Machine Learning and the Cross-Section of Cryptocurrency Returns

Machine Learning and the Cross-Section of Cryptocurrency Returns

AuthorsNusret Cakici; Syed Jawad Hussain Shahzad; Barbara Będowska-Sójka; Adam Zaremba

Paper linkhttps://papers.ssrn.com/sol3/papers.cfm?abstract_id=4295427

Invest in 574 tradable cryptocurrencies using machine learning predictions from multiple models. Form weekly long-short quintile portfolios based on expected returns. Forecast combination reduces errors and enhances alpha generation.

Machine learning captures cryptocurrency return predictability from volatility, liquidity, and past performance. Combined model forecasts exploit mispricing and generate substantial alpha, making crypto ML strategies competitive with traditional equity markets.