Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Long-only cryptocurrency portfolio management by ranking the assets: a neural network approach

Zijiang Yang

TL;DR

The paper tackles crypto portfolio management by moving from asset-specific price prediction to cross-sectional ranking of multiple assets. It uses a neural-network-based method to predict the rank of future returns and constructs a long-only portfolio with weights derived from these ranks, incorporating a decay mechanism to reduce turnover. Backtests on 2020-2023 crypto data show the rank-based MLP variants achieving high risk-adjusted performance (Sharpe ≈1.0) and robust profitability even under transaction costs, outperforming traditional benchmarks and several ML baselines. These findings highlight the value of leveraging cross-asset relationships and rank-based targets in volatile crypto markets, with practical implications for portfolio construction under varying market regimes.

Abstract

This paper will propose a novel machine learning based portfolio management method in the context of the cryptocurrency market. Previous researchers mainly focus on the prediction of the movement for specific cryptocurrency such as the bitcoin(BTC) and then trade according to the prediction. In contrast to the previous work that treats the cryptocurrencies independently, this paper manages a group of cryptocurrencies by analyzing the relative relationship. Specifically, in each time step, we utilize the neural network to predict the rank of the future return of the managed cryptocurrencies and place weights accordingly. By incorporating such cross-sectional information, the proposed methods is shown to profitable based on the backtesting experiments on the real daily cryptocurrency market data from May, 2020 to Nov, 2023. During this 3.5 years, the market experiences the full cycle of bullish, bearish and stagnant market conditions. Despite under such complex market conditions, the proposed method outperforms the existing methods and achieves a Sharpe ratio of 1.01 and annualized return of 64.26%. Additionally, the proposed method is shown to be robust to the increase of transaction fee.

Long-only cryptocurrency portfolio management by ranking the assets: a neural network approach

TL;DR

The paper tackles crypto portfolio management by moving from asset-specific price prediction to cross-sectional ranking of multiple assets. It uses a neural-network-based method to predict the rank of future returns and constructs a long-only portfolio with weights derived from these ranks, incorporating a decay mechanism to reduce turnover. Backtests on 2020-2023 crypto data show the rank-based MLP variants achieving high risk-adjusted performance (Sharpe ≈1.0) and robust profitability even under transaction costs, outperforming traditional benchmarks and several ML baselines. These findings highlight the value of leveraging cross-asset relationships and rank-based targets in volatile crypto markets, with practical implications for portfolio construction under varying market regimes.

Abstract

This paper will propose a novel machine learning based portfolio management method in the context of the cryptocurrency market. Previous researchers mainly focus on the prediction of the movement for specific cryptocurrency such as the bitcoin(BTC) and then trade according to the prediction. In contrast to the previous work that treats the cryptocurrencies independently, this paper manages a group of cryptocurrencies by analyzing the relative relationship. Specifically, in each time step, we utilize the neural network to predict the rank of the future return of the managed cryptocurrencies and place weights accordingly. By incorporating such cross-sectional information, the proposed methods is shown to profitable based on the backtesting experiments on the real daily cryptocurrency market data from May, 2020 to Nov, 2023. During this 3.5 years, the market experiences the full cycle of bullish, bearish and stagnant market conditions. Despite under such complex market conditions, the proposed method outperforms the existing methods and achieves a Sharpe ratio of 1.01 and annualized return of 64.26%. Additionally, the proposed method is shown to be robust to the increase of transaction fee.

Paper Structure

This paper contains 23 sections, 4 equations, 2 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Research
  3. Methodology
  4. Problem Setting
  5. Existing trading algorithms
  6. Benchmark
  7. Follow the winner
  8. Follow the loser
  9. Pattern matching
  10. Machine learning based trading algorithms
  11. Feature generation
  12. Transforming the return to rank
  13. Machine learning algorithms
  14. Adding the decay
  15. Evaluation Metrics
  16. ...and 8 more sections

Figures (2)

  • Figure 1: Price series for the coins(normalized to 1 at the beginning; use log scale on y axis)
  • Figure 2: Cumulative wealth of the multi-layer perceptron(MLP) trading algorithm and its excessive return(alpha) with respect to the UCRP(benchmark)