【Source】INFORMATION SYSTEMS FRONTIERS
【Abstract】Virtual investment community has become an important information source for investors. This study contributes to the related literature by investigating the endogenous interplay between investor activity on the virtual investment community and the market trading dynamics using a vector autoregressive framework to analyze an hourly dataset collected from the Bitcoin market. The main results suggest that the sentiment and the posting frequency of virtual investment community messages are largely driven by the past market outcomes, but they provide limited value-relevant information for future price prediction. It is also demonstrated that when investors express conflicting opinions, or when their discussions exhibit a lack of diversity, their incentive to trade decreases, resulting in low trading volume. Theoretical contributions and practical implications are discussed.
【Keywords】Social media analytics; Financial technology; Opinion mining; Topic modeling; Bitcoin
评论