earticle

영어

Since the end of 2017, cryptocurrencies have shown significant price fluctuations in the short and long term, drawing great interest from academia and industry by bringing unprecedented profits and losses to investors. This interest in cryptocurrency seems to be focused on realizing short-term gains using large price fluctuations rather than investments based on the essential value of assets such as stocks. Twitter data have been increasingly used as an information delivery means that has a significant impact on cryptocurrency prices. Several academic studies tried to predict cryptocurrency price fluctuations by analyzing social sentiments in Twitter contexts. We selected cryptocurrencies of major categories, collected tweets that mentioned each cryptocurrency on Twitter, and conducted sentiment analysis using various algorithms. Next, we collected cryptocurrency-related news published during the same period, classified each tweet into Markets, Business, Technology, and Policy categories based on the news data, and applied the sentiment analysis results. As a result, the tweet sentiment belonging to the market category had a correlation with the cryptocurrency price returns and there is a two-way causal relationship between tweet sentiment and cryptocurrency price returns.