s
c
ol
or
,
s
c
le
r
ot
ia
,
f
r
ui
t
-
pods
,
f
r
ui
t
s
pot
s
,
s
e
e
d,
m
ol
d
-
gr
ow
th
,
s
e
e
d
-
di
s
c
ol
or
,
s
e
e
d
-
s
iz
e
,
s
hr
iv
e
li
ng, a
nd r
oot
s
.
T
he
a
tt
r
ib
ut
e
va
lu
e
s
in
th
e
m
u
s
hr
oom
da
ta
s
e
t
w
e
r
e
c
od
e
d
num
e
r
ic
a
ll
y.
A
n
R
pr
ogr
a
m
w
a
s
w
r
it
te
n
to
r
e
pl
a
c
e
th
e
s
e
num
e
r
ic
va
lu
e
s
w
it
h
th
e
ir
tr
ue
va
lu
e
s
.
T
he
de
s
c
r
ip
ti
on
of
th
is
da
ta
s
e
t
f
r
om
th
e
U
C
I
M
a
c
hi
ne
L
e
a
r
ni
ng
R
e
pos
it
or
y
di
s
c
us
s
e
d
th
e
one
a
tt
r
ib
ut
e
(
s
ta
lk
_r
oot
)
w
he
r
e
va
lu
e
s
w
e
r
e
m
is
s
in
g.
T
hi
s
w
a
s
ve
r
if
ie
d
us
in
g
M
ic
r
os
of
t
E
xc
e
l
be
c
a
u
s
e
of
th
e
s
im
pl
ic
it
y
of
th
e
da
ta
s
e
t.
A
de
c
i
s
io
n
w
a
s
m
a
d
e
to
r
un
e
a
c
h
of
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
on
two
ve
r
s
io
ns
of
th
e
m
us
hr
oom
da
ta
s
e
t;
one
ve
r
s
io
n
w
it
h
a
ll
a
tt
r
ib
ut
e
s
in
c
lu
de
d
a
nd
a
not
he
r
ve
r
s
io
n
w
it
h
th
e
s
ta
lk
_r
oot
a
tt
r
ib
ut
e
r
e
m
ove
d.
T
he
pur
pos
e
of
c
r
e
a
ti
ng
th
e
s
e
two
ve
r
s
io
ns
w
a
s
to
in
ve
s
ti
ga
te
w
he
th
e
r
or
not
th
e
a
tt
r
ib
ut
e
w
it
h
m
is
s
in
g
va
lu
e
s
w
oul
d
s
ke
w
th
e
r
e
s
ul
ts
of
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
.
T
he
a
tt
r
ib
ut
e
va
lu
e
s
in
th
e
s
oybe
a
n
da
t
a
s
e
t
w
e
r
e
c
ode
d
num
e
r
ic
a
ll
y
a
s
w
e
ll
,
a
n
R
pr
ogr
a
m
w
a
s
w
r
it
te
n
to
r
e
pl
a
c
e
th
e
s
e
num
e
r
ic
va
lu
e
s
w
it
h
th
e
ir
tr
ue
va
lu
e
s
.
T
he
s
oybe
a
n
d
a
ta
s
e
t
w
a
s
th
e
n
e
xa
m
in
e
d
f
or
m
is
s
in
g
va
lu
e
s
,
a
l
s
o
us
in
g
a
n
R
pr
ogr
a
m
.
T
he
r
e
w
e
r
e
ni
ne
a
tt
r
ib
ut
e
s
(
ha
il
,
s
e
ve
r
it
y,
s
e
e
d_t
m
t,
le
a
f
_m
il
d,
lo
dgi
ng,
s
hr
iv
e
li
ng,
f
r
ui
ti
ng_bodie
s
,
f
r
ui
t_
s
pot
s
,
s
e
e
d_di
s
c
ol
or
)
w
he
r
e
10%
or
m
or
e
of
th
e
va
lu
e
s
w
e
r
e
m
is
s
in
g.
A
de
c
i
s
io
n
w
a
s
m
a
de
to
r
un
e
a
c
h
of
th
e
s
ix
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
on
two
ve
r
s
io
ns
of
th
e
s
oyb
e
a
n
da
ta
s
e
t,
one
ve
r
s
io
n
w
it
h
a
ll
a
tt
r
ib
ut
e
s
in
c
lu
de
d
a
nd
a
not
he
r
ve
r
s
io
n
w
it
h
th
e
s
e
ni
ne
a
tt
r
ib
ut
e
s
r
e
m
ove
d.
T
he
pur
pos
e
of
c
r
e
a
ti
ng
th
e
s
e
two
v
e
r
s
io
ns
w
a
s
th
e
s
a
m
e
a
s
t
he
ju
s
ti
f
ic
a
ti
on
f
or
th
e
s
a
m
e
m
e
th
od
in
th
e
m
us
hr
oom
da
ta
s
e
t,
to
in
ve
s
ti
ga
te
w
he
th
e
r
or
not
th
e
s
e
a
tt
r
ib
ut
e
s
w
it
h
m
is
s
in
g
va
lu
e
s
w
oul
d
s
ke
w
th
e
r
e
s
ul
t
s
of
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
.
U
pon
f
ur
th
e
r
in
ve
s
ti
ga
ti
on
o
f
th
e
s
oybe
a
n
da
ta
s
e
t,
it
w
a
s
di
s
c
ove
r
e
d
th
a
t
th
e
r
e
e
xi
s
te
d
onl
y
one
da
ta
poi
nt
f
or
th
e
2
-
4
-
d
-
in
ju
r
y
c
la
s
s
a
nd
th
a
t
m
os
t
of
th
e
a
tt
r
ib
ut
e
va
lu
e
s
w
e
r
e
m
is
s
in
g
f
or
th
is
one
da
ta
poi
nt
.
T
hi
s
da
t
a
poi
nt
w
a
s
r
e
m
ove
d
f
r
om
bot
h
ve
r
s
io
ns
of
th
e
s
oybe
a
n
da
ta
s
e
ts
in
or
de
r
to
a
voi
d s
ke
w
in
g t
he
r
e
s
ul
t
s
of
t
he
c
la
s
s
if
ic
a
ti
on a
lg
or
it
hm
s
.
3.
2
.
C
la
s
s
if
i
c
at
io
n
m
e
t
h
od
s
S
ix
di
f
f
e
r
e
nt
c
la
s
s
if
ic
a
ti
on
te
c
hni
que
s
w
e
r
e
t
e
s
te
d
in
th
is
s
tu
dy
to
bui
ld
c
la
s
s
if
ic
a
ti
on
m
ode
ls
f
or
pr
e
di
c
ti
ng dis
e
a
s
e
s
i
n
s
oybe
a
n
s
a
nd e
di
bl
e
or
poi
s
onous
f
e
a
tu
r
e
s
of
m
us
hr
oom
s
. T
he
c
la
s
s
if
ic
a
ti
on a
lg
or
it
hm
s
w
e
r
e
a
ll
tr
a
in
e
d
u
s
in
g
10
-
f
ol
d
c
r
os
s
va
li
da
ti
on
a
nd
w
e
r
e
e
xe
c
ut
e
d
us
in
g
f
unc
ti
ons
in
W
e
k
a
[
15]
.
W
it
h
10
-
f
ol
d
c
r
os
s
va
li
da
ti
on,
th
e
r
ow
s
w
it
hi
n
th
e
da
ta
s
e
ts
a
r
e
r
a
ndoml
y
r
e
or
ga
ni
z
e
d
a
nd
s
pl
it
in
to
10
f
ol
ds
of
e
qua
l
s
iz
e
[
16]
.
