Novel Approaches to Identify Clusters Using Independent Components Analysis with Application

Algorithm 2

RR2_clustering_algorithm.

	Input: : matrix of observations
	Output: C: clusters
(1)	;/^∗ execute ICA algorithm to generate ICs ^∗/
	/^∗compute estimated mixing matrix W and the corresponding de-mixing matrix A^∗/
(2)	;
(3)	;
	/^∗ determine a ranking of ICs in ^∗/
(4)	;
(5)	Perform reconstruction of the original series at arbitrary threshold levels
(6)	Compare each of the reconstructed series with original series and compute adjusted coefficient of determination
(7)	Rank the values of computed in ascending order for each variable in .
	/ ^∗Perform clustering of variables based on similar ranking patterns.^∗/
(8)	;