Table 1 Training and testing errors, learning and prediction times for the best learned FIS of ANFIS, PSOFIS and GAFIS methods.

From: Influence of number of membership functions on prediction of membrane systems using adaptive network based fuzzy inference system (ANFIS)

Methods

ANFIS

GAFIS

PSOFIS

Number of inputs

5

5

5

Maximum of iteration

500

500

500

Percentage of P

70

70

70

Clustering type

Grid partition

Subtractive clustering

Subtractive clustering

Training MSE error

0.001535665

0.006535045

0.005841158

Training RMSE error

0.039187567

0.080839622

0.076427468

Training mean error

1.86015E − 06

− 0.00145448

0.002179109

Training standard deviation (StD)

0.039221103

0.080895708

0.076461776

Training correlation coefficient (R) error

0.99832132

0.992869298

0.994394045

Training coefficient of determination (R2) error

0.996645458

0.985789443

0.988819516

Testing MSE error

0.009672048

0.00655753

0.006493703

Testing RMSE error

0.098346572

0.080978575

0.080583516

Testing mean error

− 0.007010998

0.000766209

0.002871209

Testing standard deviation (StD)

0.098155074

0.081023424

0.080580558

Testing correlation coefficient (R) error

0.989996681

0.993137387

0.99320752

Testing coefficient of determination (R2) error

0.980093427

0.98632187

0.986461178

Learning time (s)

102.6507035

85.5361454

57.1184928

Prediction time (s)

0.7148131

1.267197

0.1744804