Inferring Probabilistic Contagion Models Over Networks Using Active Queries

Abhijin Adiga, Vanessa Cedeno-Mieles, Chris J. Kuhlman, Madhav V. Marathe, S. S. Ravi, Daniel J. Rosenkrantz, Richard E. Stearns
2018 Proceedings of the 27th ACM International Conference on Information and Knowledge Management - CIKM '18  
The problem of inferring unknown parameters of a networked social system is of considerable practical importance. We consider this problem for the independent cascade model using an active query framework. More specifically, given a network whose edge probabilities are unknown, the goal is to infer the probability value on each edge by querying the system. The optimization objective is to use as few queries as possible in carrying out the inference. We present approximation algorithms that
more » ... de provably good estimates of edge probabilities. We also present results from an experimental evaluation of our algorithms on several real-world networks.
doi:10.1145/3269206.3271790 dblp:conf/cikm/AdigaCKMRRS18 fatcat:lcm6xjtb6fdrpolf3gr26l6caq