Identification of a Typical CSTR Using Optimal Focused Time Lagged Recurrent Neural Network Model with Gamma Memory Filter

Table 4

Parameters of Fully Recurrent NN (Benchmark data). Number of Exemplars: Training = 3000, Testing = 4500, Max epoh = 1000, Partially recurrent, Gamma, Norm → L-infinity.

Sr. no.	Parameters	Hidden layer 1	Hidden layer 2	Output layer
1	Neurons	16	27	1
2	Transfer function	tan h	tan h	Linear tan h
3	Learning rule	dbd	dbd	dbd