Algorithm 4 Cluster_Fusion.
Input: Dataset pp that needs fusion detection
Output: Fusion clustering results
1	for I in pp // Count any two clusters in the current dataset.
2		for j in pp
3			if there are two points with distance < dc and from two different labels
4				marked as adjacent samples and adjacent clusters;
5		end if
6	end for
7	if meet the fusion rules(The number of adjacent samples accounts for more than 2% of the total number of two adjacent clusters)
8		Update the cluster center cc to re-cluster, and perform fusion detection again after re-labeling;
9		Update dataset labels;
10	end if
11	return cluster  // Return clustering results.