基于影响力加权的在线投资者情绪对股票收益的影响
Effect of Influence-based Online Investor Sentiment on Stock Returns
王高山 1王越 1董宜麟 1张新1
作者信息
- 1. 山东财经大学管理科学与工程学院,山东济南 250014
- 折叠
摘要
为更好地捕捉市场总体情绪,本文在利用机器学习算法对在线投资者评论进行情感分类的基础上,考虑评论的阅读量、点赞量和评论量等信息,构建基于影响力加权的在线投资者情绪指数,并建立加权和未加权在线投资者情绪与股票收益的回归模型,在控制股票市值、账面市值比、Beta等变量基础上,发现在线投资者情绪对股票收益具有显著正向影响,且基于影响力加权的情绪指数相比未加权的情绪指数更能反映股票收益变化.这意味着在线评论影响力蕴含着对投资决策和市场监管有价值的信息.
Abstract
In order to better capture the overall market sentiment,this paper used machine learning algorithms to classify online investor comments and developed an influence-based investor sentiment index,taking into account information such as reading volume,number of likes,and number of comments.Furthermore,the paper developed regression models of weighted sentiment,unweighted sentiment,and stock returns.After controlling for variables such as stock market value,book-to-market ratio,and Beta,we found that online investor sentiment has a significant and positive impact on stock returns.Moreover,the influence-based sentiment index can better reflect changes in stock returns compared to the unweighted sentiment index,implying that the influence of online comments contains valuable information for investment decision-making and market regulation.
关键词
在线评论/影响力/机器学习/投资者情绪/股票收益Key words
Online comment/Influence/Machine learning/Investor sentiment/Stock returns引用本文复制引用
基金项目
教育部人文社会科学研究规划基金项目(22YJA630086)
出版年
2024
NETL
NSTL
万方数据