Table 3 Parameter settings used for developing the fuzzy models

Data preparation

Data normalization

Yes

Percentage used for training/testing the models

75%/25%

Model structure

Number of clusters

2

Shape membership functions

Gaussian

T-norm

Minimum

Model construction

Clustering method

FCM coupled with FST-PSO

Degree of fuzziness of the identified cluster structure (m)

2

FST-PSO maximum number of iterations

100

GRABS threshold [33]

0.9