Research on Trust Propagation Models in Reputation Management Systems

Table 1

Simulation configuration.

Network scale	1000
Number of malicious nodes	13.4%
Number of pretrust nodes	3%
% of spy users	3.4%
% of service requests in which good user i provides unsatisfied services	0%–10%
% of service requests in which malicious user i provides unsatisfied service	Varied in threat models
% of spy users gives honest ratings	10%, 30%, 50%, 70%
Decay factors	0.1, 0.3, 0.5, 0.7
Threshold value of conditional propagation	0.005, 0.5
Rating mechanism	Multiscale rating mechanism
Similarity algorithm	[8] [Su, et al. 2013]
Local trust algorithm	[5] [Kamvar, et al. 2003]
Initialization strategy	Varied in propagation models
Jump strategy	Varied in propagation models
Queries issued by each node	Uniform distribution 50–100.
Number of cycles	30