| With the rapid development of social media,how to identify paid posts become more and more important for social media content analysis.In this research,we identify paid posts using feature words,building text based module,machine learning.We also leverage the relationship between multi-texts to build the identification module.In addition,we detect the unusual paid posters by account attributes and the accounts relationship.The characteristics of the paid posters could be from anywhere and it is very difficult to get enough high-quality text based training samples.The outstanding contribution in this research is that we use a variety of different technologies for different type of paid posters,at the same time,we develop a unique,successful paid posts method which is more suitable for Chinese social media environment and it is proved to be effective and operational by practice. |
