Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
12
,
No.
3
,
Decem
ber
201
8
, p
p.
1273
~
1281
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v1
2
.i
3
.pp
1273
-
1281
1273
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
Decision
Making
i
n the T
ea L
eaves
Diseases
Detecti
on
Using
Mamda
ni Fu
zzy Infe
re
n
ce M
ethod
Arif
Ridh
o
Lu
bis
1
, Sant
i
Pra
yu
d
an
i
2
, Mu
h
arm
an
Lu
bis
3
,
A
l
-
Kh
owariz
mi
4
1, 2, 4
Depa
rtment of Computer En
gine
er
ing
and
In
form
at
ic
s,
Polit
e
knik
Nege
r
i
Me
dan
,
Jala
n
Alm
a
m
ater
No.
1
,
20155
,
Meda
n,
Nor
th
S
um
at
era
,
Indone
sia
3
School
of
Indus
tri
al E
ng
ineeri
ng
,
T
el
kom
Univer
sit
y
Jal
an
T
eleko
m
unika
si,
No.
1,
Bandung, 4025
7,
Indone
si
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
pr
30
, 201
8
Re
vised
Ju
l
14
,
201
8
Accepte
d
Aug
2
1
, 201
8
The
t
ea
p
la
n
ts
(Camell
i
a
Sinens
is)
are
sm
al
l
tree
spec
i
es
th
at
us
e
leave
s
and
le
af
buds
to
pro
duce
te
a
har
v
est
ed
through
a
m
onocul
tur
e
s
y
s
tem
.
It
is
an
agr
ic
u
lt
ure
p
ract
ic
e
to
cultiv
at
e
one
t
y
pes
of
cr
op
or
li
vesto
ck,
var
iet
y
or
bre
ed
on
a
f
arm
annua
l
l
y
.
More
over
,
th
e
emerg
enc
e
of
p
ests,
pa
thoge
ns
an
d
disea
ses
c
ause
serious
damage
s
to
tea
pl
ant
s
sig
nifi
c
ant
l
y
to
i
ts
produc
ti
v
i
t
y
and
qualit
y
to
op
ti
m
um
wors
t.
All
par
ts
of
the
te
a
pla
nt
such
as
l
eaves,
stems
,
roots,
flowe
rs
a
nd
fruit
s
a
re
ex
pos
ed
to
th
ese
har
m
le
ad
to
lo
ss
of
y
i
el
d
7
unti
l
10%
per
yea
r.
Th
e
in
te
nsit
y
of
th
ese
at
t
ac
k
s
var
y
gre
at
l
y
o
n
par
t
ic
ul
ar
cl
imat
e,
the
deg
ree
slop
e
and
th
e
pl
ant
m
a
te
ri
al
used.
The
r
efo
re
,
th
is
st
u
d
y
ana
l
y
z
es
t
ea
le
a
ves
as
a
comm
o
n
par
t
used
in
r
ec
ip
es
to
cr
ea
te
unique
t
ast
e
and
fl
avor
in
t
ea
produ
ct
ion
,
espe
cially
in
a
gro
-
industr
y
.
T
he
de
ci
sio
n
m
aki
ng
m
et
hod
used
is
Fuzz
y
Mam
dani
Infe
re
nce
as
one
of
m
odel
with
func
ti
on
al
hi
era
r
ch
y
wi
th
initial
input
base
d
on
esta
bli
shed
cri
t
eri
a
.
Fuz
z
y
logi
c
wil
l
provid
e
tol
er
ance
to
th
e
set
of
val
ue
,
so
tha
t
sm
al
l
cha
n
ges
will
not
result
in
signif
icant
ca
t
egor
y
d
iff
ere
nc
es,
on
l
y
a
ff
ec
t
the
m
embership
le
v
el
on
the
var
ia
bl
e
val
u
e.
Previous
m
et
h
od
using
proba
bil
ities
hav
e
show
n
78%
tea
le
av
es
have
bee
n
at
t
ac
k
ed
b
y
ca
t
e
gor
y
C
(Gra
y
Bl
ight
)
whil
e
using
Mam
dani
indi
c
at
ed
86%
of
te
a
l
ea
v
es
have
be
en
infect
ed.
In
thi
s
c
ase
,
thi
s
result
point
ed
out
th
a
t
Fuzz
y
Mam
d
ani
Infe
r
enc
es
have
m
ore
optim
al
result
compare
to
th
e
p
rev
ious method
.
Ke
yw
or
d
s
:
Decisi
on m
aking
Fu
zzy
lo
gic
Infer
e
nce m
et
h
od
Mam
dan
i
Copyright
©
201
8
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Ar
if
Ri
dho Lu
bis
,
Dep
a
rtm
ent o
f C
om
pu
t
er E
ng
i
neer
i
ng and
Inf
or
m
at
ic
s,
Po
li
te
kn
i
k Nege
ri Meda
n,
Jal
an Alm
a
m
ater
No.
1,
2015
5,
Me
da
n
, N
or
t
h
S
um
at
era,
Indonesia.
Em
a
il
:
arifr
idho@
polm
ed.
ac.id
1.
INTROD
U
CTION
To
day
the
devel
op
m
ent
of
I
T
has
bee
n
s
o
rap
i
d,
not
lim
it
ed
to
the
de
ve
lop
m
ent
of
ha
rdwar
e
an
d
so
ft
war
e
,
but
a
lso
the
c
om
pu
t
ing
m
et
ho
d
su
c
h
as
t
he
decisi
on
-
m
aking
sys
tem
.
It
is
a
bra
nch
of
sci
e
nce th
at
it
s
con
ce
pt
li
es
an
d
us
e
d
f
re
qu
e
nt
ly
between
IS
and
AI
fiel
d.
The
abili
ty
to
qu
ic
kly
decide
on
par
ti
cula
r
them
base
d
on
c
onte
xt,
ri
gh
t
on
ta
r
get
an
d
acc
ount
able
is
the
crit
ic
al
factor
t
o
be
su
ccess
in
th
e
global
com
pe
ti
ti
on
.
Hav
i
ng
a
lo
t
of
inf
orm
ation
i
s
not
suffici
ent
,
if
not
a
ble
to
do
decisi
on
-
m
akin
g
prop
e
rly
su
c
h
as
al
te
r
na
ti
ve
identific
at
ion,
weig
ht
the
ev
idence
a
nd
re
viewin
g
sc
he
m
e.
Pr
ior
to
choose
am
on
g
al
te
rn
at
ives,
sever
al
crit
erions
s
houl
d
be
est
a
blish
ed
first,
in
w
hich
it
s
ho
ul
d
be
able
t
o
a
nswer
im
po
rtant
qu
e
sti
on
a
bout
ho
w
well
an
al
te
r
native
can
s
olv
e
a
prob
le
m
at
hand
.
U
sin
g
ste
p
-
by
-
ste
p
decisi
on
-
m
aking
syst
e
m
can
help
t
o
m
ake
m
or
e
inform
ed
assessm
ent
by
orga
nizing
rel
evan
t
i
nfor
m
ation
i
n
pro
pe
r
s
tructu
re
a
nd
syst
e
m
at
ic
a
ll
y.
In
the
end, the or
ga
nizat
ion
often st
r
uggle from
the af
te
r
ef
fect o
f
an
incide
nt wh
ic
h
m
a
y creat
e d
am
age an
d ha
rm
to
the r
e
puta
ti
on
,
fina
nces a
nd worker
p
e
rfo
rm
a
nces
[13].
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1273
–
1281
1274
Tea
is
a
bev
era
ge
m
ade
by
s
te
epin
g
le
aves
in
bo
il
ing
wate
r,
wh
ic
h
low
-
gr
own
te
as
are
pro
du
ce
f
ro
m
0
-
600m
,
m
id
-
gro
wn
f
r
om
600
-
1,2
00
m
wh
il
e
the
high
-
gro
wn
te
as
a
re
c
ul
ti
vated
bet
we
en
1,200
-
2,0
00
m
[1
]
.
Me
anwhil
e,
th
e
te
a
plant
need
s
a
ho
t,
m
oist
cl
i
m
at
e
with
i
ts
sp
eci
fic
requirem
ents
are
rangin
g
te
m
pe
ratur
e
s
from
10
-
30oC,
m
ini
m
u
m
annual
p
recipit
at
io
n
of
1250
m
m
,
prefe
rab
ly
aci
dic
s
oils
an
d
i
deall
y
0.5
-
10
de
gr
ee
slop
es
[
2].
