The International Journal of Fuzzy Logic and Intelligent Systems

eISSN 2093-744X
pISSN 1598-2645

SSRF-CA algorithm.

Fix the maximum number of clusters C = Cmax and m ∈ [1, ∞);
Initialize the fuzzy partition matrix U(CL) randomly;
Set α1, α2, and α3;
Initialize the matrices Ai to the identity matrix;
Initialize the threshold ε1;
Compute the initial cardinalities Ni for 1 ≤ iC by using Eq. (10);
REPEAT
Compute the dissimilarity RAi for all clusters using Eq. (6);
Compute the membership vectors vi using Eq. (5);
For each point xk, k ∈ 1, 2, …, N
Compute the distance distik2 using Eq. (4);
Update α3(k) using Eq. (14) and (15);
Update the fuzzy membership using Eq. (7);
Update the measure distance matrices Ai using Eq. (12);
Compute the cardinalities of all clusters Ni by using equation Eq. (10);
If (Ni < ε1) remove cluster i and update the number of clusters C;
UNTIL (The max number of iterations is reached or the fuzzy memberships stabilize)
International Journal of Fuzzy Logic and Intelligent Systems 2020;20:156~167 https://doi.org/10.5391/IJFIS.2020.20.2.156
© IJFIS
© 2020. The Korean Institute of Intelligent Systems. / Powered by INFOrang Co., Ltd