W
it
h
e
a
c
h
it
e
r
a
ti
on
of
th
e
c
la
s
s
if
ic
a
ti
on
m
ode
l
tr
a
in
in
g
pr
oc
e
s
s
,
one
f
ol
d
is
us
e
d
a
s
th
e
t
e
s
t
da
ta
s
e
t
a
nd
th
e
r
e
m
a
in
in
g
9
f
ol
ds
a
r
e
us
e
d
a
s
th
e
tr
a
in
in
g
da
ta
s
e
ts
.
T
hi
s
pr
oc
e
s
s
r
e
pe
a
ts
10
ti
m
e
s
unt
il
e
a
c
h
f
ol
d
ha
s
be
e
n
us
e
d
a
s
th
e
te
s
t
d
a
ta
s
e
t.
T
he
r
e
s
ul
ti
ng
c
la
s
s
if
ic
a
ti
on
m
od
e
l
is
a
n
a
ve
r
a
ge
of
th
e
10
it
e
r
a
ti
on
s
of
th
e
tr
a
in
in
g
pr
oc
e
s
s
.
T
he
f
ol
lo
w
in
g 6 c
la
s
s
if
ic
a
ti
on me
th
ods
w
e
r
e
us
e
d i
n t
hi
s
s
tu
dy:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
10
, N
o.
1
,
M
a
r
c
h
20
2
1
:
25
7
–
26
4
260
3.
2
.
1.
A
r
t
if
ic
ia
l
n
e
u
r
al
n
e
t
w
or
k
A
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
ks
(
A
N
N
)
a
r
e
bui
lt
t
o
r
e
s
e
m
bl
e
t
he
w
a
y a
huma
n br
a
in
t
hi
nks
. A
N
N
s
c
ont
a
in
m
ul
ti
pl
e
w
e
ig
ht
e
d
c
onne
c
ti
ons
be
twe
e
n
in
put
s
a
nd
out
put
s
,
th
e
s
e
w
e
ig
ht
s
a
r
e
a
dj
us
te
d
w
he
n
bui
ld
in
g
th
e
m
ode
l
on
th
e
tr
a
in
in
g
da
t
a
in
or
de
r
to
c
or
r
e
c
tl
y
pr
e
di
c
t
c
la
s
s
la
b
e
ls
ba
s
e
d
on
th
e
in
put
da
t
a
obj
e
c
t
[
17]
.
I
n
th
is
s
tu
dy,
A
N
N
s
w
e
r
e
bui
lt
us
in
g
th
e
m
ul
ti
la
ye
r
pe
r
c
e
pt
r
on
a
lg
or
it
hm
in
W
e
ka
.
T
he
m
ul
ti
la
ye
r
pe
r
c
e
pt
r
on
a
lg
or
i
th
m
bui
ld
s
a
n
A
N
N
th
r
ough
a
pr
oc
e
s
s
c
a
ll
e
d
ba
c
kpr
opa
g
a
ti
on.
I
n
th
is
pr
oc
e
s
s
,
w
e
ig
ht
s
a
r
e
a
s
s
ig
ne
d
to
e
a
c
h
da
t
a
obj
e
c
t
in
th
e
in
put
la
ye
r
of
th
e
A
N
N
.
T
he
s
e
w
e
ig
ht
s
a
r
e
th
e
n
r
e
-
a
s
s
ig
ne
d
a
s
n
e
c
e
s
s
a
r
y
in
one
,
or
m
ul
ti
pl
e
,
hi
dde
n
la
ye
r
s
of
th
e
A
N
N
in
or
de
r
to
m
in
im
iz
e
th
e
m
e
a
n
s
qu
a
r
e
d
e
r
r
or
be
twe
e
n
th
e
c
la
s
s
la
be
l
pr
e
di
c
te
d
by
th
e
A
N
N
a
nd
th
e
tr
ue
c
la
s
s
l
a
be
l
of
th
e
gi
v
e
n
da
ta
obj
e
c
t.
T
h
e
pr
oc
e
s
s
i
s
c
a
ll
e
d
ba
c
kpr
opa
ga
ti
on
be
c
a
us
e
th
e
s
e
a
dj
us
tm
e
nt
s
to
th
e
w
e
ig
ht
s
a
r
e
done
in
th
e
ba
c
kw
a
r
ds
di
r
e
c
ti
on
s
ta
r
ti
ng
a
t
th
e
out
put
la
ye
r
,
w
hi
c
h
c
ont
a
in
s
th
e
c
la
s
s
la
be
l
s
,
a
nd
goi
ng
ba
c
k
th
r
ough
a
ll
of
t
he
hi
dde
n
la
ye
r
s
to
th
e
f
ir
s
t
hi
dde
n
la
ye
r
[
18]
.
T
he
m
ul
ti
la
ye
r
pe
r
c
e
pt
r
on
a
lg
or
it
hm
w
a
s
e
xe
c
ut
e
d
us
in
g
a
le
a
r
ni
ng
r
a
te
of
0.3,
a
m
om
e
nt
um
of
0.2,
a
nd
tr
a
in
in
g t
im
e
of
500.
3.
2
.
2.
N
aï
ve
b
aye
s
N
a
ïv
e
b
a
ye
s
is
a
pr
oba
bi
li
s
ti
c
c
la
s
s
if
ic
a
ti
on
m
e
th
od
th
a
t
us
e
s
ba
ye
s
th
e
or
e
m
.
T
h
e
na
ïv
e
ba
ye
s
c
la
s
s
if
ie
r
ta
ke
s
a
s
e
t
of
f
e
a
tu
r
e
s
f
r
om
a
da
ta
s
e
t
a
nd
de
te
r
m
in
e
s
th
e
pr
oba
bi
li
ty
of
e
a
c
h
f
e
a
tu
r
e
oc
c
ur
r
in
g
in
e
a
c
h
c
la
s
s
w
it
hi
n
th
e
da
ta
[
19]
.
F
o
r
e
a
c
h
r
ow
of
da
ta
,
th
e
va
l
ue
s
of
th
e
a
tt
r
ib
ut
e
s
a
r
e
us
e
d
to
c
a
lc
ul
a
te
th
e
pos
te
r
io
r
pr
oba
bi
li
ty
f
or
e
a
c
h
c
la
s
s
w
it
hi
n
th
e
da
t
a
s
e
t,
th
e
r
ow
of
da
ta
is
th
e
n
a
s
s
ig
ne
d
to
th
e
c
la
s
s
w
it
h
th
e
hi
ghe
s
t
pos
te
r
io
r
pr
oba
bi
li
ty
.
T
hi
s
m
e
th
od
is
r
e
f
e
r
r
e
d
to
a
s
na
ï
ve
be
c
a
us
e
it
a
s
s
um
e
s
th
a
t
a
ll
f
e
a
tu
r
e
s
of
th
e
da
ta
s
e
t
a
r
e
in
de
pe
nd
e
nt
of
one
a
not
he
r
,
w
hi
c
h
is
a
n
a
s
s
um
pt
io
n
th
a
t
is
li
ke
ly
unt
r
ue
a
nd
th
us
na
ïv
e
.
D
e
s
pi
te
th
is
a
s
s
um
pt
io
n
not
be
in
g
tr
ue
in
a
ll
c
a
s
e
s
,
na
ïv
e
ba
ye
s
h
a
s
be
e
n
s
how
n
to
be
a
s
uc
c
e
s
s
f
ul
c
la
s
s
if
ie
r
in
la
r
ge
da
ta
s
e
ts
.
T
h
e
na
ïv
e
ba
ye
s
a
lg
or
it
hm
w
a
s
e
x
e
c
ut
e
d
u
s
in
g
th
e
n
a
iv
e
ba
ye
s
c
la
s
s
if
ie
r
in
W
e
ka
.
