Skip to Main Content
Table 2

The results of temporal modeling of AO7 concentration using ANN and ANFIS method for TP6

ModelInput combinationNetwork structureRMSE (absorption)
DC
CalibrationVerificationCalibrationVerification
ANN Comb. (1) 3-4-1a 0.064 0.069 0.985 0.724 
Comb. (2) 3-4-1 0.061 0.051 0.986 0.853 
Comb. (3) 4-3-1 0.057 0.034 0.988 0.933 
Comb. (4) 5-6-1 0.045 0.058 0.992 0.806 
ANFIS Comb. (1) trimfb 0.110 0.053 0.956 0.838 
Comb. (2) trimf 0.076 0.050 0.979 0.860 
Comb. (3) gauss2mf 0.008 0.018 0.998 0.981 
Comb. (4) gussmf 0.128 0.064 0.940 0.768 
ModelInput combinationNetwork structureRMSE (absorption)
DC
CalibrationVerificationCalibrationVerification
ANN Comb. (1) 3-4-1a 0.064 0.069 0.985 0.724 
Comb. (2) 3-4-1 0.061 0.051 0.986 0.853 
Comb. (3) 4-3-1 0.057 0.034 0.988 0.933 
Comb. (4) 5-6-1 0.045 0.058 0.992 0.806 
ANFIS Comb. (1) trimfb 0.110 0.053 0.956 0.838 
Comb. (2) trimf 0.076 0.050 0.979 0.860 
Comb. (3) gauss2mf 0.008 0.018 0.998 0.981 
Comb. (4) gussmf 0.128 0.064 0.940 0.768 

aThe mentioned values on the ANN structure stand for the number of input, hidden, and output neurons, respectively.

btrimf: Triangular-shaped; gauss2mf: Gaussian combination; gaussmf: Gaussian curve membership function.

Close Modal

or Create an Account

Close Modal
Close Modal