基于文本情感的投资者情绪与股指关系研究  

Research on the Relationship Between Investor Sentiment and Stock Index Based on Textual Emotion

作　　者：李珊[1] 丁宇 陈妙苗 郑晨 LI Shan;DING Yu;CHEN Miaomiao;ZHENG Chen(College of Economics and Management,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)

机构地区：[1]南京航空航天大学经济与管理学院,南京211106

出　　处：《河南科学》2021年第5期829-837,共9页Henan Science

基　　金：中央高校基本科研业务费专项基金资助项目(NJ2019023);教育部课题基金资助项目(15YJC630122);国家社会科学基金资助项目(17BGL055)。

摘　　要：为了研究我国散户投资者情绪与股指之间的相互关系,以东方财富网"上证指数"股吧的用户发帖为样本数据,利用Bi-LSTM模型对发帖内容进行文本情感分类,在此基础上设计投资者直接情绪指标,并结合间接情绪指标构建新的综合投资者情绪指数.利用近5年的数据进行实证检验,结果表明,散户投资者的情绪与股指之间存在着长期、正向且增强的相互作用,投资者情绪受到股指的影响要大于其对股指施加的影响.研究结果证明,同时包含直接情绪因素和间接情绪因素的情绪指数,会有更好的股指拟合效果.In order to study the relationship between retail investor sentiment and stock index,the Bi-LSTM model is used to classify the textual emotions of the posting contents collected from the east wealth net.On this basis,the investor direct sentiment index is designed.Then,a new comprehensive investor sentiment index is constructed by combining the investor direct sentiment index and traditional indirect sentiment index.Using the data of the last five years to conduct empirical test,the results show that there is a long-term,positive and enhanced interaction between retail investor sentiment and the stock index,and the influence of investor sentiment on the stock index is greater than its influence on the stock index.Moreover,this research results prove that the emotional index containing both direct and indirect emotional factors will have a better stock index fitting effect.

关 键 词：投资者情绪指数 直接情绪 上证股指 文本情感分类 Bi-LSTM模型 

分 类 号：F832.5[经济管理—金融学]