M
ost
of
the
te
a
w
or
l
d
producti
on
base
d
on
F
A
O
2013’s
re
por
t
is
pro
du
c
ed
(
thousa
nd
to
ns
)
in
Fa
r
East
(3,96
5.6/7
8.31%),
w
hich
m
ajo
rity
is
in
China
(Mai
la
nd)
,
I
nd
ia
, S
rila
n
ka
,
Viet
nam
and
I
nd
onesi
a,
then
i
n
Africa
(64
9.5/12.82%
)
that
is
m
os
tl
y
in
Ken
ya
an
d
U
ganda,
f
ollo
w
ed
by
Near
E
ast
(2
53.
5/5%
)
,
Lat
in
Ca
ribb
ea
n
(95
/2%),
Ja
pa
n
(
84.7
/
1.6%)
a
nd
Ocea
nia
(
6.5/0.1%
)
[
2].
Fu
rt
her
m
or
e,
m
any
of
the
pests,
path
og
e
ns
an
d
diseases
tha
t
aff
ect
t
he
te
a
plan,
in
w
hich
it
was
rec
ord
ed
t
hat
10
34
s
pecies
of
art
hro
po
d
(m
os
tl
y
fr
om
bu
tt
erf
li
es
a
nd
m
oth
),
82
s
pec
ie
s
of
nem
at
ode
or
par
asi
te
,
m
or
e
than
400
of
f
ungal
an
d
sever
a
l
sp
eci
es
of
bact
eria
and
al
gae
[3
]
.
T
he
three
m
os
t
seriou
s
te
a
plant
d
ise
as
e
are
cam
e
ll
ia
diebac
k
an
d
ca
nk
e
r,
flo
wer
li
ght
and
root
r
ot
[4
]
.
The
biggest
chall
enge
f
or
t
ea
grow
e
rs
nowad
ay
s
,
is
to
pro
du
ce
te
a
w
it
ho
ut
pestic
ide
resid
ues
by
gro
wing
te
a
in
env
i
ronm
ents
le
ss
fav
ora
ble
to
pest
s
and
diseases,
wh
ic
h
m
ay
no
t
be
al
ways
po
ssi
ble
[5
]
.
Alte
rn
at
ively
,
it
cou
ld
be
throu
gh
c
hoos
i
ng
pe
sti
ci
des
with
low
interfer
e
nce
or
by
app
ly
in
g
pestic
ides
accor
di
ng
to
econom
ic
t
hr
es
hold
[
3]
,
[
5].
The
qual
it
y
of
le
aves,
buds,
fruit
,
flo
wers,
an
d
so
il
are
vi
br
a
nt
to
gro
w
in
wh
ic
h
their
aro
m
as
and
t
ast
es
are
i
m
pacted
by
the
rainf
al
ls
pe
rce
ntage,
con
ce
ntrati
on
of seco
ndary
m
et
abo
li
te
com
po
unds
a
nd fa
rm
er r
esi
li
ence m
anag
em
ent
[6
]
.
Pr
e
vious
r
esea
rch
[
7]
has
un
der
ta
ken
resea
rch
i
n
diag
nos
ing
le
a
f
diseas
es
an
d
te
a
pes
ts
by
usi
ng
Naïve
Ba
ye
sia
n
pro
ba
bili
ti
es
an
d
Ba
c
kw
a
r
d
C
haining
t
o
pro
du
ce
accu
r
at
e
scor
e
s.
Th
us
,
this
pap
e
r
wan
t
t
o
exp
l
or
e
f
ur
the
r
the
diag
nose
process
by
c
onduct
ing
ot
he
r
A
I
te
ch
nique,
w
hich
is
f
uzzy
lo
gic
syst
e
m
by
Mam
dan
i
Me
thod
(Min
-
Ma
x
I
nf
e
ren
ce
).
A
ct
ually
,
this
te
chn
i
qu
e
is
qu
i
te
popula
r
am
ong
aca
dem
ic
i
an
i
n
wh
ic
h
m
any
researc
hes
ha
d
bee
n
done
in
var
i
ou
s
the
m
e
and
disp
a
rate
top
ic
s
suc
h
as
pr
e
dicti
on
of
represe
ntati
ve
defor
m
at
ion
m
odulus
of
tu
nnel
s,
dam
s
and
m
ining
str
uctu
re
[
8],
m
od
el
ing
of
wate
r
m
o
vem
ent
in
no
n
-
sat
ur
at
e
d
s
oil
[9
]
,
cl
as
sific
at
ion
of
t
oddle
r
nutrit
io
na
l
sta
tus
[
10
]
,
m
od
el
li
ng
aut
o
z
oo
m
functi
on
i
n
dig
it
al
ca
m
era
[11],
delim
it
at
i
on
of
r
ur
al
an
d
urba
n
areas
on
a
dv
a
nce
d
carto
gr
a
ph
ic
vis
ualiz
at
ion
[
12
]
,
eve
n
sp
eci
fic
m
ic
ro
con
t
ro
ll
er
of
ARM
Cortex
-
M4
STM3
2F4
07V
G
[38]
an
d
m
any
m
or
e.
The
de
velo
pme
nt
from
the
scratc
h
or
ad
ding
s
om
e
functi
on
to
the
e
xisti
ng
software
will
co
ll
ide
with
the
fast
dev
el
opm
ent
of
so
ft
war
e
capa
bi
li
t
ie
s
in
softw
are
in
dustry,
w
hich
i
ncr
ease
s
ign
ific
a
ntly
thr
ough
ti
m
e
to
tim
e
as
the
re
sul
t
of
ti
gh
t com
p
et
it
i
on
i
n
the m
arket
[
19]
,
[27
-
28]
. I
n
t
he
co
nte
xt o
f
a
war
e
ness
, pers
on sho
uld
hav
e t
he
ada
pt
abili
t
y
as
their
pr
e
pa
r
at
ion
to
fit
wit
h
occurri
ng
ch
ang
e
s
or
unex
pe
ct
ed
ci
rc
um
st
ances
[21]
,
[30
-
32]
.
T
he
a
dv
a
ntages
for
us
in
g
fu
zz
y
log
ic
c
on
t
ro
l
(F
LC
)
is
beca
us
e
t
he
underst
and
a
bili
ty
,
flex
ibil
it
y,
m
od
el
lin
g
of
non
-
li
ne
ar
a
nd
com
plex
funct
ion
,
de
velo
pme
nt
of
ex
pe
rt
syst
e
m
,
colla
bo
rati
on
with
c
onve
ntion
al
c
ontr
ol
te
chn
i
ques
an
d
base
d
e
xclusive
ly
o
n natu
ral l
angua
ge
2.
LIT
ERATUR
E REVIE
W
The
pa
per
w
ork
by
Za
de
h
[
14
]
on
f
uzzy
al
gorithm
s
pr
e
sent
the
ne
w
con
ce
pt
of
f
orm
ula
ti
ng
the
con
t
ro
l
al
gorithm
through
lo
gical
r
ules
in
volvin
g
a
se
ries
of
fu
zzy
co
nd
it
ion
al
sta
te
m
e
nt,
wh
ic
h
sta
te
d
t
hat
if
a
set
of
c
ondit
ion
s
a
re
f
ulfill
ed
the
n
a
set
of
c
onseq
ue
nc
es
can
be
in
fe
rr
e
d.
In
FLC
te
rm
ino
log
y,
a
fu
zzy
con
t
ro
l
r
ule
pl
ay
the
ro
le
of
a
ntecede
nt,
w
hi
ch
is
a
conditi
on
i
n
the
ap
plica
ti
on
dom
ai
n
and
t
he
co
ns
e
quent
is
the
co
ntr
ol
act
ion
f
or
a
c
on
t
rol
le
d
syst
e
m
.
The
in
pu
ts
of
a
syst
e
m
based
on
f
uzzy
ru
le
s
m
us
t
be
pr
ovi
de
d
by
the
f
uzzy
se
ts
for
a
dju
stm
ent
in
fu
zzi
ficat
ion
of
the
c
risp
inputs
a
nd
vic
e
ve
rsa
for
t
he
outp
uts.
