Boosted Fuzzy Granular Regression Trees

Algorithm 1

Clustering algorithm with automatic optimization of cluster centers.

Input: instance set , maximum iteration , threshold value
Output: optimization cluster center set
(1)	Remove instances missing some attribute values.
(2)	Normalize each attribute value into .
(3)	Let evaluated set be an empty set.
(4)	Initialize current iteration.
(5)	WHILE OR
(6)	. Initialize the current cluster center set.
(7)	. Initialize the number of cluster center.
(8)	Random choose 1 instance point as cluster center.
(9)
(10)
(11)	WHILE
(12)
(13)	The probability of that is selected as next cluster center
(14)	p = GenProb(); Random generate a probability.
(15)	IF THEN
(16)	END WHILE
(17)	. Calculate the loss function value and cluster center in this iteration and update the evaluated set.
(18)	Update the current iteration.
(19)	END WHILE
(20)	. In the evaluated set , choose the cluster center set with minimum loss function value.
Return the optimization cluster centers and their number.
(21)	, ( represents the number of elements of the set).