Portfolio Optimization for Cointelated Pairs: Financial Mathematics or Machine Learning?
20 Pages Posted: 21 Sep 2017 Last revised: 30 Jul 2019
Date Written: May 9, 2019
We investigate the problem of dynamic portfolio optimization in continuous-time, finite-horizon setting for a portfolio of two stocks. The stocks follow the Cointelation model recently introduced , . The proposed optimization methods are twofold. First, in what we call a Dynamics Switching approach, we compute the optimal weights using the mean- variance criterion and the power utility maximization. We show that dynamically switching between these two optimal strategies by introducing a triggering function can further improve the portfolio returns. We contrast this with the Machine Learning clustering methodology inspired by the band-wise Gaussian mixture model , . Though both of these methods have their benefits, we show that Machine Learning approach yields better result most of the time.
Keywords: Cointelation, Portfolio Optimization, Mean- Variance Criterion, Power Utility Maximization, Band-wise Gaussian Mixture, LSTM, Partial Differential Equation, Deep Learning, Cryptocurrency, Bitcoin, Altoin, Pairs Trading
Suggested Citation: Suggested Citation