Paper: arXiv 2505.24831

Authors: Ruixue Jing, Ryota Kobayashi, Luis Enrique Correa Rocha

Abstract

The emerging cryptocurrency market presents unique challenges for investment due to its unregulated nature and inherent volatility. However, collective price movements can be explored to maximise profits with minimal risk using investment portfolios. In this paper, we develop a technical framework that utilises historical data on daily closing prices and integrates network analysis, price forecasting, and portfolio theory to identify cryptocurrencies for building profitable portfolios under uncertainty. Our method utilises the Louvain network community algorithm and consensus clustering to detect robust and temporally stable clusters of highly correlated cryptocurrencies, from which the chosen cryptocurrencies are selected. A price prediction step using the ARIMA model guarantees that the portfolio performs well for up to 14 days in the investment horizon. Empirical analysis over a 5-year period shows that despite the high volatility in the crypto market, hidden price patterns can be effectively utilised to generate consistently profitable, time-agnostic cryptocurrency portfolios.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 6.0/10 (capped from 7.0: transaction costs not modelled; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 7.0)

Why this score: The paper integrates advanced network theory (Louvain algorithm), consensus clustering, and ARIMA forecasting, representing moderate-to-high mathematical density; its empirical rigor is supported by a 5-year backtest, specific investment horizon (14 days), and multiple data-driven validation steps.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Identify Stable Crypto Clusters for<br>Profitable, Low-Risk Portfolios"] --> B{"Data Input<br>5 Years Daily Closing Prices"}
  B --> C["Network Analysis<br>Louvain Algorithm for Community Detection"]
  C --> D["Consensus Clustering<br>Identify Robust & Temporally Stable Clusters"]
  D --> E["Selection & Prediction<br>Select Assets & Forecast with ARIMA<br>14-Day Horizon"]
  E --> F["Portfolio Theory Application<br>Construct Optimized Portfolio"]
  F --> G["Key Outcome<br>Consistently Profitable<br>Time-Agnostic Portfolios"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes an integrated framework that combines network analysis, ARIMA price forecasts, consensus clustering, and modern portfolio theory to build profitable cryptocurrency portfolios.

Source sentence
This paper introduces an integrated framework that combines network analysis, price forecasting, and portfolio theory to identify stable groups of highly correlated cryptocurrencies for profitable portfolio construction.

Assumes 5

  • Cryptocurrency daily closing prices are stationary over the sample period. statistical
    Source sentence
    Augmented Dickey‑Fuller (ADF) test results confirm that the log returns ri,t of all examined cryptocurrencies are stationary, rejecting the null hypothesis at p < .001.
  • Missing price data for more than ten days leads to exclusion of a cryptocurrency. data
    Source sentence
    Cryptocurrencies with missing prices for more than ten days are excluded.
  • ARIMA models provide useful forecasts of future log returns. model
    Source sentence
    By incorporating predicted returns into the correlation network, the proposed P(ARIMA) strategy improves the detection of forward‑looking interdependencies and provides better‑performing portfolios over extended holding horizons.
  • Trading can be executed at daily closing prices with manageable slippage and bid‑ask spreads. execution
    Source sentence
    These returns also provide a buffer against typical cryptocurrency trading costs (approximately 0.10%–0.20% per side on major exchanges), such that at shorter horizons the strategies can remain profitable after transaction fees, provided that execution slippage and bid–ask spreads are well managed.
Show 1 more
  • The cryptocurrency market exhibits high volatility but the identified clusters remain persistent across regimes. market
    Source sentence
    Central to our methodology is the identification of persistent clusters of highly interdependent cryptocurrencies via consensus clustering. These clusters capture structural co‑movement patterns that remain stable across market regimes.

Data

Backtest sample: 2017-11 – 2022-04 · cryptocurrency market · crypto · daily

  • Cryptocurrency price data · various online sources market-prices public
    15.11.2017 to 15.04.2022 · 157 cryptocurrencies
    Source sentence
    The dataset comprises daily closing market prices (at midnight) of 5,450 cryptocurrencies

Costs

Unclear - sources disagree The platform classifier and the paper-text extraction give different answers, so no value is asserted.

Platform classifier: Not modeled

Paper text extraction: Partial - Typical trading costs are about 0.10%–0.20% per side; the paper discusses that returns can remain profitable after these fees but does not embed them in headline performance numbers.

Source sentence (the authors' words)
These returns also provide a buffer against typical cryptocurrency trading costs (approximately 0.10%–0.20% per side on major exchanges), such that at shorter horizons the strategies can remain profitable after transaction fees, provided that execution slippage and bid–ask spreads are well managed.

Finds 5

  • Predictive consensus‑clustering portfolios deliver consistently positive and stable performance up to a 14‑day holding horizon, with favourable gain‑loss asymmetry and tighter tail‑risk control. performance
    Source sentence
    Our empirical results show that predictive consensus‑clustering portfolios maintain consistently positive and stable performance up to a 14‑day horizon, exhibit favourable gain‑loss asymmetry, and achieve tighter tail‑risk control.
  • The P(ARIMA) strategy outperforms the baseline historical‑only strategy, especially for longer holding periods. performance
    Source sentence
    Our portfolio strategies that incorporate time‑series prediction of log returns outperform the Baseline strategy relying only on historical data, particularly over longer horizons.
  • Consensus clustering frequently identifies Bitcoin as part of the largest and most persistent clusters, reflecting its role as a market‑wide reference asset. statistical
    Source sentence
    Our consensus clustering frequently identifies Bitcoin (BTC) as part of the largest and most persistent clusters, reflecting its role as a dominant asset and market‑wide reference point.
  • Incorporating predicted returns strengthens gain‑loss asymmetry, reduces tail‑risk exposure, and maintains profitability as horizons increase. performance
    Source sentence
    prediction‑enhanced clustering systematically strengthens gain‑loss asymmetry, reduces tail‑risk exposure, and maintains profitability as horizons increase.
  • High computational costs of ARIMA forecasting may not justify adoption compared to simpler strategies. negative
    Source sentence
    On the other hand, the high computational costs of forecasting models may not justify their adoption compared to our simpler strategies.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 37 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (71)

