A Clustering-Guided Integer Brain Storm Optimizer for Feature Selection in High-Dimensional Data

Algorithm 1

The proposed IFC.

Input: The original feature set, Fset;
Output: The feature clustering result,
(1)	Use the SU measure to evaluate the importance of each feature in Fset;
(2)	Sort all features in Fset in the decreasing order of SU values, denoted the sorted result by Fsorted;
(3)	Set i = 1;
(4)	While |Fsorted| > 0%|Fsorted| is the size of Fsorted
(5)	Set the first feature in Fsorted to be the i-th cluster center, Centeri
(6)	Initialize ;
(7)	For j = 1: |Fsorted|
(8)	Calculate the correlation between Centeri and the j-th feature in Fsorted;
(9)	If the correlation > η, then save the j-th feature into;
(10)	Endfor
(11)	Reset , and i = i + 1.
(12)	Endwhile