A Wrapper Feature Subset Selection Method Based on Randomized Search and Multilayer Structure

Algorithm 1

Description of the MLFSSM algorithm.

	Algorithm: MLFSSM
	Input:
	dataset X;
	input feature set F;
	Begin
	For l = 1, …, L
	Calculate for f ∈ F using equation (1);
	Calculate posl,f for f ∈ F using equation (3);
	Generate M · l-layer’s feature subsets by pos values;
	For m = 1, …, M
	Construct classification models using ;
	Endfor;
	If the accuracy rate of top T · l-layer’s feature subsets reaches 100%
	Break;
	Endfor;
	Return the topmost subset of l-layer;
	End