On the Accurate Identification of Network Paths Having a Common Bottleneck

Table 4

Results of FlowMate and our SVD based approach with three correct clusters ({1, 2, 3}, {4, 5, 6, 7, 8}, and {9, 10, 11, 12}). Our SVD based approach always reported the accurate clusters whereas false sharing and wrong splitting were observed in FlowMate results.

Output clusters	AI	Interpretation
{1, 2}, {3}, {4, 5, 6, 8}, {7}, {9, 10, 11, 12}	0.833	2 errors: false splitting
{1, 2, 3}, {4, 5, 6}, {7, 8}, {9, 10, 11}, {12}	0.833	2 errors: false spitting
{1, 2, 3, 4, 5, 6, 7, 8, 10}, {9, 11, 12}	0.500	6 errors: false sharing
{1, 2, 3, 4}, {5, 6, 11}, {7, 8}, {9, 10, 12}	0.750	3 errors: 2 false splittings and 1 false sharing
{1, 2, 12}, {3, 11}, {4, 5, 6, 7, 8, 10}, {9}	0.667	4 errors: 1 false splitting and 3 false sharings
{1, 2, 10}, {3}, {4, 5, 6}, {7}, {8, 11}, {9, 12}	0.583	5 errors: 3 false splittings and 2 false sharings