T
h
e
na
ïv
e
ba
ye
s
c
la
s
s
if
ie
r
in
W
e
ka
us
e
s
e
s
ti
m
a
to
r
c
la
s
s
e
s
.
A
ba
tc
h
s
iz
e
of
100
w
a
s
us
e
d
w
it
hout
ke
r
ne
l
e
s
ti
m
a
ti
on
or
s
upe
r
vi
s
e
d di
s
c
r
e
ti
z
a
ti
on.
3.
2
.
3.
k
-
n
e
ar
e
s
t
n
e
ig
h
b
or
T
he
k
-
ne
a
r
e
s
t
ne
ig
hbor
(
K
N
N
)
a
lg
or
it
hm
a
s
s
ig
ns
c
la
s
s
la
be
ls
t
o
r
ow
s
w
it
hi
n
a
da
ta
s
e
t
ba
s
e
d
on
th
e
c
la
s
s
l
a
be
l
s
of
t
r
a
in
in
g da
ta
t
ha
t
a
r
e
s
im
il
a
r
[
17
]
. T
he
K
N
N
a
lg
or
it
hm
w
or
ks
by s
e
a
r
c
hi
ng t
he
t
r
a
in
in
g da
ta
f
o
r
k t
r
a
in
in
g t
upl
e
s
t
ha
t
a
r
e
c
lo
s
e
s
t
to
t
he
t
e
s
t
da
ta
t
upl
e
a
nd
a
s
s
ig
ns
t
he
t
e
s
t
tu
pl
e
a
c
la
s
s
l
a
be
l
ba
s
e
d on the
c
la
s
s
la
be
ls
of
th
os
e
c
lo
s
e
s
t
tr
a
in
in
g
tu
pl
e
s
.
T
he
c
lo
s
e
ne
s
s
of
a
tr
a
in
in
g
tu
pl
e
to
a
te
s
t
tu
pl
e
is
de
te
r
m
in
e
d
u
s
in
g
a
di
s
ta
nc
e
f
unc
ti
on,
s
uc
h
a
s
E
uc
li
de
a
n
di
s
ta
nc
e
.
K
N
N
w
a
s
im
pl
e
m
e
nt
e
d
in
W
e
k
a
f
or
th
is
e
xpe
r
im
e
nt
u
s
in
g
th
e
in
s
ta
nc
e
ba
s
e
d
le
a
r
ne
r
(
I
B
K
)
a
lg
or
it
hm
.
T
he
I
B
K
a
lg
or
it
hm
w
a
s
e
xe
c
ut
e
d
us
in
g
th
e
E
uc
li
de
a
n
di
s
ta
nc
e
f
unc
ti
on, a
ba
tc
h s
iz
e
of
100, a
nd k
=
1.
3.
2
.
4.
S
u
p
p
or
t
ve
c
t
or
m
ac
h
in
e
S
uppor
t
ve
c
to
r
m
a
c
hi
ne
(
S
V
M
)
is
a
s
upe
r
vi
s
e
d
m
a
c
hi
ne
le
a
r
ni
ng
a
lg
or
it
hm
us
e
d
in
c
la
s
s
if
ic
a
ti
on
a
nd
r
e
gr
e
s
s
io
n.
S
V
M
s
w
e
r
e
f
ir
s
t
pr
e
s
e
nt
e
d
by
V
la
di
m
i
r
V
a
p
ni
k
a
nd
hi
s
c
ow
or
ke
r
s
,
B
e
r
nha
r
d
B
os
e
r
a
nd
I
s
a
be
ll
e
G
uyun,
a
t
th
e
c
om
put
a
ti
ona
l
le
a
r
ni
ng
th
e
or
y
(
C
O
L
T
-
9
2)
c
onf
e
r
e
nc
e
[
20
]
.
I
n
th
is
a
lg
or
it
hm
,
tr
a
in
in
g
da
ta
is
tr
a
ns
f
or
m
e
d
to
a
hi
ghe
r
di
m
e
ns
io
n.
A
li
ne
or
hype
r
pl
a
ne
s
e
pa
r
a
te
s
th
e
c
la
s
s
e
s
of
d
a
ta
f
r
om
e
a
c
h
ot
h
e
r
.
T
he
li
ne
or
hype
r
pl
a
ne
a
r
e
f
ound
us
in
g
s
uppor
t
ve
c
to
r
s
.
S
uppor
t
ve
c
to
r
s
a
r
e
th
e
poi
nt
s
c
lo
s
e
s
t
to
th
e
hype
r
pl
a
ne
.
S
V
M
s
a
r
e
hi
ghl
y
a
c
c
ur
a
te
,
w
hi
c
h
m
a
ke
s
up
f
or
th
e
s
lo
w
s
pe
e
d
a
s
s
oc
ia
te
d
w
it
h
th
e
m
.
I
n
th
is
s
tu
dy,
S
V
M
s
w
e
r
e
bui
lt
us
in
g
th
e
s
e
que
nt
ia
l
m
in
im
a
l
opt
im
i
z
a
ti
on
(
S
M
O
)
a
lg
or
it
hm
in
W
e
ka
.
T
he
S
M
O
a
lg
or
it
hm
us
e
s
th
e
c
om
pl
e
xi
ty
p
a
r
a
m
e
te
r
,
a
l
s
o
known
a
s
th
e
C
pa
r
a
m
e
te
r
,
to
c
ont
r
ol
th
e
f
le
xi
bi
li
ty
of
th
e
pr
oc
e
s
s
in
dr
a
w
in
g
th
e
li
ne
b
e
twe
e
n
c
la
s
s
e
s
[
21]
;
th
e
C
pa
r
a
m
e
te
r
us
e
d
w
a
s
1.0.
T
he
P
ol
yK
e
r
ne
l
de
f
a
ul
t
w
a
s
us
e
d, w
hi
c
h s
e
pa
r
a
te
s
t
he
c
la
s
s
e
s
by
a
c
ur
ve
d l
in
e
[
21]
.
3.
2
.
5.
D
e
c
is
io
n
t
r
e
e
A
de
c
is
io
n
tr
e
e
is
a
s
tr
uc
tu
r
e
th
a
t
c
ont
a
in
s
in
te
r
na
l
node
s
th
a
t
d
e
not
e
a
tt
r
ib
ut
e
s
,
br
a
nc
he
s
th
a
t
de
not
e
th
e
out
c
om
e
of
a
te
s
t
on
a
n
obs
e
r
va
ti
on
a
nd
le
a
f
node
s
th
a
t
de
not
e
th
e
c
la
s
s
la
be
l
[
17]
.
T
he
to
p
node
of
th
is
tr
e
e
-
li
ke
s
tr
uc
tu
r
e
is
th
e
r
oot
node
.
I
n
or
de
r
to
de
te
r
m
in
e
th
e
c
la
s
s
of
a
n
obs
e
r
va
t
io
n,
th
e
de
c
i
s
io
n
tr
e
e
is
f
ol
lo
w
e
d,
s
ta
r
ti
ng
a
t
th
e
r
oo
t,
m
ovi
ng
dow
n
to
th
e
le
a
f
node
s
.
T
he
de
c
is
io
n
tr
e
e
a
lg
or
it
hm
w
a
s
im
pl
e
m
e
nt
e
d
in
W
e
ka
f
or
th
is
s
tu
dy
us
in
g
th
e
J
48
de
c
is
io
n
tr
e
e
a
lg
or
it
h
m
.
T
he
J
48
a
lg
or
it
hm
w
a
s
e
xe
c
ut
e
d
us
in
g
a
ba
tc
h
s
iz
e
of
100,
th
e
m
in
im
a
l
of
obj
e
c
t
s
of
2,
w
it
hout
u
s
in
g
unpr
un
e
d
tr
e
e
s
,
a
c
onf
id
e
nc
e
in
te
r
va
l
of
0.25,
s
ubt
r
e
e
r
a
is
in
g a
nd w
it
hout
bi
na
r
y s
pl
it
s
[
22]
.
