Research Article

The Comprehensive Contributions of Endpoint Degree and Coreness in Link Prediction

Table 2

AUC on the twelve benchmark networks with L = 100, where L denotes the number of the candidate links to measure the prediction accuracy in each data set. Every data point is an average over 30 independent realizing processes, and every point represents a random 90%–10% division of training set and testing set. The values in the parentheses indicate the corresponding optimal number of steps. All the present results represent the optimal cases by (if any) adjusting the coefficients.

AUCCNAARALPSRWCSRWHSRWSHIHCHIDCHI

USair0.9777710.9842430.9866110.9669170.988464 (4)0.987171 (6)0.987346 (5)0.987975 (5)0.987793 (5)0.988243 (5)
Yeast0.7369460.7370780.7370750.7428130.740707 (3)0.740543 (3)0.740576 (3)0.740644 (3)0.740553 (3)0.740622 (3)
Food0.6163910.6176810.6195210.8278790.749576 (15)0.745886 (15)0.745153 (15)0.747621 (15)0.763479 (5)0.765610 (5)
Power0.6796130.6797230.679680.7639820.888964 (15)0.888625 (15)0.888724 (15)0.888844 (15)0.925632 (15)0.947420 (15)
NS0.9902270.9903450.9903550.9939980.994864 (15)0.994783 (15)0.994802 (15)0.994836 (15)0.995283 (15)0.995562 (8)
Jazz0.9722420.9764380.9813340.9353420.981301 (2)0.981 (2)0.982014 (2)0.981992 (2)0.981740 (2)0.982131 (2)
E-mail0.8819550.8831620.8824710.9422830.947927 (15)0.946909 (15)0.947445 (15)0.947803 (15)0.954343 (8)0.955154 (7)
Slavko0.9640260.9659340.9656780.9590880.970067 (6)0.968883 (7)0.969502 (6)0.969898 (6)0.970442 (5)0.970861 (5)
UCsocial0.8130940.8174050.817640.9485160.939542 (15)0.936628 (15)0.937387 (15)0.938368 (15)0.947678 (11)0.948375 (9)
Infec0.9623180.9642230.9642090.9707870.977274 (8)0.976935 (10)0.977107 (9)0.977311 (9)0.979832 (5)0.980139 (5)
Eurosis0.9552690.956590.9560810.9784580.981863 (14)0.981535 (15)0.98186 (15)0.981694 (15)0.984722 (5)0.985111 (5)
CE0.9515450.9770890.979050.9323160.982963 (5)0.982176 (7)0.982329 (6)0.982663 (5)0.984026 (5)0.983915 (5)