Abstract
AbstractThis paper proposes the use of social media as a proxy for financial information. Using an extended sample of 53,580,759 tweets and employing text analysis tools (Latent Dirichlet Allocation and Term Frequency–Inverse Document Frequency), we determine the information being exchanged on any given day. We train machine‐learning classifiers and forecast crypto price movements for more than 8000 cryptocurrencies and gauge market efficiency through successful forecasts based on public information. We propose various metrics of market efficiency for cryptocurrency assets and demonstrate that market efficiency is higher during the first 6 months after the Initial Coin Offering. We also examine the efficiency behavior of individual currencies during crisis periods.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