3.2.
6
. R
an
d
om
f
or
e
s
t
A
r
a
ndom
f
or
e
s
t
is
a
c
ol
le
c
ti
on
of
de
c
i
s
io
n
tr
e
e
s
.
E
a
c
h
de
c
is
io
n
tr
e
e
w
it
hi
n
th
e
r
a
ndom
f
or
e
s
t
ge
n
e
r
a
te
s
a
c
la
s
s
pr
e
di
c
ti
on;
th
e
c
la
s
s
w
it
h
th
e
la
r
ge
s
t
num
be
r
be
c
om
e
s
th
e
pr
e
di
c
ti
on
of
th
e
r
a
ndom
f
or
e
s
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
P
la
nt
di
s
e
as
e
pr
e
di
c
ti
on u
s
in
g c
la
s
s
if
ic
at
io
n al
gor
it
hm
s
(
M
ar
ia
M
or
gan
)
261
[
23
]
.
I
n
or
de
r
f
or
th
is
a
lg
or
it
hm
to
be
e
f
f
ic
ie
nt
,
th
e
in
di
vi
dua
l
m
ode
ls
m
us
t
not
be
c
or
r
e
la
te
d
or
s
houl
d
h
a
ve
a
lo
w
c
or
r
e
la
ti
on.
T
he
r
e
a
r
e
two
m
e
th
ods
u
s
e
d
to
e
n
s
ur
e
th
a
t
th
e
in
di
vi
dua
l
de
c
is
io
n
tr
e
e
m
ode
ls
a
r
e
not
to
o
c
lo
s
e
ly
c
or
r
e
la
te
d
w
it
h
e
a
c
h
ot
he
r
.
O
n
e
m
e
th
od
is
b
a
ggi
ng.
E
a
c
h
in
di
vi
dua
l
tr
e
e
s
e
le
c
t
s
a
r
a
ndom
s
a
m
pl
e
f
r
om
th
e
da
ta
s
e
t
w
it
h
r
e
pl
a
c
e
m
e
nt
[
23]
.
T
he
s
e
c
ond
m
e
th
od
is
r
a
ndom
li
ne
a
r
c
om
bi
na
t
io
ns
of
th
e
a
tt
r
ib
ut
e
s
.
T
hi
s
m
e
th
od
us
e
s
ne
w
a
tt
r
ib
ut
e
s
th
a
t
a
r
e
a
li
ne
a
r
c
om
bi
na
ti
on
of
th
e
e
xi
s
ti
ng
a
tt
r
ib
ut
e
s
[
17
]
.
T
hi
s
a
ls
o
he
lp
s
to
r
e
duc
e
c
or
r
e
la
ti
on
be
twe
e
n
c
la
s
s
if
ie
r
s
.
T
h
e
r
a
ndom
f
or
e
s
t
a
l
gor
it
hm
in
W
e
ka
w
a
s
us
e
d
in
th
is
s
tu
dy.
T
he
r
a
ndom
f
or
e
s
t
a
lg
or
ig
hm
us
e
s
th
e
num
F
e
a
tu
r
e
s
va
lu
e
of
0,
w
hi
c
h
s
e
le
c
ts
th
e
num
be
r
of
a
tt
r
ib
ut
e
s
c
ons
id
e
r
e
d
a
t
e
a
c
h s
pl
it
poi
nt
. T
he
a
lg
or
it
hm
w
a
s
e
xe
c
ut
e
d w
it
h a
ba
g
s
iz
e
pe
r
c
e
nt
of
100%
, w
hi
c
h c
r
e
a
te
s
a
ne
w
r
a
ndom
s
a
m
pl
e
th
e
s
a
m
e
s
i
z
e
a
s
th
e
tr
a
in
in
g
s
a
m
pl
e
.
T
he
N
um
I
te
r
a
ti
ons
va
lu
e
w
a
s
100,
w
hi
c
h
s
e
ts
th
e
num
be
r
of
ba
gs
or
i
te
r
a
ti
ons
t
o 100.
3.
3
.
P
e
r
f
or
m
an
c
e
e
val
u
at
io
n
s
T
he
f
ol
lo
w
in
g
s
e
ve
n
m
e
a
s
ur
e
s
w
e
r
e
us
e
d
to
e
va
lu
a
t
e
th
e
pe
r
f
or
m
a
nc
e
of
th
e
s
ix
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
on
th
e
s
oybe
a
n
a
nd
m
us
hr
oom
da
ta
s
e
ts
,
th
e
s
e
m
e
a
s
ur
e
s
w
e
r
e
s
e
le
c
te
d
b
a
s
e
d
on
th
e
ir
us
e
in
a
s
im
il
a
r
s
tu
dy
w
hi
c
h
us
e
d
c
la
s
s
if
ic
a
ti
on
f
unc
ti
ons
in
W
e
k
a
f
or
pl
a
nt
di
s
e
a
s
e
de
te
c
ti
on
on
a
da
ta
s
e
t
of
pl
a
nt
im
a
ge
s
[
4]
.
A
c
c
ur
a
c
y
:
A
pe
r
c
e
nt
a
ge
c
a
lc
ul
a
te
d
by
di
vi
di
ng
th
e
num
be
r
of
c
or
r
e
c
tl
y
c
la
s
s
if
ie
d
da
ta
poi
n
ts
by
th
e
to
ta
l
num
be
r
of
da
ta
poi
nt
s
a
nd mul
ti
pl
yi
ng by 100.
M
e
a
n a
bs
ol
ut
e
e
r
r
or
:
T
he
m
e
a
n a
bs
ol
ut
e
e
r
r
or
(
M
A
E
)
i
s
c
a
lc
ul
a
te
d by ta
ki
ng t
he
s
um
of
t
he
a
bs
ol
ut
e
e
r
r
or
s
di
vi
de
d by the
numbe
r
of
non
-
m
is
s
in
g da
ta
poi
nt
s
.
T
r
ue
pos
it
iv
e
r
a
te
:
T
h
e
T
P
r
a
te
is
c
a
l
c
ul
a
te
d
by
di
vi
di
ng
th
e
n
um
be
r
of
tr
ue
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
by
th
e
s
um
of
th
e
num
be
r
of
tr
ue
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
a
nd
th
e
num
be
r
of
f
a
ls
e
ne
ga
ti
ve
c
la
s
s
if
ic
a
ti
ons
.
(
T
P
/(
T
P
+
F
N
)
)
.
F
a
ls
e
pos
it
iv
e
r
a
te
:
T
he
F
P
r
a
te
is
c
a
lc
ul
a
te
d
by
di
vi
di
ng
th
e
n
um
be
r
of
f
a
ls
e
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
by
th
e
s
um
of
th
e
num
be
r
of
f
a
ls
e
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
a
nd
th
e
num
be
r
o
f
t
r
ue
ne
ga
ti
ve
obs
e
r
va
ti
ons
.
(
F
P
/(
F
P
+
T
N
)
)
.
P
r
e
c
is
io
n:
P
r
e
c
is
io
n
is
c
a
lc
ul
a
t
e
d
by
di
vi
di
ng
th
e
num
be
r
of
tr
ue
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
by
th
e
s
um
o
f
t
he
numbe
r
of
t
r
ue
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
a
nd t
he
numbe
r
of
f
a
ls
e
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
. (
T
P
/(
T
P
+
F
P
)
)
.
