GSPSO-LRF-ELM: Grid Search and Particle Swarm Optimization-Based Local Receptive Field-Enabled Extreme Learning Machine for Surface Defects Detection and Classification on the Magnetic Tiles

Algorithm 1

: GS optimization for the parameters C and K.

(1)	Setting the maximum, minimum and step size of C and K to get array A and array
(2)	B respectively. The numerical angle in array A is 1: m, and the numerical angle in
(3)	array B is 1: n,
(4)	Bastacc = 0;
(5)	For C = 1: m %C is the balance parameter
(6)	{
(7)	For K = 1: n %K is the number of feature maps
(8)	{
(9)	Substituting C and K into the ELM-LRF algorithm, the classification
(10)	accuracy of the algorithm is obtained;
(11)	If Acc (C, K) > Bastacc
(12)	Bestacc = Acc (C, K);
(13)	BestC = C;
(14)	BestK = K;
(15)	End
(16)	}
(17)	}