A data-driven approach to forecasting Bitcoin price using on-chain metrics

Cryptocurrencies represent a rapidly expanding market, every day more and more people are looking for new investment opportunities and new strategies to identify trends and their prices. Most investors mainly use technical analysis to operate on the markets. Bitcoin is one of the most valuable cryptocurrencies and was introduced in 2008 by Satoshi Nakamoto with the introduction of the blockchain, considering the growing value based on blockchain, today bitcoin is considered to large extent. In addition to revolutionizing the idea of money, the blockchain has opened the door to the study of new metrics to be able to identify the market direction of a specific cryptocurrency. These metrics are called chain metrics and are mainly based on the study of transactions and tokenomics of the blockchain itself. The following research aims to identify and couple on-chain metrics with machine learning and deep reinforcement learning models, in order to predict the price of bitcoin for short and long term price models.