R
e
c
a
ll
:
R
e
c
a
ll
is
c
a
lc
ul
a
te
d
by
di
vi
di
ng
th
e
num
be
r
of
tr
ue
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
by
th
e
s
um
of
th
e
num
be
r
of
t
r
ue
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
a
nd
th
e
numb
e
r
of
f
a
ls
e
ne
ga
ti
ve
c
la
s
s
if
ic
a
ti
ons
. (
T
P
/(
T
P
+
F
N
)
)
.
F
-
M
e
a
s
ur
e
:
T
he
F
-
M
e
a
s
ur
e
is
c
a
lc
ul
a
t
e
d
by
m
ul
ti
pl
yi
ng
th
e
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
,
di
vi
di
ng
th
is
va
lu
e
by
th
e
s
um
of
th
e
pr
e
c
i
s
io
n
a
nd
r
e
c
a
ll
,
a
nd
f
in
a
ll
y
m
ul
ti
pl
yi
ng
th
is
num
be
r
by
two.
(
2*(
(
pr
e
c
is
io
n*r
e
c
a
ll
)
/(
pr
e
c
is
io
n+
r
e
c
a
ll
)
)
)
.
4.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
T
he
a
lg
or
it
hm
s
w
e
r
e
te
s
te
d
on
bot
h
va
r
ia
ti
ons
of
th
e
m
us
hr
oo
m
da
ta
s
e
t,
one
w
it
h
th
e
a
tt
r
ib
ut
e
w
it
h
m
is
s
in
g
va
lu
e
s
r
e
m
ove
d
a
nd
one
w
it
h
a
ll
a
tt
r
ib
ut
e
s
in
c
lu
de
d.
T
he
m
e
a
s
ur
e
s
pr
e
vi
ous
ly
de
s
c
r
ib
e
d
w
e
r
e
r
e
por
te
d
f
or
e
a
c
h
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
.
T
he
r
e
s
ul
ts
f
or
bo
th
va
r
ia
ti
ons
of
th
e
m
us
hr
oom
da
ta
s
e
t
w
e
r
e
s
im
il
a
r
,
s
o
th
e
r
e
por
te
d
r
e
s
ul
ts
a
r
e
f
r
om
th
e
ve
r
s
io
n
of
th
e
d
a
ta
s
e
t
w
it
h
a
ll
a
tt
r
ib
ut
e
s
in
c
lu
d
e
d.
A
ll
of
th
e
s
ix
a
lg
or
it
hm
s
te
s
te
d
pe
r
f
or
m
e
d
e
xt
r
e
m
e
ly
w
e
ll
on
th
e
m
u
s
hr
oom
da
ta
s
e
t
w
it
h
a
lm
o
s
t
a
ll
a
c
c
ur
a
c
y
va
lu
e
s
a
t
100%
.
T
he
na
ïv
e
ba
ye
s
a
lg
or
it
hm
pe
r
f
or
m
e
d
th
e
w
or
s
t
on
th
is
da
ta
s
e
t
w
it
h
a
n
a
c
c
ur
a
c
y
of
95.83%
,
w
hi
c
h
is
s
ti
ll
a
good a
c
c
ur
a
c
y l
e
v
e
l.
T
a
bl
e
1
s
how
s
f
or
r
e
s
ul
ts
f
or
t
he
m
u
s
hr
oom
da
ta
s
e
t.
T
a
bl
e
1
.
R
e
s
ul
ts
f
or
m
us
hr
oom
da
ta
s
e
t
P
a
r
a
m
e
t
e
r
C
l
a
s
s
i
f
i
c
a
t
i
on M
e
t
hod
ANN
N
a
ï
ve
B
a
ye
s
KNN
S
V
M
D
e
c
i
s
i
on T
r
e
e
R
a
ndom
F
or
e
s
t
A
c
c
ua
r
c
y
100.00%
95.83%
100.00%
100.00%
100.00%
100.00%
M
A
E
0.00
0.04
0.00
0.00
0.00
0.00
TP
-
R
a
t
e
1.00
0.96
1.00
1.00
1.00
1.00
FP
-
R
a
t
e
0.00
0.04
0.00
0.00
0.00
0.00
P
r
e
c
i
s
i
on
1.00
0.96
1.00
1.00
1.00
1.00
R
e
c
a
l
l
1.00
0.96
1.00
1.00
1.00
1.00
F
-
M
e
a
s
ur
e
1.00
0.96
1.00
1.00
1.00
1.00
T
he
a
lg
or
it
hm
s
w
e
r
e
te
s
te
d
on
two
va
r
ia
ti
ons
of
th
e
s
oybe
a
n
da
ta
s
e
t,
one
w
it
h
a
ny
a
tt
r
ib
ut
e
s
th
a
t
c
ont
a
in
e
d
10%
or
m
or
e
m
is
s
in
g
va
lu
e
s
r
e
m
ove
d
a
nd
one
w
it
h
a
ll
a
tt
r
ib
ut
e
s
in
c
lu
de
d.
T
he
r
e
s
ul
ts
f
or
bot
h
va
r
ia
ti
ons
of
th
e
s
oybe
a
n
da
t
a
s
e
t
w
e
r
e
s
im
il
a
r
,
s
o
th
e
r
e
por
te
d
r
e
s
ul
ts
a
r
e
f
r
om
th
e
ve
r
s
io
n
of
th
e
da
ta
s
e
t
w
it
h
a
ll
a
tt
r
ib
ut
e
s
in
c
lu
de
d.
A
ll
of
th
e
s
ix
a
lg
or
it
hm
s
,
e
x
c
e
pt
f
or
de
c
is
io
n
tr
e
e
,
p
e
r
f
or
m
e
d
w
e
ll
on
th
e
s
oybe
a
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
10
, N
o.
1
,
M
a
r
c
h
20
2
1
:
25
7
–
26
4
262
da
ta
s
e
t
w
it
h
a
c
c
ur
a
c
y
va
lu
e
s
f
a
ll
in
g
in
th
e
r
a
nge
of
89.22
-
91.83
%
.
T
he
d
e
c
is
io
n
tr
e
e
a
lg
o
r
it
hm
ha
d
a
r
e
por
te
d
a
c
c
ur
a
c
y
of
82.68%
f
or
th
e
s
oybe
a
n
da
t
a
s
e
t,
s
o
th
is
w
a
s
th
e
w
or
s
t
-
pe
r
f
or
m
in
g
a
lg
or
it
hm
in
c
la
s
s
if
yi
ng
di
s
e
a
s
e
s
i
n t
he
s
oyb
e
a
n da
ta
s
e
t.
T
a
bl
e
2
s
how
s
f
or
r
e
s
ul
ts
f
or
t
he
s
oybe
a
n da
t
a
s
e
t.
T
a
bl
e
2
.
R
e
s
ul
ts
f
or
s
oybe
a
n d
a
ta
s
e
t
P
a
r
a
m
e
t
e
r
C
l
a
s
s
i
f
i
c
a
t
i
on M
e
t
hod
ANN
N
a
ï
ve
B
a
ye
s
KNN
S
V
M
D
e
c
i
s
i
on T
r
e
e
R
a
ndom
F
or
e
s
t
A
c
c
ua
r
c
y
91.18%
90.20%
91.83%
89.22%
82.68%
89.54%
M
A
E
0.01
0.01
0.01
0.1
0.02
0.04
TP
-
R
a
t
e
0.91
0.9
0.92
0.89
0.83
0.9
FP
-
R
a
t
e
0.91
0.01
0.01
0.01
0.02
0.01
P
r
e
c
i
s
i
on
0.91
0.92
0.92
0.89
0.82
0.9
R
e
c
a
l
l
0.91
0.9
0.92
0.89
0.83
0.9
F
-
M
e
a
s
ur
e
0.91
0.9
0.92
0.89
0.83
0.89
I
n
c
om
pa
r
in
g
th
e
r
e
s
ul
ts
f
or
bo
th
da
ta
s
e
ts
,
A
N
N
a
nd
K
N
N
pe
r
f
or
m
e
d
be
s
t
on
th
e
s
oybe
a
n
da
ta
s
e
t,
w
hi
le
a
ll
m
e
th
ods
ot
he
r
th
a
n
na
ïv
e
b
a
ye
s
p
e
r
f
or
m
e
d
a
t
100%
a
c
c
ur
a
c
y
on
th
e
m
u
s
hr
oom
da
ta
s
e
t.