N
orm
al
l
y,
fu
zzy
l
og
ic
is
us
e
d
to
t
ra
ns
la
te
qua
ntit
ie
s
ex
pr
e
ssed
in
li
nguisti
cs,
su
c
h
a
s
vehi
cl
e
sp
eed
de
cl
ared
cat
egorical
ly
, w
hic
h
are slo
w
, q
uic
k,
f
a
st an
d
tur
bo. I
t c
an a
lso b
e use
d
to
p
r
ov
i
de
the d
e
gr
ee to
wh
ic
h a value
is
true
an
d
false.
Th
us
,
t
his
appr
oach
has
adv
a
ntage
s
ov
er
the
re
su
lt
s
relat
ed
to
hum
an
cogniti
ve
trai
ts,
especial
ly
in
sit
uations
that
in
vo
l
ves
th
e
f
orm
at
ion
of
c
on
c
ept,
the
rec
ogni
ti
on
of p
at
te
rn
an
d
decisi
on
-
m
aki
ng
in unce
rtai
n or
un
cl
ea
r
e
nv
i
ronm
ents [
20]
.
In
ge
ner
al
,
the
FLC
syst
em
consi
sts
of
f
ou
r
m
ai
n
par
ts
nam
el
y
fu
zzi
ficat
ion
inte
rf
ace
,
fu
z
zy
r
ule
base,
f
uzzy
in
fer
e
nce
e
ng
i
ne
an
d
defuzzifi
cat
ion
inte
rf
ac
e.
Ma
m
dan
i
and
As
sil
ia
n
[
15]
intr
oduce
d
n
ew
te
chn
iq
ue
i
n
f
uz
zy
infer
e
nce,
as
the
their
firs
t
appro
ac
h
to
c
on
t
ro
l
the
c
ombinati
on
of
ste
a
m
eng
ine
a
nd
bo
il
e
r
m
achines
by
s
ynthesiz
in
g
a
set
of
li
nguisti
c
con
tr
ol
r
ules
der
ive
d
f
r
om
exp
e
rience
d
hu
m
an
operat
or
s
.
I
n
add
it
io
n,
the
li
nguisti
c
va
riab
le
w
ho
se
va
lu
es
are
ex
press
ed
i
n
sta
te
m
ent,
w
ord
or
se
nt
ences
i
n
t
he
form
of
conditi
ons
in
natu
ral
la
ngua
ge
was
de
fine
d
by
s
uitable
m
e
m
ber
sh
ip
f
un
ct
io
n
(MF)
,
wh
ic
h
is
a
cu
rv
e
t
hat
determ
ines
how
each
point
is
m
app
ed
in
the
input
sp
ac
e
to
a
m
e
m
be
rsh
i
p
value
be
twe
en
0
a
nd
1
[22].
Howe
ver,
so
m
e
people
try
to
fin
d
phil
oso
phic
al
ans
wer
on
the
fun
dam
e
ntal
quest
ion
on
w
hy
a
syst
e
m
based
on
fuzzy
ru
le
s
work
well
f
or
a
wide
range
of
pract
ic
al
pro
blem
s.
The
firs
t
at
tem
pt
to
answ
er
t
his
quest
ion
is
qu
a
ntit
at
ively
dem
on
s
trat
ed
by
W
an
g
an
d
Buckley
[16
-
17
]
w
her
e
the
y
fo
und
ou
t
that
a
par
ti
cula
r
FLC
syst
e
m
s
cl
ass
is
un
ive
rsal
ap
pro
xim
a
tors,
wh
ic
h
has
ce
rt
ai
n
capa
bili
ty
of
a
ppr
oach
i
ng
any
real
c
onti
nu
ous
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Decisi
on M
aking in t
he
Te
a L
eaves
Disea
ses
D
et
ect
io
n U
sing M
amdani F
uz
zy
In
fe
re
nce
…
(
Arif
Ridho Lu
bis
)
1275
functi
on
in
a
com
pact
set
with
an
ar
bitrar
y
pr
eci
sio
n
an
d
accu
racy.
It
is
ch
aracte
riz
ed
by
the
Ga
us
sia
n
m
e
m
ber
sh
ip
f
unct
ions,
fu
zzy
conj
un
ct
io
n an
d
im
plica
ti
on
prod
uct and ce
nt
er ar
ea
of
defu
zzi
ficat
ion
.
The
f
uzzifica
ti
on
i
nterf
ace
i
nvol
ves
a
f
un
ct
i
on
s
t
o
m
easur
e
the
values
of
i
nput
va
riables,
creati
ng
a
scal
e
m
app
ings
that
transf
er
s
the
ran
ge
of
input
value
s
va
riables
to
the
corres
pondin
g
unive
rse
of
disc
ourse
and
c
onduct
a
fu
zzi
ficat
io
n
functi
on
that
trans
f
or
m
it
into
ade
quat
e
li
ng
uisti
c
val
ue
to
be
disp
la
ye
d
as
a
certai
n
f
uzzy
s
et
.
The
ru
le
ba
se
relat
ed
t
o
th
e
knowle
dge
f
or
a
ppl
ic
at
ion
do
m
ai
n
an
d
th
e
co
ncurr
e
nt
c
on
t
ro
l
obj
ect
ive
.
It
c
onsist
s
of
a
data
base
a
nd
a
li
nguisti
c
c
on
tr
ol
ru
le
base,
w
hi
ch
prov
i
des
im
portant
def
i
niti
on
to
unde
rstan
d
t
he
instr
uctio
n
lo
gic
a
nd
data
m
anip
ulati
on
in
a
FLC,
w
hich
char
act
e
rizes
t
he
c
ontrol
goa
ls
an
d
po
li
cy
of
the
dom
ai
n
exp
erts
[18].
In
oth
e
r
hand,
the
f
uzz
y
infer
ence
e
ngine
is
the
core
of
a
FLC,
w
hich
ha
s
the
abili
ty
to
s
i
m
ulate
the
m
e
chan
ism
in
hu
m
an
decisi
on
-
m
aking
t
hro
ugh
fu
zzy
c
once
pts
by
infe
rr
i
ng
fu
zzy
con
t
ro
l
act
io
ns
that
us
e
f
uzz
y
i
m
plic
ation
and
the
of
inf
eren
ce
r
ules.
S
ince
sev
eral
li
nguisti
c
va
riab
le
s
ar
e
involve
d
i
n
the
antece
den
ts
an
d
t
he
c
on
cl
us
io
ns
of
a
r
ule,
t
he
f
uzzy
syst
em
is
cal
le
d
a
s
th
e
m
ulti
–
inp
ut
m
ul
ti
–
ou
t
pu
t t
ype
[1
8]
.
Mam
dan
i
m
eth
od
is
m
os
t
of
te
n
cat
eg
or
i
zed
as
a
f
orm
of
ap
prox
i
m
at
e
reaso
ning,
w
hic
h
ha
s
been
cal
le
d
as
the
process
that
al
lo
w
po
s
s
ible
im
pr
eci
se
co
nclusi
on
is
infe
rr
e
d
from
inade
quat
e
c
ol
le
ct
ion
pr
em
ise
[23].
This
cl
assifi
ca
ti
on
a
nd
a
set
of
IF
-
THE
N
ru
le
s
can
easi
ly
m
ake
so
m
eo
ne
belie
ve
t
ha
t
this
appr
oach
can
pro
vid
e
the
l
og
i
cal
i
m
plica
ti
on
of
a
set
of
ru
l
es
that
a
re
us
e
d
for
it
s
c
onstr
uction,
al
th
ough
on
ly
appr
ox
im
at
ely.
I
n
par
ti
cular
,
So
m
e
al
so
ar
gued
that
it
m
i
gh
t
be
te
m
pte
d
to
belie
ve
i
n
this
m
et
ho
d
as
the
sh
are
d
trut
h
of the p
rem
ise
s
ensures
the
trut
h
of
the
co
nclu
sion
s
[
24
-
26
]
.
I
n
lo
gics
that
r
ecognize
d
the
l
evel o
f
par
ti
al
truth,
it
is
exp
ect
ed
th
at
if
the
inp
uts
of
the
syst
em
is
true
up
to
ce
rtai
n
point,
the
n
the
outp
uts
of
the
syst
e
m
m
us
t
also
co
rresp
ond
to
the
sam
e
lev
el
of
certai
nt
y.
More
ov
e
r,
t
he
an
te
ce
de
nts
and
t
he
co
ns
e
qu
e
nts
can
al
so
be
c
om
bin
ed
pro
po
s
it
ion
s
that
incl
ud
e
t
he
lo
gical
connecti
on
of
AND
or
OR.
