Does Big Data Matter?: Predicting Stock Returns using Online Stock Message Boards

초록

Can online stock message boards predict the direction of stock returns? We investigate whether these boards contain predictive power regarding the direction of stock returns. Moreover, we verify the effectiveness of machine learning techniques for extracting predictive information from such message boards. Finally, we aim to identify the informational content of online these message boards using the implications of microstructure theory. Our results suggest that online stock message boards do contain predictive information, which can be extracted using machine learning techniques. Furthermore, this information is closely linked to public/private information arriving at the market.

키워드