Table 1 Describes the networks’ structural properties, where V represents the number of vertices, E represents the number of edges, \(Gc_V\%\) and \(Gc_E\%\) denote the percentage of vertices and edges in the giant component, respectively, \(\langle k \rangle \) represents the average node degree, \(\beta _{th}\) indicates the epidemic threshold of a network as measured by largest eigenvalue, \(p_c\) is the percolation threshold, given by Eq. (4), and \(tune_1\) and \(tune_2\) denote the best tunable parameters selected for the proposed ngsc method.
From: Best influential spreaders identification using network global structural properties
Classification | Network | V | E | \(\beta _{th}\) | \(p_c\) | \(Gc_V\%\) | \(Gc_E\%\) | \(\langle k \rangle \) | \(\langle k^2 \rangle \) | \(tune_1\) | \(tune_2\) |
|---|---|---|---|---|---|---|---|---|---|---|---|
Social | PolBlogs61 | 3982 | 6803 | 0.011 | 0.012 | 96.12 | 99.44 | 27.31 | 2219.35 | 0.9 | 0.4 |
Ego-Facebook60 | 4039 | 88,234 | 0.036 | 0.009 | 99.9 | 99.9 | 43.69 | 4656.14 | 0.9 | 0.2 | |
Advogato53 | 6541 | 51,127 | 0.011 | 0.012 | 77.08 | 97.07 | 15.633 | 1255.84 | 0.9 | 0.4|0.2 | |
Brightkite63 | 58,228 | 214,078 | 0.0098 | 0.016 | 99.97 | 99.99 | 7.35 | 468.42 | 0.9 | 0.6 | |
Epinions62 | 75,879 | 508,837 | 0.004 | 0.005 | 99.99 | 100 | 13.412 | 1966.47 | 0.9 | 0.2 | |
Collaboration | Netscience52 | 1589 | 2742 | 0.052 | 0.168 | 25.91 | 33.32 | 3.74 | 26.05 | 0.2 | 0.9 |
CA-GrQc64 | 5242 | 14,496 | 0.011 | 0.063 | 92.29 | 96.44 | 5.53 | 93.25 | 0.2 | 0.9 | |
CA-HepTh64 | 9877 | 25,998 | 0.016 | 0.087 | 99.41 | 99.74 | 5.26 | 65.89 | 0.4 | 0.9 | |
CA-CondMat64 | 23,133 | 93,497 | 0.013 | 0.047 | 98.35 | 98.71 | 8.08 | 178.19 | 0.2 | 0.9 | |
Citation | Cora66 | 23,166 | 91,500 | 0.031 | 0.044 | 65.27 | 77.74 | 7.70 | 182.30 | 0.2 | 0.9 |
Cit-HepTh65 | 27,770 | 352,807 | 0.013 | 0.009 | 97.62 | 98.87 | 25.37 | 2697.53 | 0.9 | 0.2 | |
Neural | Celegans67 | 2325 | 5110 | 0.037 | 0.026 | 99.71 | 99.88 | 8.94 | 358.49 | 0.9 | 0.4 |