  1. Hierarchical structure in financial markets (1999)
  2. Analyzing and modeling real-world phenomena with complex networks: a survey of applications (2011)
  3. Sampling of temporal networks: Methods and biases (2017)
  4. The global migration network of sex-workers (2022)
  5. Do accrual minimise (maximise) stock risk (return)?: E vidence from I ndonesia (2017)
  6. Portfolio Selection (2008)
  7. PyPortfolioOpt: portfolio optimization in Python (2021)
  8. A network-based strategy of price correlations for optimal cryptocurrency portfolios — → 2304.02362
and 63 more
  1. Fast unfolding of communities in large networks (2008)
  2. Price prediction of cryptocurrency: an empirical study (2019)
  3. Crypto Currency Market Price Prediction Using Data Science Process (2022)
  4. A deep learning-based cryptocurrency price prediction scheme for financial institutions (2020)
  5. Cryptocurrency-portfolios in a mean-variance framework (2019)
  6. Community detection in cryptocurrencies with potential applications to portfolio diversification (2023)
  7. Using sentiment analysis to predict interday Bitcoin price movements (2018)
  8. The dynamics of cryptocurrency market behavior: sentiment analysis using Markov chains (2022)
  9. Cryptocurrency portfolio management with deep reinforcement learning (2017)
  10. Using Principal Component Analysis on Crypto Correlations to Build a Diversified Portfolio (2021)
  11. Maximizing portfolio profitability during a cryptocurrency downtrend: A Bitcoin Blockchain transaction-based approach (2023)
  12. Crypto currency portfolio allocation Using Machine Learning (2021)
  13. Cryptocurrencies & the challenge of global governance (2018)
  14. The impact of COVID-19 on cryptocurrency markets: A network analysis based on mutual information (2022)
  15. Portfolio Construction: A Network Approach (2023)
  16. Prospects for crypto-currency and blockchain technologies in financial markets (2018)
  17. Evolutionary dynamics of the cryptocurrency market (2017)
  18. Cryptocurrency market contagion: Market uncertainty, market complexity, and dynamic portfolios (2019)
  19. The Emergence of New Blockchains and Crypto-Assets (2022)
  20. Investigating the problem of cryptocurrency price prediction: a deep learning approach (2020)
  21. Cryptocurrency price prediction using traditional statistical and machine-learning techniques: A survey (2021)
  22. Neural network techniques for financial performance prediction: integrating fundamental and technical analysis (2004)
  23. Cryptocurrencies as a financial asset: A systematic analysis (2019)
  24. Cryptocurrencies: Key risks and challenges (2022)
  25. Risk parity portfolio vs. other asset allocation heuristic portfolios (2011)
  26. Applying Artificial Intelligence in Cryptocurrency Markets: A Survey (2022)
  27. Equity portfolio diversification (2008)
  28. Optimal asset allocation and risk shifting in money management (2007)
  29. Value investing: investing for grown ups? (2012)
  30. Do mutual funds time the market? Evidence from portfolio holdings (2007)
  31. Asset management: A systematic approach to factor investing (2014)
  32. Quantitative equity investing: Techniques and strategies (2010)
  33. A new look at the statistical model identification (1974)
  34. Dynamics of market correlations: Taxonomy and portfolio analysis (2003)
  35. Visualization of a stock market correlation matrix (2014)
  36. Stock market networks: The dynamic conditional correlation approach (2012)
  37. Are simple technical trading rules profitable in bitcoin markets? (2024)
  38. Consensus clustering in complex networks (2012)
  39. Sampling of temporal networks: Methods and biases (2017)
  40. Analyzing and modeling real-world phenomena with complex networks: a survey of applications (2011)
  41. TensorFlow: a system for Large-Scale machine learning (2016)
  42. Financial and market risks of bitcoin adoption as legal tender: evidence from El Salvador (2024)
  43. A well-conditioned estimator for large-dimensional covariance matrices (2004)
  44. Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models (2002)
  45. Dynamic equicorrelation (2012)
  46. L 1 Regularization for High-Dimensional Multivariate GARCH Models (2024)
  47. Financial time series forecasting with deep learning: A systematic literature review: 2005–2019 (2020)
  48. Quantitative risk management: concepts, techniques and tools-revised edition (2015)
  49. Measuring market risk (2007)
  50. Measuring systemic risk (2017)
  51. An introduction to omega (2002)
  52. Trading systems and methods (2013)
  53. The evaluation and optimization of trading strategies (2011)
  54. A tool for filtering information in complex systems (2005)
  55. Correlation filtering in financial time series (2005)
  56. Mastering Bitcoin: Programming the open blockchain (2023)
  57. Bitcoin and cryptocurrency technologies: a comprehensive introduction (2016)
  58. Bitcoin, gold and the dollar–A GARCH volatility analysis (2016)
  59. Some stylized facts of the Bitcoin market (2017)
  60. Structure and dynamics of financial networks by feature ranking method (2021)
  61. Cryptocurrency pump-and-dump schemes (2018)
  62. Risk factors in cryptocurrency pricing (2025)
  63. When do improved covariance matrix estimators enhance portfolio optimization? An empirical comparative study of nine estimators (2011)

1 of 71 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.