Page 64 - Proceeding The 2nd International Seminar of Science and Technology : Accelerating Sustainable Innovation Towards Society 5.0
P. 64
nd
The 2 International Seminar of Science and Technology
“Accelerating Sustainable innovation towards Society 5.0”
ISST 2022 FST UT 2022
Universitas Terbuka
1 -1
2
b.[∑ m (X ij -V kj ) ] η-1
j=1
t ik =[1 + ( ) ] (7)
γ k
where the value γ can be found by Equation (8).
k
2
w
∑ n i=1 [(μ ik ) (∑ m (X ij -V kj ) )]
j=1
γ = K ∑ n (μ ik ) w (8)
k
i=1
4. Fixed cluster center ( V ) with Equation (9).
kj
w
∑ n (aμ ik +bt ik η )X ij
V = i=1 w η (9)
kj
n
∑
i=1 (aμ ik +bt ik )
5. Repair matrix peculiarity relatively (μ )and the absolute
ik
uniqueness matrix ( t ) according to the formula stated in the
ik
FPCM algorithm.
6. Check condition stop:
a. If | Pt – Pt-1 | < ε or t > Max iteration, then stop
b. If no: t = t+1 then repeat step 4
2.6 Validity Cluster
One method for knowing the quality of the clusters formed is with see
score validity clusters. The value of the cluster validity index in the
fuzzy method exists two that is Partition Coefficient (PC) and Modified
Partition Coefficient (MPC) [11]. MPC is index validity results repair
from the PC, because on the PC the value is tend experience change.
Following formula for look for MPC value as in Equation (10), where n
= many studies, c = many group, U = score membership object i-th
ik
with center group to -k.
C
MPC =1- (1-PC(c)) (10)
C-1
As for the formula for look for PC values in Equation (11):
1 c n 2
PC = ∑ k=1 ∑ i=1 U ik (11)
n
3 RESULT
3.1 Application Method
3.1.1 Grouping with Fuzzy C-Means (FCM)
Table 1. Grouping results with FCM.
Cluster MPC Number of Iterations Computing Time
ISST 2022 – FST Universitas Terbuka, Indonesia 43
International Seminar of Science and Technology “Accelerating Sustainable
Towards Society 5.0