Towards Exploring the Influence of Community Structures on Information Dissemination in Sina Weibo Networks

Algorithm 2

Null model of IPT-based information propagation analysis. Firstly, a randomized copy of the IPT is generated; then, the corresponding R-value and Z-value in the extended significant profile are calculated by equations (6)–(8); finally, the obtained R-value and Z-value are drawn on the same plane, so as to acquire the for analyzing the laws and patterns of information propagation.

Input:, an IPT.
Output: the extended significant profile () of .
(1)	The randomized copy of is generated by NMRC-IPT detailed in Algorithm 1.
(2)	Community detection in and by utilizing the method detailed in Section 4.2. And, the community partition of and is and , respectively.
(3)	for each do
(4)	The R-value of is calculated by equation (6).
(5)	.
(6)	The Z-value of is calculated by equation (8).
(7)	.
(8)	end for
(9)	for each do
(10)	The R-value of is calculated by equation (6).
(11)	.
(12)	The Z-value of is calculated by equation (8).
(13)	.
(14)	end for
(15)	The mean value of , , and is calculated as , , , and , respectively.
(16)	The R-value and Z-value of degree correlation and the R-value and Z-value of the cascade rate of IPPs are also calculated, and plotting them on a same plan figure together with , , , and , as shown in Figure 3, i.e., the extended significant profile.
(17)	return the ESP calculated in 16.