摘要
研究新闻对股票市场的影响是当前研究的热点之一,本文使用宏观财经新闻的话题来研究财经新闻对股票收益率的影响。本文使用LDA方法对宏观财经新闻的话题进行提取,并研究新闻话题分布对于行业板块股票收益率的影响,实证结果证实了中国股票市场的"媒体效应"。在此基础上,本文探讨了行业收益率和新闻话题之间的关系,并选取了制造业和金融业两个代表性行业,通过财经新闻话题分布的变化来预测两个行业收益率的情况,策略的结果表明通过财经新闻话题分布来构建预测模型可以获得超额的收益率。
The impact of news on the stock market is one of the recent hot research topics, this paper aims to investigate whether latent topics in financial news can impact the return of the stock market. By introducing text mining techniques and Latent Dirichlet Allocation ( LDA ) classification method, we extract topics ’ distribution from comprehensive macroeconomics news and research the impact of financial news on the whole market. Then the relationship of financial topics and return in industry level are further discussed. The empirical study confirms that the distribution of financial topics can affect the distribution of sector return. Accordingly an investment strategy of sector allocation based on topics* distribution is constructed and an obviously higher return can be obtained comparing with random strategy.
作者
龙文
毛元丰
管利静
崔凌逍
Long Wen;Mao Yuanfeng;Guan Lijing;Cui Lingxiao(School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190;Research Center on Fictitious Economy & Data Science, Chinese Academy of Sciences, Beijing 100190;Key Laboratory of Big Data Mining & Knowledge Management, Chinese Academy of Sciences, Beijing 100190)
出处
《管理评论》
CSSCI
北大核心
2019年第5期18-27,共10页
Management Review
基金
国家自然科学基金项目(71771204)