I
t
is
to
b
e
e
xpe
c
te
d
th
a
t
m
os
t
of
th
e
c
la
s
s
if
ic
a
ti
on
m
e
th
ods
w
oul
d
p
e
r
f
or
m
be
s
t
on
th
e
m
us
hr
oom
da
ta
s
e
t
be
c
a
us
e
th
e
r
e
a
r
e
onl
y
two
c
la
s
s
e
s
pr
e
s
e
nt
in
th
is
da
ta
s
e
t,
w
hi
le
th
e
s
oybe
a
n
da
ta
s
e
t
th
a
t
w
a
s
te
s
te
d
ha
s
18
c
la
s
s
e
s
.
A
s
m
e
nt
io
ne
d
in
th
e
li
te
r
a
tu
r
e
r
e
vi
e
w
s
e
c
ti
on,
na
ïv
e
ba
ye
s
a
nd
K
N
N
a
r
e
not
ty
pi
c
a
ll
y
u
s
e
d
in
a
gr
ic
ul
tu
r
a
l
s
tu
di
e
s
.
T
hi
s
is
a
n
in
te
r
e
s
ti
ng
poi
nt
to
c
ons
id
e
r
be
c
a
us
e
n
a
ïv
e
ba
ye
s
w
a
s
th
e
onl
y
a
lg
or
it
hm
th
a
t
di
d
not
pr
oduc
e
100%
a
c
c
ur
a
c
y
in
th
e
m
u
s
hr
oom
da
ta
s
e
t,
but
a
ls
o
in
te
r
e
s
ti
n
g
to
not
e
be
c
a
u
s
e
K
N
N
w
a
s
one
of
th
e
b
e
s
t
pe
r
f
or
m
in
g a
lg
or
it
hm
s
i
n t
he
s
oybe
a
n
da
ta
s
e
t
[
11]
. A
lt
hough
th
e
pe
r
f
or
m
a
nc
e
of
K
N
N
on
th
e
s
oybe
a
n da
ta
s
e
t
s
e
e
m
to
c
onf
li
c
t
w
it
h
th
e
pr
e
vi
ous
li
te
r
a
tu
r
e
,
th
e
r
e
s
ul
ts
s
how
th
a
t
A
N
N
w
a
s
one
of
th
e
to
p
pe
r
f
or
m
in
g
a
lg
or
it
hm
s
in
th
e
s
oybe
a
n
da
ta
s
e
t
c
om
pa
r
e
d
to
th
e
ot
he
r
a
lg
or
it
hm
s
,
a
nd
th
is
c
onf
ir
m
s
w
ha
t
w
a
s
f
ound
in
th
e
F
e
br
ua
r
y 2019 s
tu
dy me
nt
io
ne
d i
n t
he
l
it
e
r
a
tu
r
e
r
e
vi
e
w
s
e
c
ti
on
[
2]
.
F
ur
th
e
r
c
om
pa
r
is
on
of
th
e
r
e
s
ul
ts
f
or
bot
h
da
ta
s
e
ts
r
e
ve
a
le
d
a
not
he
r
m
a
jo
r
di
f
f
e
r
e
nc
e
be
twe
e
n
th
e
two
da
ta
s
e
t
s
.
T
he
m
us
hr
oom
da
ta
s
e
t
is
ba
la
n
c
e
d,
w
it
h
th
e
ob
s
e
r
va
t
io
ns
be
in
g
e
qua
ll
y
di
s
tr
ib
ut
e
d
a
m
ong
th
e
two
c
la
s
s
e
s
,
poi
s
onou
s
a
nd
e
di
bl
e
.
T
he
ba
la
nc
e
d
di
s
tr
ib
ut
io
n
of
th
e
m
us
hr
oom
da
ta
s
e
t
is
s
how
n
in
F
ig
ur
e
1.
T
he
s
oybe
a
n
da
ta
s
e
t,
how
e
ve
r
,
is
im
ba
la
nc
e
d
a
m
ong
th
e
di
s
e
a
s
e
c
la
s
s
e
s
.
A
s
s
how
n
in
F
ig
ur
e
2,
th
e
r
e
a
r
e
4
c
la
s
s
e
s
th
a
t
c
ont
a
in
a
m
uc
h
hi
ghe
r
pe
r
c
e
nt
a
ge
of
th
e
obs
e
r
va
ti
o
ns
c
om
pa
r
e
d
to
th
e
ot
he
r
c
la
s
s
e
s
.
T
he
di
s
e
a
s
e
c
la
s
s
e
s
w
it
h
th
is
hi
gh
pe
r
c
e
nt
a
ge
of
obs
e
r
va
ti
ons
,
in
F
ig
ur
e
2
(
D
1
th
r
ough
D
4)
,
a
r
e
phyt
opht
ho
r
a
-
r
ot
,
b
r
ow
n
-
s
pot
,
a
lt
e
r
na
r
ia
le
a
f
-
s
pot
,
a
nd
f
r
og
-
e
ye
-
le
af
-
s
pot
.
B
e
c
a
us
e
th
e
s
oybe
a
n
da
ta
s
e
t
is
im
ba
la
nc
e
d,
m
e
a
s
ur
e
s
ot
he
r
th
a
n
a
c
c
ur
a
c
y
ne
e
de
d
to
be
c
on
s
id
e
r
e
d
to
de
te
r
m
in
e
if
th
e
im
ba
la
nc
e
of
th
e
da
ta
s
e
t
w
a
s
s
ke
w
in
g
th
e
r
e
s
ul
ts
f
or
e
a
c
h
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
.
I
n
th
e
c
a
s
e
of
im
ba
la
nc
e
d
da
ta
s
e
ts
w
it
h
a
la
r
ge
num
b
e
r
of
va
lu
e
s
,
th
e
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
va
lu
e
s
c
a
n
be
e
va
lu
a
te
d
to
de
te
r
m
in
e
th
e
pe
r
f
or
m
a
nc
e
of
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
[
24]
.
A
s
s
how
n
in
th
e
r
e
s
ul
ts
f
or
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
t
e
s
te
d
on
th
e
s
oybe
a
n
da
ta
s
e
t
in
T
a
bl
e
2,
a
ll
of
th
e
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
v
a
lu
e
s
a
r
e
c
lo
s
e
to
1.
R
e
f
e
r
r
in
g
ba
c
k
to
th
e
pa
r
a
m
e
t
e
r
e
va
lu
a
ti
ons
s
e
c
ti
on
of
th
is
pa
pe
r
,
th
is
in
di
c
a
te
s
th
a
t
th
e
num
be
r
of
tr
ue
pos
it
iv
e
c
la
s
s
if
ic
a
t
io
ns
a
r
e
m
uc
h
l
a
r
ge
r
th
a
n
th
e
num
be
r
of
f
a
l
s
e
ne
ga
ti
ve
a
nd
f
a
ls
e
pos
it
iv
e
c
la
s
s
if
ic
a
ti
ons
.
I
f
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
w
e
r
e
b
e
in
g
s
ke
w
e
d
by
th
e
im
ba
la
nc
e
in
th
e
da
ta
s
e
t,
our
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
va
lu
e
s
w
oul
d be
m
uc
h
lo
w
e
r
.
