A System Fault Diagnosis Method with a Reclustering Algorithm

Algorithm 1

The improvement of K-means.

	Require: the initial training set , is the ith log message in C; maximum number of clusters max_k; the threshold of the sum of squares of distance between data point and cluster center min_dist; the distance threshold between two cluster centers center_list.
Ensure: log vector matrix D
(1)	function Improvement K-means
(2)	for ido
(3)	ifthen
(4)	set K = i
(5)	end if
(6)	end for
(7)
(8)	While centers do
(9)	fordo
(10)	divide int the closest cluster cluster[k]
(11)	end for
(12)	end for
(13)	ifthen
(14)	merge and
(15)	end if
(16)	end while
(17)	return D
(18)	end function