The
inter
pret
at
ion
of
Mam
dan
i
m
et
h
od
as
m
oto
r
f
or
in
fer
e
nce
pre
serv
at
io
n
t
ru
t
h
is
certai
nly
cha
ll
eng
e
by
the
e
vid
e
nce
a
nd
it
is
not
sh
are
d
by
e
ve
ry
academ
ic
ian
,
but
sti
ll
it
is
reasonably
well
-
kn
own
f
or
to
the
point
it
per
m
ea
te
s
m
any
si
m
ulati
on
app
li
cat
ion
s [2
4].
3.
RESEA
R
CH MET
HO
DOL
OGY
This
researc
h
m
et
ho
d
co
ns
is
ts
of
se
ver
al
s
ta
ges
nam
ely
l
it
eratur
e
st
ud
y
,
data
colle
ct
ion,
pro
blem
analy
sis
an
d
pro
blem
so
lving.
This
st
ud
y
ha
s
obj
ect
ive
to
i
m
ple
m
ent
Fu
zzy
Mam
dan
i
m
et
ho
d
t
o
dete
rm
ine
the
ty
pe
of
dis
ease
an
d
pest
i
n
te
a
le
aves
ba
sed
on
se
ve
ral
ind
ic
at
ors
or
s
ym
pto
m
s.
The
ty
pe
of
disease
us
ed
in
this
resea
r
ch
a
re
li
m
i
te
d
to
Tea
Bl
ist
er
Bl
ig
ht
(E
xoba
sidium
ve
xans),
Lea
f
S
po
t
(Cyl
indro
cl
adium
il
ic
ic
ola),
Gr
a
y
Bl
igh
t
(P
est
al
otia
theae),
Re
d
Ro
ot
Rot
(G
a
node
rm
a
ps
e
oduf
e
rr
e
um),
Bl
ack
Ro
ot
Rot
(Ros
el
li
nia
arc
uata)
a
nd
the
t
ype
of
pests
na
m
el
y
Tea
Mosq
uito
Bu
g
(
Hel
op
el
ti
s
s
pp.)
,
T
ea
To
rtrix
(Ho
m
on
a
coffeari
a)
a
nd
Walke
r
(
Hyp
osi
dr
a
ta
la
ca)
.
T
her
e
are
ste
p
by
ste
p
to
do
prob
le
m
analy
sis
in
this
stu
dy,
wh
ic
h
are
decidin
g
t
he
fuzzy
va
riabl
e,
dete
rm
ining
the
f
uzzy
set
and
dom
ai
n,
f
uzzifica
ti
on
to
dev
el
op
m
e
m
b
ersh
i
p
functi
on
an
d
countin
g
the
val
ue,
creati
ng
f
uz
zy
ru
le
,
cond
ucting
in
fer
e
nt
ia
l
syst
e
m
by
c
al
culat
ing
α
-
pr
edicat
e
in
eac
h
ru
le
by
MI
N
im
plica
ti
on
a
nd
usi
ng
MA
X
m
et
hods
f
or
co
m
po
sit
ion
al
f
or
al
l
r
ules,
la
stl
y
defuzzifi
cat
io
n wit
h
ce
ntr
oid
m
et
ho
ds.
4.
DISCU
SSI
ON A
ND R
ES
UL
TS
In
F
LC,
in
pu
t
var
ia
bles
us
e
d
are
inte
rv
al
s
,
so
the
i
nput
in
the
f
or
m
of
a
str
ic
t
nu
m
ber
m
us
t
be
conve
rted
int
o
fu
zzy
nu
m
ber
s
as
can
be
see
n
in
the
pr
e
viou
s
T
able
1
that
sp
eci
fy
the
wei
gh
t
disease
a
nd
pe
st
(
WD
C)
an
d
w
ei
gh
t
in
dicat
or
(
WI
)
.
T
he
sli
ght
cha
nges
i
n
t
he
values
is
no
t
necessa
ry
res
ulti
ng
t
he
diffe
ren
ces
in
the
cat
eg
or
y
bu
t
on
ly
af
fec
t
the
degree
of
m
e
m
ber
sh
ip.
I
n
this
case,
t
he
cat
egory
3
is
us
e
d
f
or
e
xam
ple
in
creati
ng
a
set
of
va
riable
an
d
fu
zzy
i
nput/o
utp
ut.
T
he
value
is
giv
e
n
t
o
eac
h
disease
a
nd
pest
are
di
ff
e
re
nt
but
a
set
of
f
uzzy
to
be
us
e
d
ba
sed
on
pre
vious
resea
rc
h
as
the
baseli
ne
f
or
this
re
searc
h
f
or
the
pur
pose
of
com
par
ison
[
7]
.
The
r
ule
f
or
m
e
m
ber
sh
ip
will
be
cal
c
ulate
d
in
the
E
qu
at
ion
1
-
4
f
or
r
especti
ve
cat
e
gory
C
(Gray
Bl
igh
t)
in
the
in
dicat
or
based
on
weig
hting
for
in
dicat
or
(C
1
-
C
4)
.
The
Fig
ure
1
s
howe
d
the
us
e
d
f
or
m
in
the
ap
plica
ti
on
to
li
st
the
data
in
f
uzzy
set
s
on
diseas
e
and
pest
of
te
a
le
aves
in
gen
e
ral
base
d
on
t
he
cat
egorizat
ion
in
the
T
able 1
. A
fter
that,
t
he
fu
zzy
set
s w
as
created b
ase
d
on
it
s
sp
eci
fic d
ise
ases
a
nd
pe
sts,
a
s
sh
ow
n
in
Fi
gure
2.
T
he
Fi
gure
3
s
howe
d
th
e
us
ed
form
in
the
app
li
cat
io
n
to
analy
ze
di
sease
an
d
pest
of
te
a
le
aves
based
on
obser
vation
by
ent
ry
the
da
ta
directl
y
th
rou
gh
certai
n
value
i
n
e
ver
y
res
pected
in
di
cat
or
base
d on f
uzzy m
e
tho
d.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1273
–
1281
1276
Table
1.
A Set
of Tea
Diaseas
e an
d
Pe
s
t
No
Diseas
e/Pest
Ind
icato
r
W
DP
WI
1
Tea
Blis
te
r
Blig
h
t
/
Exob
a
sid
iu
m V
exan
s
L
i
t
t
l
e
p
a
l
e
g
r
e
e
n
s
p
o
t
s
0
.8
0
.1
T
r
a
n
s
l
u
c
e
n
t
l
i
g
h
t
o
n
y
o
u
n
g
le
a
v
e
s
0
.3
In 5
-
6
days,
the sp
o
ts ex
ten
d
s to
0.6
-
1
.3
c
m
0
.5
T
h
e
s
u
r
f
a
c
e
o
f
th
e
s
p
o
ts
w
a
s
c
o
v
e
r
ed
i
n
w
h
i
t
e
g
r
a
y
d
u
s
t
0
.7
T
h
e
s
u
r
f
a
c
e
o
f
th
e
s
p
o
ts
p
r
o
t
ru
d
e
s
d
o
w
n
w
a
r
d
0
.8
2
L
e
a
f
S
p
o
t
/
Cylin
d
ro
cla
d
iu
m
ilicico
la
A
t
t
a
c
k
o
n
th
e
t
i
p
o
f
the leaf
0
.7
0
.2
Leaves ap
art
f
ro
m
th
e stalk
0
.3
The sh
o
o
ts d
ry
ou
t
0
.5
B
r
o
w
n
s
p
o
t
s
o
n
t
h
e
l
e
a
v
e
s
0
.7
3
G
r
a
y
B
l
i
g
h
t
/
Pesta
lo
tia
thea
e
G
r
a
y
s
p
o
t
s
o
n
t
h
e
l
e
a
v
e
s
w
i
t
h
b
r
o
w
n
e
d
g
e
s
0
.5
0
.1
Mus
h
roo
m
s sp
r
ead to
sh
o
o
ts
0
.4
The sh
o
o
ts d
ry
ou
t
0
.6
Twigs
bro
k
en
and
y
ello
win
g
0
.8
4
R
e
d
R
o
o
t
R
o
t
/
Ga
n
o
d
erma
p
seo
d
u
ferr
eu
m
The leaves
turn
y
el
lo
w
0
.6
0
.2
The leaves
with
e
r
0
.3
The leaves
f
all ou
t
0
.4
Plan
ts d
ie
0
.6
There
a
re
red th
rea
d
s in
r
o
o
t su
rf
ace
0
.7
W
o
o
d
on
the sick
r
o
o
t is so
f
t and
dra
ws wate
r
wh
en
press
ed
0
.9
5
B
l
a
c
k
R
o
o
t
R
o
t
/
R
o
sellin
ia
a
rcu
a
ta
The leaves
turn
y
el
lo
w
0
.2
0
.2
The leaves
with
e
r
0
.4
The leaves
f
all ou
t
0
.5
Plan
ts d
ie
0
.6
There
a
re
b
lack
f
u
n
g
i thread
s in
r
o
o
t su
rf
ace
0
.7
There
a
re
b
lack
do
ts o
n
woo
d
of
the roo
ts
0
.8
6
T
e
a
M
o
s
q
u
i
t
o
B
u
g
/
H
elo
p
eltis
sp
p
Attack
occu
r
o
n
le
av
es o
r
sh
o
o
ts
0
.7
0
.2
Black
sp
o
t
o
n
tea
leav
es
0
.4
Bran
ch
es o
r
b
u
d
s h
av
e con
cave sp
o
t
s
0
.6
Twigs
withered
an
d
dry
0
.8
7
T
e
a
T
o
r
t
r
i
x
/
Ho
mo
n
a
coffea
ria
There
a
re
b
ite
m
a
r
k
s o
n
the leaves
0
.5
0
.2
Lar
v
ae
eat tea
leav
es
0
.4
Attack
in
g
y
o
u
n
g
leaves
0
.5
Leaves p
erfo
rated
0.
