| Summary: | A series of source locating algorithms have been brought up in recent studies. However,
less work has been done to compare their performance on online social networks as
well as the performance on multiple source detection problems. This work aims to give
a systematic view of the source location problem, especially for online social networks.
In this work, we first briefly discuss the property of online social networks. As the first
step to locating source is to describe the spreading of the rumor, we studied the SI model
and spreading delay models for rumor spreading. We also show that the SI model and
the geometric delay model are equivalent. Then, we classify the algorithms we are
going to study into 3 categories, maximum likelihood estimator, network centrality
measurement and time reverse algorithm. We take Netsleuth as an example of
maximum likelihood estimator and analyze its proof in detail. Finally, we use both
simulation and real rumor spreading data to test the performance of these algorithms.
Keywords: Source location, rumor spreading, online social networks
|
|---|