T
hu
s
,
a
lt
hough
th
i
s
is
a
m
a
jo
r
di
f
f
e
r
e
nc
e
be
twe
e
n
our
da
ta
s
e
ts
,
th
e
im
ba
la
nc
e
f
e
a
tu
r
e
of
th
e
s
oybe
a
n
da
ta
s
e
t
di
d
not
ha
ve
a
n
a
dve
r
s
e
e
f
f
e
c
t
on
th
e
r
e
s
ul
ts
of
e
a
c
h
c
la
s
s
if
ic
a
ti
on a
lg
or
it
hm
.
I
n
f
ut
ur
e
s
tu
di
e
s
of
th
is
ki
nd,
da
ta
s
e
ts
th
a
t
a
r
e
im
ba
la
nc
e
d
m
a
y
ne
e
d
a
ddi
ti
ona
l
d
a
ta
pr
e
pa
r
a
ti
on
t
e
c
hni
que
s
a
s
to
not
s
ke
w
th
e
r
e
s
ul
ts
of
th
e
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
.
T
w
o
pot
e
nt
ia
l
te
c
hni
q
ue
s
f
or
ha
ndl
in
g
im
ba
la
nc
e
d
da
ta
s
e
ts
a
r
e
ove
r
s
a
m
pl
in
g
a
nd
unde
r
s
a
m
pl
in
g
[
25]
.
I
n
ove
r
s
a
m
pl
in
g,
s
ynt
he
ti
c
d
a
ta
is
ge
ne
r
a
te
d s
o
th
a
t
a
ddi
ti
ona
l
obs
e
r
va
ti
ons
a
r
e
pr
e
s
e
nt
in
th
e
m
in
or
it
y
c
la
s
s
a
nd
th
e
di
s
tr
ib
ut
io
n
of
th
e
da
ta
is
m
or
e
e
qua
l
a
m
ong
th
e
c
la
s
s
e
s
.
I
n
unde
r
s
a
m
pl
in
g,
obs
e
r
va
ti
ons
a
r
e
r
e
m
ove
d
f
r
om
t
he
m
a
jo
r
it
y c
la
s
s
e
s
t
o m
a
ke
t
he
di
s
tr
ib
ut
io
n of
da
ta
m
or
e
e
qua
l
a
m
ong the
c
la
s
s
e
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
P
la
nt
di
s
e
as
e
pr
e
di
c
ti
on u
s
in
g c
la
s
s
if
ic
at
io
n al
gor
it
hm
s
(
M
ar
ia
M
or
gan
)
263
F
ig
ur
e
1.
M
us
hr
oom
da
ta
s
e
t
c
la
s
s
di
s
tr
ib
ut
io
n
F
ig
ur
e
2. S
oybe
a
n da
ta
s
e
t
c
la
s
s
di
s
tr
ib
ut
io
n
5.
C
O
N
C
L
U
S
I
O
N
I
n
th
is
pa
pe
r
,
w
e
te
s
te
d
A
N
N
,
na
ïv
e
ba
ye
s
,
K
N
N
,
S
V
M
,
de
c
is
i
on
tr
e
e
,
a
nd
r
a
ndom
f
or
e
s
t
c
la
s
s
if
ie
r
s
to
pr
e
di
c
t
di
s
e
a
s
e
pr
e
s
e
n
c
e
in
a
m
u
s
hr
oom
da
ta
s
e
t
a
nd
c
la
s
s
if
y di
s
e
a
s
e
in
a
s
oyb
e
a
n
d
a
ta
s
e
t.
I
n
th
e
m
us
hr
oom
da
ta
s
e
t,
w
e
f
ound
th
a
t
a
ll
c
la
s
s
if
ie
r
s
,
e
xc
e
pt
f
o
r
na
ïv
e
ba
ye
s
,
pe
r
f
or
m
e
d
a
t
100%
a
c
c
ur
a
c
y.
T
hi
s
is
a
li
ke
ly
r
e
s
ul
t
gi
ve
n
a
da
ta
s
e
t
w
it
h
onl
y
two
c
la
s
s
e
s
.
I
n
th
e
s
oybe
a
n
d
a
t
a
s
e
t,
w
e
h
a
ve
s
how
n
th
a
t
A
N
N
a
nd
K
N
N
a
r
e
th
e
be
s
t
c
la
s
s
if
ie
r
s
in
t
e
r
m
s
of
a
c
c
ur
a
c
y,
but
th
a
t
A
N
N
is
li
ke
l
y
th
e
be
tt
e
r
c
hoi
c
e
s
in
c
e
K
N
N
c
la
s
s
if
ic
a
ti
on
is
not
ty
pi
c
a
ll
y
us
e
d
f
or
pl
a
nt
da
ta
s
e
t
s
.
W
e
a
l
s
o
s
how
e
d
th
a
t
th
e
i
m
ba
la
nc
e
of
th
e
s
oybe
a
n
da
ta
s
e
t
di
d
not
a
f
f
e
c
t
th
e
r
e
s
ul
ts
of
th
e
c
la
s
s
if
ic
a
ti
on
m
e
th
ods
,
li
ke
ly
be
c
a
us
e
a
la
r
ge
a
m
ount
of
da
ta
is
pr
e
s
e
nt
.
I
n
th
e
m
us
hr
oom
da
ta
s
e
t,
w
e
us
e
d
c
la
s
s
if
ic
a
ti
on
to
de
te
r
m
in
e
if
a
di
s
e
a
s
e
w
a
s
pr
e
s
e
nt
or
not
(
e
di
bl
e
or
poi
s
onous
)
a
nd
in
th
e
s
oybe
a
n
da
ta
s
e
t,
w
e
us
e
d
c
la
s
s
if
ic
a
ti
on
to
de
te
r
m
in
e
w
hi
c
h
di
s
e
a
s
e
w
a
s
pr
e
s
e
nt
.
T
he
pur
pos
e
of
th
e
s
e
e
xpe
r
im
e
nt
s
w
a
s
to
c
om
e
up
w
it
h
c
la
s
s
if
ic
a
ti
on
m
e
th
od
s
th
a
t
c
a
n
be
us
e
d
on
da
t
a
s
e
t
s
f
or
pl
a
nt
s
or
f
ungi
th
a
t
c
ont
a
in
r
e
a
l
m
e
a
s
ur
e
m
e
nt
s
in
s
te
a
d
of
im
a
g
e
s
.
T
he
f
in
di
ngs
in
th
is
pa
pe
r
c
a
n
b
e
r
e
pe
a
te
d
on
s
im
il
a
r
f
ungi
or
pl
a
nt
da
ta
s
e
ts
but
m
a
y
a
l
s
o
be
e
xt
e
nd
e
d
to
tr
a
in
in
g
c
la
s
s
if
ic
a
ti
on
a
lg
or
it
hm
s
f
or
pr
e
di
c
ti
ng
di
s
e
a
s
e
pr
e
s
e
nc
e
or
di
s
e
a
s
e
c
l
a
s
s
if
ic
a
ti
on i
n huma
n or
a
ni
m
a
l
da
ta
s
e
ts
w
it
h r
a
w
m
e
a
s
ur
e
m
e
nt
s
.
R
E
F
E
R
E
N
C
E
S
[1]
G.
Prem,
et
al
,
“Plant
Disease
Prediction
Using
Machine
Learning
Al
gorithms,”
Internati
onal
Journal
of
Computer
Applications
, vol. 182, no. 25, pp. 1
–
7
,
2018.
DOI: 10.5120/ij
ca2018918049.
[2]
N.
Kanaka
Durga
&
G.
