Detecting dietary preference of social media users in China via sentiment analysis
Qingqing Zhou, Chengzhi Zhang · Proceedings of the Association for Information Science and Technology · 2017
ABSTRACT Dietary preferences are linked to all aspects of the human culture. Currently, researches on dietary preferences are mainly based on questionnaires, etc., which are mature and feasible. However, high cost, small scale and long timeframes are difficult to avoid. With the rapid development of social media, massive dietary reviews are shared in social media. Therefore, researches on users' dietary preferences by mining data from social media may overcome the disadvantages of traditional methods. In this paper, we use microblogs from weibo.com (one of the most popular social media platforms in China) to detect dietary preferences of social media users in China via sentiment analysis. Specifically, we compared four different aspect extraction methods and chose an optimal one to obtain aspects about dietary preferences. Secondly, sentiment polarities of the aspects and dishes are identified by sentiment classification. Empirical analysis on 3,975,800 microblogs presents that social media users in China are not satisfied with the overall status quo of dietary. In addition, experimental results show that semantic information is useful in extracting dietary aspects.