多因子投资模式下的因子挖掘——基于舆情的因子研究与策略

Factor Mining Under Multi-factor Investment Mode——Factor Research and Strategy Based on Public Opinion

以东方财富个股股吧评论内容作为原始语料,结合自然语言处理技术构建显性舆情因子后发现,舆情因子与股票收益率具有较强的相关性,因子测试结果表明,使用舆情因子构建投资策略具有可行性。经将舆情因子加入多因子模型中形成指数增强策略,回测结果表明,结合舆情因子的多因子投资策略可以显著获得超过基准的超额收益。利用技术指标作为隐性舆情的替代,探讨隐性情绪指标在Barra模型中的利用以期获得更完善的情绪策略研究,结果表明结合隐性策略的Barra模型可以获得较显性策略且更稳健的回测效果。

It is found that public opinion factors have a good correlation with stock returns.The factor test results show that it is feasible to use public opinion factors to construct investment strategies.And the factor values are positively related with stock returns.This paper adds public opinion factors to the multi-factor model to form an index enhancement strategy.The results show that the multi-factor investment strategy combined with public opinion factors can significantly obtain excess returns that exceed the benchmark.This article also uses technical indicators as a substitute for recessive emotions,and explores the use of recessive emotion indicators in the Barra model to obtain a more complete emotional strategy research.The results show that the Barra model combined with recessive strategies can achieve more robustness than explicit strategies.