6
Leaf b
u
d
s b
ald
0
.8
8
W
a
l
k
e
r
/
Hyp
o
sid
ra
tala
ca
The b
u
d
s are
rolle
d
up
0
.3
0
.3
There
a
re
f
in
e thre
ad
s o
n
the sh
o
o
ts
0
.5
There
a
re
d
a
m
ag
es
on
the p
art
o
f
that is rolled
up
0
.8
Figure
1.
A
ppli
cat
ion
F
orm
I
nterf
ace
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Decisi
on M
aking in t
he
Te
a L
eaves
Disea
ses
D
et
ect
io
n U
sing M
amdani F
uz
zy
In
fe
re
nce
…
(
Arif
Ridho Lu
bis
)
1277
Figure
2. A
Set
of
F
uzzy
Gr
ay
Bl
igh
t
µA1
[
x
]
=
{
0
,
≤
0
≥
0
.
4
−
0
0
.
1
−
0
,
0
≤
≤
0
.
1
0
.
4
−
0
.
4
−
0
.
1
,
0
.
1
≤
≤
0
.
4
(1)
µA2
[
x
]
=
{
0
,
≤
0
.
1
≥
0
.
6
−
0
.
1
0
.
4
−
0
.
1
,
0
.
1
≤
≤
0
.
4
0
.
6
−
0
.
6
−
0
.
4
,
0
.
4
≤
≤
0
.
6
(2)
µA3
[
x
]
=
{
0
,
≤
0
.
4
≥
0
.
8
−
0
.
4
0
.
6
−
0
.
4
,
0
.
4
≤
≤
0
.
6
0
.
8
−
0
.
8
−
0
.
4
,
0
.
6
≤
≤
0
.
8
(3)
µA4
[
x
]
=
{
0
,
≤
0
.
6
−
0
.
6
0
.
8
−
0
.
6
,
0
.
6
≤
≤
0
.
8
1
,
≤
0
.
8
(4)
Figure
3. Pro
ble
m
A
naly
sis f
or Gray
Bl
ig
ht
Disease
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1273
–
1281
1278
Table
2.
Set o
f Fu
zzy
Inp
ut and
Ou
t
pu
t
Set of
Data
Inp
u
t Co
d
e
Ou
tp
u
t Co
d
e
Do
m
ain
Less (
L)
X1
Y1
0
-
45
Eno
u
g
h
(
E)
X2
Y2
40
-
75
Go
o
d
(
G)
X3
Y3
70
-
100
In
m
e
m
ber
sh
ip
functi
on,
it
ha
s
interval
0
t
o
1
an
d
f
or
c
urv
e
x,
val
ue
of
e
ach
va
riable
f
r
om
0
to
100.
Ther
e
f
or
e,
t
o
determ
ine
the
po
i
nt
of
it
s
m
e
m
ber
sh
ip
us
i
ng
the
tra
pezo
i
d
curve
re
pr
e
ssion
beca
us
e
it
can
be
m
easur
ed
acc
ordi
ng
ly
an
d
pr
eci
sel
y
in
reg
a
rd
to
t
he
lim
it
s
of
eac
h
dom
ain
.
O
n
eac
h
sid
e,
it
us
es
a
curve
to
te
rm
inate
the variable
of
a f
uz
zy
zo
ne
. In
add
it
ion
, th
e
m
e
m
b
ers
hip f
unct
io
ns
for
eac
h
set
are:
a.
A
Set
of Less
(L)
µA1
[
xi
]
=
{
0
,
≤
−
−
,
≤
≤
(5)
b.
A
Set
of E
nough (E
)
µA2
[
x
]
=
{
(
−
)
/
(
−
)
,
≤
≤
1
,
≤
≤
(
−
)
/
(
−
)
,
≤
≤
(6)
c.
A
Set
of Go
od
(
G)
µA3
[
x
]
=
{
(
−
)
/
(
−
)
,
≤
≤
1
,
≥
(7)
Wh
e
re:
a = Mi
nim
u
m
value o
f
le
ss
b
= Ma
xim
al
v
al
ue
of less
c = Mi
nim
u
m
value o
f
e
noug
h
d
= Ma
xim
al
v
al
ue
of e
noug
h
On
ce
t
he
va
ri
ables
an
d
set
are
f
or
m
ed,
th
e
app
li
cat
io
n
of
t
he
im
plica
ti
on
f
unct
ion
i
s
perform
ed.
Fo
r
e
xam
ple
there
is
a
ca
se
on
G
ray
Bl
ig
ht
as
f
ollo
ws,
the
e
xtent
t
o
w
hich
the
te
a
le
aves
a
re
i
ns
ula
te
d
by
the
disease
an
d
pest
of
Pes
ta
loti
a
theae
i
f
the
value
is
m
at
ched
.
A
cru
ci
al
co
ns
e
quence
of
the
MAX
aggre
gation
m
et
hod
i
s
that
IF
-
T
HEN
ru
le
s
in
a
Mam
dan
i
syst
e
m
hav
e
eff
ect
s
on
th
e
final
ou
t
pu
t
of
the
syst
e
m
that cannot
be
a
naly
zed inde
pende
ntly
o
f othe
r
I
F
-
THE
N
r
ules t
ha
t
m
ay
b
e fire
d sim
ultaneo
us
ly
in
the
syst
e
m
[
24
]
,
[
26]
,
[35
-
37]
.
Table
3.
A
Set
of Tea
Disease
and Pe
s
t
No
Co
d
e
Ind
icato
r
Valu
e
3.
C1
G
r
a
y
s
p
o
t
s
o
n
t
h
e
l
e
a
v
e
s
w
i
t
h
b
r
o
w
n
e
d
g
e
s
73
C2
Mus
h
roo
m
s sp
r
ead to
sh
o
o
ts
83
C3
The sh
o
o
ts d
ry
ou
t
80
C4
Twigs
bro
k
en
and
y
ello
win
g
76
The rule
de
rive
d from
the d
at
a
v
al
ue
of e
nd
-
di
seased
disease
is
1.
If
C
1
is e
noug
h,
C
2
is
good,
C3 is
good, a
nd c is
go
od, th
e
n pr
e
dicat
e v
al
ue (PV) is
at
ta
cked
2.
If
C
1
is
good,
C2 is
good, C
3 i
s go
od, a
nd C4
is
good,
the
n PV
is at
ta
c
ked
Af
te
r
g
et
ti
ng t
he rule t
hen lo
ok for m
e
m
ber
sh
ip
v
al
ue
a
nd
i
m
plica
ti
on
v
al
ue:
A.
