Looking for Information Source from Online Social Network with Incomplete Observation

Online social network has become an important media for information spreading. To target the origin of rumors and well understand the cause of event, information source in social network should be found out. Unfortunately, because of the huge and complex structure of social network, it is nearly impossible to monitor the complete online social network. To solve this problem, the paper gives a method to look for information source with incomplete observation. The basis of the method is the correlation between activated time and network topology structure. Some experiments are executed and the effect of different sampling methods was analyzed as well. Experiments on BA, WS and Facebook show that the method can produce satisfied results with incomplete observation.