Random Fuzzy Granular Decision Tree

Algorithm 1

Adaptive global random clustering.

	Input: instance set , maximum iterations
	Output: optimal cluster center set and its number
(1)	Remove instances of missing feature values.
(2)	Normalize feature values of instances into .
(3)	Initialize evaluation set.
(4)	Set current iteration.
(5)	While
(6)	Let the current cluster center be an empty set.
(7)	Randomly generate the quantity of cluster center in
(8)	Initialize the quantity of cluster center.
(9)	Randomly select an instance as a cluster center in the dataset.
(10)
(11)	While
(12)
(13)	. Calculate the probability selected of as the next cluster center.
(14)	p = GenProbability (); Randomly generate a probability.
(15)	If then
(16)	End while
(17)	The standard deviation of intercluster and one of inner-cluster are calculated and stored and updated in the evaluation set.
(18)	Update current iterations.
(19)	End while
(20)	In the evaluation set , choose the cluster center set with the largest ratio.
(21)	, (here, represents the quantity of elements of the set). Return optimal parameters.