If
C1
is En
ough C2
G
ood
, C3 Good, and C4 Goo
d,
T
hen
P
V
A
gainst. Aft
er th
e m
e
m
ber
sh
ip
value
is
ob
ta
ine
d, the
n l
ook for the
im
plica
ti
on
value
(
MI
N). a
nd th
e MIN
v
al
ue of
this rule:
α
_pred
ic
at
e
1 = μC∩
μB
∩ μ
B ∩ μB = m
in (
0.4
.1,1,1
)
=
0.4
B.
If
C1
G
ood
C
2
G
ood,
G
ood
C3,
a
nd
C
4
Goo
d,
T
he
n
P
V
A
gainst
.
A
fter
the
m
e
m
ber
sh
ip
value
is
ob
ta
ine
d, the
n l
ook for the
im
plica
ti
on
value
(
MI
N). a
nd th
e MIN
v
al
ue of
this rule:
α
_pred
ic
at
e
1 = μC∩
μB
∩ μB ∩ μB = m
in (0
.6
.
1,1,1
)
=
0.6
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Decisi
on M
aking in t
he
Te
a L
eaves
Disea
ses
D
et
ect
io
n U
sing M
amdani F
uz
zy
In
fe
re
nce
…
(
Arif
Ridho Lu
bis
)
1279
Af
te
r
the
im
plica
ti
on
value
i
s
ob
ta
ine
d,
the
nex
t
ste
p
is
the
com
po
si
ti
on
of
t
he
ru
le
that
ta
kes
the
MAX valu
e
of
the im
plica
t
ion value
that e
xis
ts:
Figure
4. G
raphic Re
su
lt
for C
om
po
sit
ion
R
ules
(a1
-
70)
/
(75
-
70)
=
0.6 => a
1 = 7
3
(8)
(a2
-
70)
/
(75
-
70)
=
0
=> a
2
=
70
(9)
Hav
i
ng
obta
in
ed
the
value
of
a
1
a
nd
a2
then
t
he
m
e
m
ber
sh
ip
f
unc
ti
on
ca
n
be
f
or
m
ed
base
d
on
t
he
deco
m
po
sit
io
n resu
lt
as
foll
ows:
µ
[
x
]
=
{
0
,
≤
70
(
70
−
)
/
(
75
−
70
)
,
70
≤
≤
73
0
.
6
,
≥
73
(10)
The
a
naly
sis
r
esult
can
help
decisi
on
m
aking
in
determ
i
ning
the
c
urre
nt
sta
tus
of
te
a
le
aves
f
or
furthe
r
act
ion.
The
la
st
pro
cess
is
us
in
g
centr
oid
m
et
ho
d
for
de
f
uzz
yfi
cat
ion
,
wh
i
ch
in
dicat
ed
86%
of
cat
egory C (
Gray
Bl
igh
t)
with
the
fo
ll
owin
g resu
lt
:
X*
=
1
+
2
+
3
1
+
2
+
3
X*
=
∫
0
70
0
∫
(
70
−
)
(
75
−
70
)
73
70
∫
0
.
6
100
73
=
(
73
∗
0
)
+
(
(
0
+
0
.
6
)
∗
(
73
−
70
)
)
+
(
(
1
00
−
73
)
∗
0
.
6
)
X*
=
0
+
(
−
64
,
8
)
+
(
15
30
)
0
+
(
0
.
9
)
+
(
16
.
2
)
X*
=
1465
.
2
17
.
1
X
*=
86.
212
Figure
5. Pro
ble
m
Result
f
or
Com
po
sit
ion
R
ules
by Ap
plica
ti
on
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
12
, N
o.
3
,
Dece
m
ber
2
01
8
:
1273
–
1281
1280
The
Fig
ur
e
5
sh
owe
d
the
fina
l
resu
lt
in
the
a
naly
sis
pr
oce
ss
fo
r
te
a
le
aves
base
d
on
F
uzz
y
Mam
dan
i
Infer
e
nce,
wh
i
ch
is
i
m
ple
m
e
nted
in
the
a
ppli
cat
ion
to
he
lp
far
m
ers
diagnose
the
ty
pe
of
disea
se
an
d
pest
,
wh
ic
h
at
ta
cke
d
the
te
a
le
ave
s.
In
la
r
ge
scal
e,
after
t
he
tria
l
and
er
ror
te
st,
i
t
is
exp
ect
e
d
t
he
ap
plica
ti
on
c
an
be
us
e
d
f
or
decisi
on
m
aking
pro
cess
to
f
ur
th
er
reduce
the
bur
den
i
n
ga
rd
e
ni
ng
m
anag
em
ent.
On
the
oth
e
r
hand,
the
eff
ic
ie
ncy
al
so
bec
om
e
t
he
pri
m
ary
ob
je
ct
ive
to
sel
ect
sp
eci
fic
ap
plica
ti
on
to
be
use
d.
T
he
com
pa
rison
with
pr
e
vious
m
et
ho
d
al
so
ha
s
been
cal
culat
ed
that
sh
owe
d
78
%,
lo
wer
ef
fici
ency
com
par
e
to
fu
zzy
m
et
hods
as foll
ows:
P(
C|
C1)=
P
(
C
|
C1
)
∗
C
P
(
C
|
C1
)
∗
C
=
0
.
1
∗
0
.
8
0
.
1
∗
0
.
8
=
1
P(
C|
C2)=
P
(
C
|
C2
)
∗
C
P
(
C
|
C2
)
∗
C
=
0
.
4
∗
0
.
8
0
.
4
∗
0
.
8
=
1
P(
C|
C3)=
P
(
C
|
C3
)
∗
C
P
(
C
|
C3
)
∗
C
=
0
.
6
∗
0
.
8
0
.
6
∗
0
.
8
=
1
P(
C|
C3)=
P
(
C
|
C4
)
∗
C
P
(
C
|
C4
)
∗
C
=
0
.
8
∗
0
.
8
0
.
8
∗
0
.
8
=
1
Pr
oba
bili
ty
To
ta
l = P(C|C
1) +
P(
C|
C2) + P
(C|C
3)
+
P
(C|C
4)
= 4
(N
P|C
1)
/
Pro
ba
bili
ty
To
ta
l = 73
/ 4
=
18.25
(N
P|C
2)
/
Pro
ba
bili
ty
To
ta
l = 83 / 4
=
20.75
(N
P|C
3)
/
Pro
ba
bili
ty
To
ta
l = 80 / 4
=
20
(N
P|C
4)
/
Pro
ba
bili
ty
To
ta
l = 76 / 4
=
19
Total
A
tt
acke
d C
ount =
18.2
5 + 2
0.7
5
+
20
+ 19 =
78%
5.
CONCL
US
I
O
N
In
s
umm
ary,
t
his
stu
dy
us
e
s
the
Ma
m
dan
i
fu
zzy
in
fe
re
nce
m
e
tho
d
withi
n
the
6
(six)
phases
nam
el
y
the
determ
inatio
n
of
the
f
uz
zy
var
ia
bles,
f
uzzy
set
s
a
nd
fu
zzy
dom
ai
ns
,
ad
justm
ent
fo
r
f
uzzifica
ti
on,
t
he
form
ation
of
f
uzzy
r
ule
in
th
e
form
of
IF
-
T
HEN,
in
fer
e
nc
e
us
in
g
Ma
m
dan
i
m
e
tho
d
t
ha
t
is
m
in
i
m
plication
functi
on
(m
inim
u
m
)
and
m
a
x
r
ule
c
om
po
sit
ion
(m
axi
m
u
m
)
and
la
stl
y,
defuzzifi
cat
io
n.
I
n
a
dd
it
io
n,
work
i
ng
with
dif
fer
e
nt
im
plications
and
at
the
sa
m
e
tim
e,
the
po
s
sibil
it
y
of
com
bin
ing
the
res
ult
gi
ven
by
th
e
i
m
plica
ti
on
s
of
these,
there
fore,
offe
r
a
so
li
d
basis
f
or
a
m
or
e
accurate
resu
lt
s
of
com
pu
ta
ti
onal
ap
plica
ti
on
syst
e
m
to
be
dev
el
op
e
d.
M
oreo
ver,
it
can
be
im
pr
ov
e
d
by
add
in
g
new
i
m
plica
ti
on
s,
by
us
in
g
m
or
e
fu
zzy
m
at
ching
te
ch
niques
or
by
ot
her
a
ggre
gate
op
e
rato
rs
to
obta
in
a
n
ov
e
rall
crisp
act
io
n
f
ro
m
t
ho
se
pro
vid
e
d,
separ
at
el
y
th
rough
e
ve
ry
im
p
li
cat
ion
.
