Link Prediction Methods and Their Accuracy for Different Social Networks and Network Metrics

Table 7

Pearson Correlation of prediction methods accuracy and network metrics.

	CN	JI	PA	AA		Cosine	Sor	HPI	HDI	LHN	Average
GCC	0.68	0.79	0.05	0.68	0.47	0.80	0.81	0.73	0.74	0.27	0.60
ACC	0.75	0.68	0.43	0.76	0.39	0.70	0.65	0.67	0.68	0.30	0.60
Network Density	0.52	0.58	0.18	0.52	0.09	0.61	0.61	0.48	0.52	−0.12	0.40
Gini	0.45	0.30	0.94	0.46	0.49	0.29	0.25	0.36	0.37	0.57	0.45
Diameter	−0.67	−0.61	−0.77	−0.68	−0.51	−0.61	−0.52	−0.61	−0.63	−0.39	−0.60
ASP	−0.63	−0.55	−0.79	−0.65	−0.29	−0.57	−0.52	−0.52	−0.56	−0.18	−0.53

This table shows the correlation between prediction methods accuracy and network metrics calculated with Pearson’s linear correlation coefficient. The number is within the range of where 1 is completely positive correlation, 0 is no correlation, and is completely negative correlation.