Twitter Homophily: Network Based Prediction of User’s Occupation

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2633–2638 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 2633 Twitter Homophily: Network Based Prediction of User’s Occupation Jiaqi Pan∗ University of Electronic Science and Technology of China jiaqi.pan1019@gmail.com Rishabh Bhardwaj∗ National University of Singapore rishabhbhardwaj 15@gmail.com Wei Lu Singapore University of Technology and Design luwei@sutd.edu.sg Hai Leong Chieu DSO National Laboratories chaileon@dso.org.sg Xinghao Pan DSO National Laboratories pxinghao@dso.org.sg Ni Yi Puay DSO National Laboratories pniyi@dso.org.sg Abstract In this paper, we investigate the importance of social network information compared to con- tent information in the prediction of a Twitter user’s occupational class. We show that the content information of a user’s tweets, the pro- file descriptions of a user’s follower/following community, and the user’s social network pro- vide useful information for classifying a user’s occupational group. In our study, we ex- tend an existing dataset for this problem, and we achieve significantly better performance by using social network homophily that has not been fully exploited in previous work. In our analysis, we found that by using the graph convolutional network to exploit social ho- mophily, we can achieve competitive perfor- mance on this dataset with just a small fraction of the training data. 1 Introduction Twitter (http://twitter.com) is a microblogging ser- vice launched in 2006, where, a user can pub- lish messages with up to 280 characters, called “tweets”. Unlike many other social networking platforms, such as Facebook and LinkedIn, Twitter does not provide structured fields for users to fill in personal information. However, a user can write a 160-character-long small public summary about itself called a “Bio”. Besides linguistic informa- tion from tweets and Bios, online social media is a rich source of network information. People’s personal networks are homogeneous, i.e., friends share more attributes such as race, ethnicity, re- ligion, and occupation–known as the homophily principle (McPherson et al., 2001). Such network information has been utilized in friend recommen- dation (Guy et al., 2010), community detection ∗Equal Contribution; work performed while both authors were visiting Singapore University of Technology and Design (SUTD). T1: Day at the races. T2: The new pitch era starts here. Bio Tweets Groundsman. Wolves fan. Horse racing enthusiast. A Commercial and Domestic Grounds Maintenance company T1: Pitch at wembley looks great. T2: Spurs will have best stadium in uk. Follow Figure 1: User and Network information on Twitter Mi- croblog. (Yang and Leskovec, 2013), etc. Figure 1 shows two users connected on Twitter. By looking at their Bio and tweets, it can be inferred that these users share the same occupational interest. Profiling users can enhance service quality and improve product recommendation, and hence is a widely studied problem. User occupational class prediction is an important component of user pro- filing and a sub-task of user demographic fea- ture prediction. Existing approaches to predicting Twitter users’ demographic attributes explore, se- lect, and combine various features generated from text and network to achieve the best predictive per- …

Content truncated to 500 words