Automated Detection of Chinese Government Astroturfers Using Network and Social Metadata

Authored by: Blake Miller

Awards: Best paper at the PolText 2016 conference in Dubrovnik, Croatia

Abstract: I present a method for automatically detecting pro-government astroturfers in China (colloquially referred to as the Fifty Cent Party), using comment metadata from a dataset of 45 million news media comments posted on 4 million news articles from 19 popular news websites in China. I estimate that approximately 15% of all comments made on these 19 news websites are made by government astroturfers. This method of comment propaganda detection is automated, and does not require manual human labeling. Instead, data are labeled according to metadata characteristic of the work procedures and behavioral patterns of government astroturfers. Models trained on these metadata predict posts from a leaked dataset of government astroturfers with as high as 94.1% accuracy. This method allows researchers timely access to government astroturfed commentary from China. Additionally, this method allows for prediction of astroturfers' bureaucratic affiliation using social network data, and can facilitate exploration of how this information control tactic varies both functionally and spatially.

Read Full Paper Here