题目：Predicting Stock Price Movement Using Social Network Analytics: Posts are Sometimes Less Useful
报告人：李皖昀助理教授
时间：4月12日（周五）10:00-11:30
地点：Room 501, Jiageng Building 2/嘉庚二号楼501教室
报告内容简介：Recent studies have used social network data to predict stock price movements. We suggested that this method may not consistently perform effectively during instances of social contagion. We proposed capturing individuals’ conversation characteristics, which model interactions in which individuals respond to each other’s posts and expand on ideas, in addition to overall network characteristics. Drawing on social contagion theory, we identified three conversation conditions—argument similarity, sentiment similarity, and conversation size—and associated these conditions with the likelihood of an abrupt change in stock price. We developed three hypotheses and tested their associations with an abrupt change in stock price. Our social network data came from StockTwits. We extracted 18 million posts for 859 initial public offerings launched between 2008 and 2017. Our empirical results showed that a network characterized by large conversations and posts that have similar arguments or sentiments is positively associated with the likelihood of an abrupt change in stock price in the subsequent week. Theoretically, our study introduced a level of measure (conversation) that has been overlooked and contributed to theorizing the conversation conditions associated with an abrupt change in stock price, which manifests when market individuals fail to predict stock prices accurately. In practical terms, our findings led us to advise practitioners against routinely relying on social network data for making predictions.
报告人简介：李皖昀，现任厦门大学财务管理与会计研究院助理教授，博士毕业于澳大利亚国立大学商学院。其研究兴趣主要包括财务会计，大数据与数据分析，会计信息系统。