Cryptocurrency Valuation and Machine Learning

As in stock and commodity markets, investors and financial analysts are constantly seeking a reliable indicator to assess the price of cryptocurrency assets. However, there are currently no convincing proxies for the fundamentals of cryptocurrency assets. We propose a new methodology for cryptocurrency valuation by considering insights from classic Monetary Economics and Machine Learning in data science. We then illustrate and evaluate the methodology using big data of Bitcoin as a case study. The results indicate a significant improvement from other existing approaches in explaining asset valuation and informing investment strategies. Moreover, we propose an automated trading strategy based on the simple indicator alone for further industry applications and document its outstanding performance in comparison to the conventional strategies.