Anurhada
,
“Plant
Disease
Identification
Using
SVM
and
ANN
Algorithms,”
Internati
onal
Journal
of Recent
Technol
ogy and
Engineeri
ng (IJRTE
)
, vol. 7, no. 5S4
, 2019.
[3]
H.
Al
-
Hiary,
et
al
.
,
“Fast
and
Accurat
e
Detection
and
Classification
of
Plant
Diseases,”
Internati
onal
Journal
of
Computer
Application
s,
vol 17, no. 1, 2011.
DOI: 10.5120/218
3
-
2754.
[4]
R.
Ramya,
et
al.
,
“A
Review
of
Different
Classification
Techniques
in
Machine
Learning
Using
Weka
for
Plant
Dise
ase Detecti
on,”
Internati
onal Resear
ch Journal
of Engin
eering and
Technology (IR
JET),
vol 5, no.5
, 2018.
[5]
G.
H.
Lincoff,
The
Audubon
Society
field
guide
to
North
American
mushrooms
.
Alfred
A.
Knopf;
distributed
by
Random Ho
use
, 1981.
0
500
1000
1500
2000
2500
3000
3500
4000
4500
P
oi
s
onous
E
di
bl
e
N
um
be
r
of
i
ns
t
a
nc
e
s
C
l
a
ss
l
a
be
l
0
5
10
15
20
25
30
35
40
45
N
um
be
r
of
i
ns
t
a
nc
e
s
C
l
a
ss
l
a
be
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
10
, N
o.
1
,
M
a
r
c
h
20
2
1
:
25
7
–
26
4
264
[6]
G.
H.
Lincoff
,
“
UCI
Mac
hine
Learning
Repository
,”
University
of
California,
Irvine,
School
of
Information
and
Computer Sc
iences
, 1981. [Online]. Available: https://archive.ics.uci.edu/ml/datasets/mushroom.
[7]
North
Carolina
Soybeans
,
“
Uses
of
Soybeans
,”
2019.
[Online]
.
Avail
able:
https://ncsoy.org/media
-
resources/uses
-
of
-
soybeans/
.
[8]
L.
V.
Madden,
G.
Hughes,
and
M.
E.
Irwin,
“Coupling
Disease
-
Pro
gress
-
Curve
and
Time
-
of
-
Infection
Function
s
for
Predicting
Yi
eld
Loss
of
Crops
,
”
Phytopath
ology,
vol
90,
no.
8,
2000.
https://doi.org/10.1094/PHYTO.2000.90.8.788.
[9]
W.
Pavan,
C.
W.
Fraisse,
an
N.
A.
Peres,
“Development
of
a
web
-
based
disease
forecasting
system
for
strawberrie
s,”
Computers
and Elec
tronics in
Agriculutu
re,
vol 75, no. 1, 2011.
[10]
R.N Strang
e and P.
R. Scott, “
Plant Di
sease:
A Thre
at to Glob
al Food
Security,
”
Annual Review of Phytopathology,
vol. 43, no.1, pp. 83
-
116, 2005.
[11]
D.
C.
Corrales,
J.
C.
Corrales
and
A.
Figueroa
-
Casas
,
“Towards
det
ecting
crop
disease
s
and
pest
by
supervised
learning,
” Ing.
Univ.
, vol. 19, no. 1, pp. 207
-
228.
[12]
U.
Ayub
and
S.
A.
Moqurrab
,
“Predicti
ng
crop
diseases
using
data
mining
approaches:
Classi
fication,
”
2018
1st
Internati
onal Con
ference on Pow
er, Energy an
d Smart Gr
id (ICPESG)
, 20
18.
[13]
D.
Dua,
and
C.
Graff,
“
UCI
Machine
Learning
Repository
,”
Universit
y
of
California,
Irvine,
School
of
Information
and Com
puter Sci
ences
, 2019. [Online]. Available:
http://archive.ics.uci.edu/ml
[14]
R.S.
Michalski
and
R.L.
Chilausky
,
“
UCI
Machine
Learning
Reposit
ory
,”
University
of
California,
Irvine,
Schoo
l
of
Information
and
Computer
Sciences
,
1980.
[Online].
Available:
https://archi
ve.ics.uci.edu/ml/datasets/Soybean+
(Large)
[15]
E.
Frank,
M.
A.
Hall,
and
I.
H.
Witten
,
“
The
WEKA
Workbench.
Online
Appendix
for
Data
Mining:
Practical
Machine
Learning
Tools
and
Techniques
,
"
Data
Mining:
Practical
Machine
Learning
Tools
and
Techniques,
Morgan
Kaufmann,
Fourth
Edition,
2016.
[Online].
Available
:
xxxxxhttps://www.cs.waikato.ac.nz/ml/weka/Witten_et_al_2016_appe
ndix.pdf
[16]
R.
Bouckae
rt,
WEKA
Manual
for
Version
3
-
7
-
8
,
2013.
[Online].
Available:
http://statweb.stanford.edu/~
lpekelis/13_datafest_cart/Weka
Manual
-
3
-
7
-
8.pdf
[17]
J.
Han,
M.
Kamber
and
J.
Pei,
Data
Mining:
Concepts
and
Techniques.
Amsterdam:
Elsevier
/Morgan
Kaufmann
,
2012.
[18]
R.
Rojas R
, “
The
Backpro
pagation
Algorithm
,”
Neural Networks. Sprin
ger, Berlin, Heidelberg
, 1996.
[19]
D.
Soni
,
“
Introductio
n
to
Naive
Bayes
Classifi
cation
,”
Towards
Data
Science
,
16
July
2019.
[Online].
Available
:
https://towardsdatascience.com/introduction
-
to
-
naive
-
bayes
-
clas
sification
-
4cffabb1ae54
[20]
R.
Gandhi,
“
Support
Vecto
r
Machin
e
-
Introductio
n
to
Machine
Lea
rning
Algorithm
s
,”
Towards
Data
Science
,
7
June
2018.
[Online].
Available:
https://towardsdatascience.com/supp
ort
-
vector
-
machine
-
introduction
-
to
-
machine
-
learning
-
algori
thms
-
934a444fca47
[21]
J. Brownlee,
How To Use Classi
fication Machine Learning
Algorithm
s in Weka
, 25 July 2016. [Online]. Available
:
https://machinelearningmastery.com/use
-
classifi
cation
-
machine
-
learning
-
algorit
hms
-
weka/
[22]
V.
Koblar,
Study
Programme:
Information
and
Communicatio
n
Technolo
gies
,
2012.
[Online].
Available:
https://pdfs.semanticscholar.org/94a2/9d5a74ed9ac656de
4e55c71fac92d07795ef.pdf
[23]
T.
Yiu
,
“
Understanding
Random
Forest
,”
Towards
Data
Scie
nce
,
12
June
2019
.
[Online].
Available:
https://towardsdatascience.com/understanding
-
ran
dom
-
forest
-
58381e0602d2
[24]
S.
Lador,
“
What metri
cs shoul
d be used
for evalu
ating a
model o
n an im
balanced dat
a set?
,”
Towards Data Science
,
5
Sep
2017.
[Online].
Available:
https://towardsdatascience.com/what
-
metrics
-
should
-
we
-
use
-
on
-
imbalance
d
-
data
-
set
-
preci
sion
-
recall
-
roc
-
e2e79252ae
[25]
W.
B
ard,
“
Having
an
Imbalanced
Dataset?
Here
Is
How
You
Can
F
ix
It
,”
Towards
Data
Science
,
22
Feb
2019.
[Online].
Available:
https:/
/towardsdatas
cience.com/havi
ng
-
an
-
imba
lanced
-
dataset
-
here
-
is
-
how
-
you
-
can
-
solve
-
it
-
1640568947e
Evaluation Warning : The document was created with Spire.PDF for Python.