On
e
of
the
c
oncer
n
of
our
fu
t
ur
e
stud
y
is
t
o
e
xpan
d
t
his
syst
e
m
by
include
un
ce
rtai
nty
ab
ou
t
t
he
series
of
f
uzzy
m
e
m
ber
sh
ip
f
un
ct
io
n
asso
ci
at
ed
wit
h
the
li
nguisti
c
te
rm
s.
It
al
s
o
wan
t
t
o
c
om
par
e
the
e
ff
i
ci
e
ncy
of
this
m
et
hod
with
oth
ers
in
sim
il
a
r
ob
j
ect
or
ta
rg
et
t
o
gras
p
m
or
e
unde
rstan
dab
le
of its st
re
ng
t
hs an
d weak
ness
es.
REFERE
NCE
S
[
1
]
Hicks
A.
“
Review
of
Glob
al
Te
a
Producti
on
and
the
Im
pact
on
Industr
y
o
f
the
As
ia
n
Econom
ic
Situatio
n”
.
Ret
ri
eve
d
f
rom
h
tt
p://citese
erx
.
i
st.psu.
edu/
v
ie
wd
oc/
download
?
do
i=
10.
1
.
1.
543
.
89
31&re
p=r
ep1&
t
y
pe
=pdf
[
2
]
Kaison
C.
“
W
orl
d
Te
a
Producti
o
n
and
Tra
de
Cur
ren
t
and
Future
Deve
lopment”.
Food
and
Agriculture
Or
ganizati
on
of
th
e
Unit
ed
Na
ti
ons
.
Tra
d
e
a
nd
Marke
ts Di
visio
n,
Rom
e
2015
.
[
3
]
Ah
m
ed
M.
“
Ec
o
frie
ndl
y
pest
m
ana
gement
of
t
ea
in
Bang
la
desh”
.
Tw
o
and
a
Bud
59
(2012), 11
-
16
.
[
4
]
Cle
m
son.
“
Camell
i
a
Dise
ase
s &
Insec
t
Pests”
.
H
ome
&
Gar
den
Information
Cen
t
er
(
HG
IC
)
,
2013.
[
5
]
Danz
inge
r
HL.
“
Disea
ses a
nd
Pe
sts of
Te
a
:
Over
vie
w a
nd
Pos
sibi
li
ties of
I
nt
egr
ated
Pest
and
Dise
ase
Mana
g
ement
”.
Journal
of
Agriculture in
th
e
Tr
o
pic
s and
Sub
tropic
s
101
(Apr
il
20
00)
13
-
38.
[
6
]
Ahm
ed
S.
“
Te
a
and
the
Ta
st
e
of
Cli
m
at
e
Chang
e:
Understa
ndin
g
Im
pac
ts
of
Envi
ronm
ent
al
Var
ia
ti
on
on
Bot
anica
l
Quali
t
y
”
.
Herbal
gram
103
(2014
),
44
-
51
.
[
7
]
Ta
rig
an
MP
.
“
Im
ple
m
ent
asi
Metode
Na
ïve
B
a
y
esia
n
d
an
Ba
ckwa
rd
Chai
n
in
g
pada
Sis
te
m
Pakar
Diagnos
a
Pen
y
ak
it
/H
ama Tana
m
an Te
h
”.
Bac
he
lor E
ss
ay
(2016).
Univ
ersi
ta
s Sum
at
era Ut
ara
.
[
8
]
Rez
a
ei
M,
As
ad
iz
ad
eh
M,
Majd
i
A,
Hos
saini
MF
.
“
Predic
ti
on
o
f
rep
rese
nt
ative
def
orm
at
ion
m
odulus
of
longwa
ll
pane
l
roof
roc
k
strat
a
using
M
a
m
dani
fuz
z
y
s
y
s
te
m
”.
In
te
rnatio
nal
Journal
of
Mini
ng
Sc
ie
nc
e
and
Technol
ogy
25
(2015)
23
-
30.
[
9
]
Afons
o
ACM
,
Nett
o
AM
,
de
Vasconc
el
os
E
.
“
Fuzz
y
Log
ic
A
ppli
ed
to
the
M
odel
ing
of
W
ate
r
D
y
namics
in
an
Oxisol
in
North
e
aste
rn
Br
azil”.
R
ev
ista
Brasileira de
C
ie
n
ci
a
do
S
olo
38(2)
Apr
20
04.
[
1
0
]
Perm
at
asa
ri
D
,
Aziz
ah
IN,
Had
ia
t
HL,
Ab
adi
AM
.
“
Cla
ss
ifi
cation
of
Toddler
Nutrit
ion
al
Sta
tus
using
Fuz
z
y
Infe
ren
c
e
S
y
ste
m
(FIS
)”.
AIP C
onfe
renc
e Proceedings
1868,
04
0007
(2017).
[
1
1
]
El
amvaz
u
thi
I,
Vasant
P,
W
ebb
J.
“
The
Applic
ation
of
Mam
dani
Fuzz
y
Model
for
Auto
Zoom
Fun
ct
ion
of
a
Digit
a
l
Camera
”
.
I
JCSI
S
6(3)
2009
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Decisi
on M
aking in t
he
Te
a L
eaves
Disea
ses
D
et
ect
io
n U
sing M
amdani F
uz
zy
In
fe
re
nce
…
(
Arif
Ridho Lu
bis
)
1281
[
1
2
]
Paszto
V,
Br
y
c
h
tova
A,
Tucek
P
,
Mare
k
L
,
Buri
a
n
J.
“
Us
ing
a
Fu
zzy
Inf
ere
n
ce
S
ystem
to
Del
imit
Rural
and
U
rba
n
Munici
palities i
n
the Czech
R
epu
bli
c
in
2010
”.
Jo
urnal
of
Maps
1
1(2)
2015.
[
1
3
]
Ahlan
AR,
Lubi
s
M,
Lubi
s
AR.
“
Inform
at
ion
Secur
ity
A
ware
ness
at
the
Know
le
dge
-
Ba
sed
Instit
uti
on:
Its
Antec
ed
ent
s
and
Mea
sures”
.
Proce
dia
Computer
Sci
en
ce
72
(201
5),
361
-
373
.
[
1
4
]
Za
deh
LA.
“
Calculus
of
fuz
z
y
re
stric
ti
ons”
.
in
L
.
A.
Zadeh,
K.
-
S
.
Fu,
K.
Tanaka
&
M.
Shimura
(
e
ds)
,
Fuzzy
Set
s
and
the
ir
Appl
i
cat
ion
s to
Cogni
ti
v
e
an
d
Dec
ision
Proc
esses,
A
cade
mic
Press
,
1975,
Ne
w York,
pp.
1
–
3
9.
[
1
5
]
Mam
dani
EH.
“
Applic
a
ti
on
of
f
uzzy
log
ic
to
ap
proximate
r
ea
so
ning
using
li
ngu
isti
c
s
y
nth
esis”.
IEE
E
Tr
ansacti
o
ns
on
Computers
1
977,
26(12):
118
2
–
1191.
[
1
6
]
W
ang,
LX.
(199
2
)
.
“
Fuzzy
S
y
s
t
ems
are
Univer
sal
Approxim
at
or
s”,
Proc.
of
I
EEE
Inte
r.
Conf
.
o
n
Fuzzy
Syste
ms
,
San
Diego
,
US
A,
pp
.
1163
–
1170
.
[
1
7
]
Buckl
e
y
JJ
.
(199
3).
“
Sugeno
t
y
p
e
cont
ro
ll
ers
are u
nive
rsal
cont
r
ollers”,
Fuzzy
S
et
s
and
Syste
ms
Vol
.
53:
299
–
304.
[
1
8
]
Ian
cu
I
.
“
A
Mam
dani
T
y
pe
Fuz
z
y
Logic
Contro
ll
er
”.
Fuz
z
y
Log
ic
–
Contro
ls,
C
once
pts,
The
or
ies
and
Applicati
o
n”
.
Prof.
El
mer Dad
ios (
ed.
)
.
ISN
B: 978
-
953
-
51
-
0396
-
7.
InT
ec
h
201
2.
[
1
9
]
Ibti
sam
bt.
