Credible Influence Analysis in Mass Media Using Causal Inference

The mass media has recorded major events around the world for a long time, which is very helpful in describing the dynamic changes in all aspects of human society, including the analysis of national influence using news data. Due to the publicity and significance of mass media, the results of influence analysis must be reliable. However, the current most influence analysis methods are mainly concentrated on social media networks and cannot simply be transferred to mass media. Due to the causality as the main driving factor of influence, we introduced the causal inference method convergent cross mapping, combined with the existing general influence analysis method, proposed a credible influence analysis method in mass media. This method can filter out non-causal influences, making the results more credible. We conducted experiments on the GDELT datasets, and the results proved the effectiveness and reliability of the proposed credible influence analysis in mass media.