Table 6.

Cross validation results of (CIS+RR) vs. (SP+PP) in terms of F-Measure, Precision, and Recall (± standard deviation) averaged on 3-folds.

	Identity	D	BC	CC	E	All-graphs
UNWEIGHTED
F-Measure	0.91 (±0.1)	0.68 (±0.02)	0.77 (±0.02)	0.78 (±0.02)	0.77 (±0.01)	0.83 (±0.01)
Precision	0.92 (±0.1)	0.7 (±0.01)	0.78 (±0.01)	0.78 (±0.01)	0.78 (±0.02)	0.83 (±0.01)
Recall	0.91 (±0.0)	0.68 (±0.02)	0.77 (±0.02)	0.78 (±0.02)	0.77 (±0.02)	0.83 (±0.01)
WEIGHTED
F-Measure	0.97 (±0.01)	0.99 (±0.0)	0.85 (±0.01)	0.83 (±0.02)	0.84 (±0.02)	0.92 (±0.01)
Precision	0.97 (±0.01)	0.99 (±0.0)	0.85 (±0.01)	0.84 (±0.02)	0.85 (±0.02)	0.92 (±0.01)
Recall	0.97 (±0.01)	0.99 (±0.0)	0.85 (±0.01)	0.83 (±0.02)	0.84 (±0.02)	0.92 (±0.01)

Rows report achieved results using unweighted graphs with unweighted features (upper) and using weighted graphs with weighted features (lower) [Degree (D), Betweenness Centrality (BC), Clustering Coefficient (CC), Local Efficiency (E), with all graph-metrics (all-graphs)] and without features (identity).