Online public opinion risk warning based on bayesian network modeling

The state of online public opinion as the most important index which can measure the social conditions and public opinion has caught more and more attention in managerial practices and academic research. Aiming at the online public opinion which has complex object and action, hard predicting interactions relationships and hard measuring interactions degree, the authors present a new method which combine the dynamic Bayesian network modeling method to situation assessment of the online public opinion basing on the three characteristic of dynamic Bayesian network—the expression ability of complex relationship, the expression ability of uncertainties of the probabilities, and the inferential capability of causal relationship. Through the study of expert experience and the simulating of key indicator data, we build the situation assessment model of the online public opinion and forecast the situation assessment based on 60 cases. © 2016, UK Simulation Society. All rights reserved.