Link prediction for ex ante influence maximization on temporal networks

Applied Network Science

Table 1 Summary statistics and temporal network measures for datasets under consideration

Dataset	n	m	\({\bar{p}}\)	T(p)	fNT	fLT	FNT	FLT	DA
Synthetic	500	5174	0.002	20(10)	1.00	0.53	0.75	0.00	0.00
Reality	64	722	0.024	24(20)	0.97	0.39	0.52	0.00	\(-0.17\)
Email4	168	3250	0.038	39(30)	0.98	0.88	0.78	0.12	\(-0.30\)
HS 1	312	2242	0.005	20(16)	0.98	0.74	0.39	0.02	0.09
Hospital	75	1139	0.052	16(12)	0.83	0.64	0.29	0.02	\(-0.18\)
Office	92	755	0.042	7(6)	0.98	0.69	0.63	0.05	\(-0.06\)
CopenB	703	21,318	0.002	100(90)	0.94	0.64	0.27	0.00	0.08
College	1899	13,838	0.000	50(40)	0.93	0.92	0.02	0.00	\(-0.19\)

n: number of nodes; m: number of unique links; \({\bar{p}}:\) average edge density across time steps; T(p): number of time steps (number used for training time); fNT: fraction of nodes present at half the sampling time; fLT: fraction of unique links present at half the sampling time; FNT: fraction of nodes present in the first and last 5% of the sampling time; FLT: fraction of unique links present in the first and last 5% of the sampling time; DA: degree assortativity