Causal Related Rumors Controlling in Social Networks of Multiple Information

There is a huge amount of information generated in online social networks, which is filled with a lot of rumors. The spread of a rumor often leads to the generation of a causal related rumor, and when users believe the first kind of rumor, the probability of being influenced by another causal related rumor is larger. Therefore, the influence probability will change with the process of rumor spreading. In this paper, we design the Causal Rumors Enhance Cascade (
$CREC$
) model to describe the spreading process of causal related rumors. Then we attempt to select a set of seed users that minimizes the number of users expected to be influenced by rumors, which we call the Causal Related Rumors Controlling (CRRC) problem. The main challenges of this problem are that the influence probability is constantly changing during the spread process, so the reverse sampling technique cannot be used, and the greedy mechanism is not suitable for massive-scale datasets. For the sake of overcoming these challenges and solving the problem, we put forward the Degree Trigonometric Metrology (DTM) algorithm, which uses the property of three-directed circles in the directed network to select seed nodes. Finally, experiments on three massive-scale datasets show that our algorithm outperforms the other algorithms.