Ya
acob
N.
and
Lub
is
M.
“
Multi
Age
nt
Inte
gr
at
ion
M
odel
for
Agen
E
xec
ut
ive
Inform
at
ion
S
y
s
te
m
:
A
Case
Stud
y
in
KU
IS”.
IIUM
P
ress
.
Kuala
Lu
m
pur.
ISBN
9789674180843
in
Kart
iwi
,
Mir
a
a
nd
M.Z
.
M
Khed
her
,
Akram
,
eds.
(20
11)
Data ma
nag
ement:
issues,
ch
al
l
enge
s a
nd
opp
ortuni
ties.
[
2
0
]
Abrori
M,
Prihama
y
u
AH
.
“
Aplika
si
Logi
k
a
Fuzz
y
Metod
e
Ma
m
dani
dal
am
Pengambila
n
Kepu
tusan
Penent
ua
n
Jum
la
h
Produks
i”
.
Kaun
ia
Vol
.
XI (2)
Octob
er
2
015/1436:
91
-
99
.
[
2
1
]
Ahlan
AR,
Lubis
M.
“
Inform
at
i
on
Secur
i
t
y
Aw
ar
ene
ss
in
Univ
ersity
:
M
ai
nt
ai
ni
ng
Learna
bi
li
t
y
,
Perform
anc
e
an
d
Adapta
bi
li
t
y
thr
ough
Role
s of
R
esponsibil
ity
”
.
I
nt.
Con
f. of
IAS
2011.
[
2
2
]
Maha
la
kshm
i
P,
Gane
san
K.
“
Mam
dani
Fuzz
y
R
ule
base
d
Mode
l
to
Cla
ss
if
y
Sit
e
s
for
Aquac
ult
ur
e
Deve
lopment
”
.
Indian
J
.
Fi
sh
6
2(1):
110
-
115
,
2
015.
[
2
3
]
Pal
SK
,
Mand
al
DP
.
“
Fuzz
y
Logic
and
Approxim
at
e
Rea
soning
: A
n
Overvi
ew”
.
I
ETE
J
.
Re
s
.
37
(
2015)
548
–
560.
[
2
4
]
Izq
uie
rdo
SS
,
I
z
quie
rdo
LR
.
“
Mam
dani
Fuzz
y
S
y
stems
for
Mode
li
ng
and
Sim
ulat
ion:
A
Crit
i
cal
As
sessment”
.
SSR
N
El
e
ct
ronic
Journ
al
(2
017),
DO
I: 10.2139/
ss
rn.
29
00827
[
2
5
]
Klawonna
F,
Novák
V.
“
The
R
el
a
ti
on
b
et
we
en
Infe
ren
c
e
and
In
te
rpol
at
ion
in
th
e
Fram
ework
of
Fuzz
y
S
y
stems
”
.
Fuzzy
Set
s S
yst
.
81
(1996)
331
–
3
54.
[
2
6
]
Klir
GJ
,
Yuan
B
.
“
Fuzz
y
Sets
an
d
Fuzz
y
Logic:
The
or
y
and
Applicati
ons”
.
Prentice
Hall
PTR,
Up
per
Saddle
Rive
r
,
New Jerse
y
,
199
5.
[
2
7
]
Ibti
sam
bt.
Yaa
cob
N.
and
Lub
is
M.
“
Data
Acc
ura
c
y
and
Age
nt
Inte
gr
at
ion
in
Exe
cu
ti
ve
Infor
m
at
ion
S
y
stem:
A
Case
Stud
y
in
KU
IS”.
IIUM
P
ress
.
Kuala
Lu
m
pur.
ISBN
9789674180843
in
Kart
iwi
,
Mir
a
a
nd
M.Z
.
M
Khed
her
,
Akram
,
eds.
(20
11)
Data manag
ement:
issues,
ch
al
l
enge
s a
nd
opp
ortuni
ties.
[
2
8
]
Lubi
s
AR,
Fa
ch
riz
a
l
F,
Lubi
s
M.
“
The
Eff
ect
of
Socia
l
Medi
a
to
Cul
tura
l
Hom
ec
om
ing
Tra
di
ti
on
of
Com
put
er
Student
s in
Med
an”
.
Proc
edi
a
C
omputer
Scienc
e
,
124
,
pp
.
423
-
4
28,
2017
.
[
2
9
]
Lubi
s
AR.
“
Im
ple
m
ent
ation
of
Di
vide
and
Conque
r
Te
ch
nique:
Ka
rat
suba
Algori
th
m
for
Multi
plicc
at
ion
B
ig
Int
ege
r
with
Ph
y
ton”.
Ju
rnal
Saintek
vo
l. 28
(2)
,
2014
.
[
3
0
]
Olive
ir
a
DN
,
de
Li
m
a
Hee
n
GA
,
Alm
ei
da
OM
.
“
Design
and
Im
ple
m
ent
ation
o
f
a
Mam
dani
Fuzz
y
Infe
ren
c
e
S
y
st
e
m
on
an
FP
GA
us
ing
VH
DL”
.
An
nual
Confe
renc
e
of
the
North
Ame
rican
Fuzzy
Information
Proc
essing
Soci
ety
–
NAF
IPS
,
Jul
y
20
10.
[
3
1
]
Ta
rig
an
P,
Sinur
at
S,
Sinambela
M.
“
Im
ple
m
ent
a
ti
on
of
a
Mam
d
ani
fu
zzy
logic
cont
roller
for
bu
il
ding
aut
om
atio
n
using e
l
ec
tron
ic
cont
rol
base
d
on
AT89S51”.
IEEE
TIME
-
E
2015.
[
3
2
]
Hasan
MH
,
Abdul
Azi
z
I
,
Jaa
far
J,
Abdul
R
ahi
m
L,
Man
y
i
el
JM
A.
“
A
compara
tive
stud
y
of
Ma
m
dani
and
Suge
no
fuz
z
y
m
odel
s
for
qualit
y
of
web
services monitoring”.
I
JA
CS
A
vol
.
8(9)
,
2017
.
[
3
3
]
Hakim
L,
Kus
um
asa
ri
TF,
Lub
i
s
M.
“
Te
xt
m
ining
of
UU
-
ITE
I
m
ple
m
ent
at
ion
i
n
Indone
sia”.
Jo
urnal
of
Phy
si
cs:
Confe
renc
e
S
eri
es
1007
(1)
,
201
8.
[
3
4
]
Lubi
s
AR,
Fachr
iz
a
l
F,
Lubi
s
M,
Ta
hir
HM
.
“
W
ire
le
ss
service
at
publi
c
univ
ersity:
A
surve
y
of
us
ers
per
ce
p
ti
on
o
n
s
ec
urity
aspe
ct
s”
.
IE
EE ICOIACT
2018.
[
3
5
]
Ta
n
L
.
“
Ta
k
agi
-
Sugeno
and
Ma
m
dani
Fuzz
y
Co
ntrol
of
a
R
esort
Mana
gement
S
y
stem”
.
Master
Thesis
,
Novem
ber
2011.
Bl
eki
nge
I
nstit
ute of Te
chn
olog
y
.
[
3
6
]
Kam
boj
V,
Kau
r
A
.
“
Com
par
ison
of
const
ant
Sugeno
-
t
y
p
e
an
d
Mam
dani
-
t
y
p
e
fuz
z
y
inf
ere
n
ce
s
y
s
te
m
for
l
ea
d
sensor”.
I
JSCE
,
vol.
3
(2), Ma
y
2
013.
[
3
7
]
W
ang
C.
“
A
st
ud
y
of
m
embership
func
ti
ons
on
Mam
dani
-
t
y
p
e
fuz
z
y
infe
r
en
ce
s
y
st
em
for
industri
al
de
ci
sio
n
-
m
aki
ng”.
Theses and
Diss
ertat
io
ns
,
Paper
1665,
2015.
[
3
8
]
W
y
rwol
B,
Hr
ynkie
wic
z
E
.
“
Im
ple
m
e
ntation
of
the
FITA
fuz
z
y
infe
ren
cce
s
y
st
e
m
on
the
spec
if
i
c
m
ic
roc
ont
rolle
r
pla
tform”.
IF
AC
-
Pape
rs
OnL
ine
vol.
48(4)
,
pp
.
1
65
-
169,
2015
.
Evaluation Warning : The document was created with Spire.PDF for Python.