Public/private tweet classification method to prevent posting to wrong accounts
Yuri Nozaki, Tetsuji Satoh · 2017
Corporations, municipalities, and governments have public Twitter accounts from which they send advertisements, publicity, and announcements. If such public account holders also have private accounts, posts might be sent from the wrong account. In this paper, we propose a classification method for public/private tweets using machine learning. As one classification feature, we use the tweet format and end-of-sentence expressions because these features are not confined to tweet content. We experimentally compared the classifiers built for multiple account holders and one general classifier.