Network reconstruction via density sampling

Applied Network Science

Table 1 Statistical indicators used to evaluate the performance of our sampled-based reconstruction method, for different cardinalities n of the known subset I. Results are shown together with the 95% confidence intervals (not shown whenever their difference affects the significant digits beyond the third one)

WTW	n=5 (CI 95%)	n=10 (CI 95%)	n=20 (CI 95%)	n=50 (CI 95%)	n=100 (CI 95%)
2000 - True positive rate	0.794 [0.772,0.816]	0.779 [0.765,0.793]	0.804 [0.796,0.812]	0.801 [0.797,0.806]	0.801 [0.799,0.804]
2000 - Specificity	0.700 [0.669,0.731]	0.742 [0.726,0.758]	0.721 [0.710,0.731]	0.728 [0.723,0.734]	0.729 [0.726,0.733]
2000 - Positive predicted value	0.796 [0.784,0.808]	0.810 [0.803,0.817]	0.799 [0.795,0.804]	0.802 [0.800,0.805]	0.802 [0.801,0.803]
2000 - Accuracy	0.755 [0.750,0.760]	0.763 [0.762,0.766]	0.769 [0.768,0.770]	0.771	0.771
2000 - Cosine similarity	0.712	0.712	0.712	0.712	0.712
e-MID
1999 - True positive rate	0.641 [0.601,0.673]	0.633 [0.614,0.653]	0.633 [0.620,0.646]	0.637 [0.623,0.643]	0.636 [0.632,0.640]
1999 - Specificity	0.839 [0.823,0.856]	0.856 [0.848,0.864]	0.860 [0.854,0.865]	0.860 [0.857,0.863]	0.861 [0.859,0.862]
1999 - Positive predicted value	0.623 [0.611,0.637]	0.632 [0.625,0.639]	0.633 [0.628,0.638]	0.632 [0.623,0.635]	0.633 [0.631,0.634]
1999 - Accuracy	0.785 [0.780,0.790]	0.795 [0.794,0.796]	0.798 [0.797,0.799]	0.799 [0.798,0.800]	0.799
1999 - Cosine similarity	0.810 [0.805,0.815]	0.814 [0.811,0.816]	0.816 [0.815,0.817]	0.817	0.817

The considered cardinalities n=5,10,20,50,100 correspond to percentages ranging from ≃2% to ≃50% of the total number of nodes. As reference values, the link density is c=0.578 for the WTW (in the year 2000) and c=0.274 for e-MID (in the year 1999)