Paper: arXiv 2308.01013

Authors: Anoop C, Neeraj Negi, Anup Aprem

Abstract

Identifying the structural dependence between the cryptocurrencies and predicting market trend are fundamental for effective portfolio management in cryptocurrency trading. In this paper, we present a unified Bayesian framework based on potential field theory and Gaussian Process to characterize the structural dependency of various cryptocurrencies, using historic price information. The following are our significant contributions: (i) Proposed a novel model for cryptocurrency price movements as a trajectory of a dynamical system governed by a time-varying non-linear potential field. (ii) Validated the existence of the non-linear potential function in cryptocurrency market through Lyapunov stability analysis. (iii) Developed a Bayesian framework for inferring the non-linear potential function from observed cryptocurrency prices. (iv) Proposed that attractors and repellers inferred from the potential field are reliable cryptocurrency market indicators, surpassing existing attributes, such as, mean, open price or close price of an observation window, in the literature. (v) Analysis of cryptocurrency market during various Bitcoin crash durations from April 2017 to November 2021, shows that attractors captured the market trend, volatility, and correlation. In addition, attractors aids explainability and visualization. (vi) The structural dependence inferred by the proposed approach was found to be consistent with results obtained using the popular wavelet coherence approach. (vii) The proposed market indicators (attractors and repellers) can be used to improve the prediction performance of state-of-art deep learning price prediction models. As, an example, we show improvement in Litecoin price prediction up to a horizon of 12 days.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced mathematical concepts including dynamical systems, Lyapunov stability analysis, Bayesian inference with Gaussian Processes, and potential field theory, indicating high mathematical density. Empirically, it validates its model using historical cryptocurrency data across multiple crash periods, compares results with established wavelet coherence methods, and demonstrates practical improvements in prediction tasks, showcasing significant implementation and backtest-ready analysis.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Characterize Crypto Market Dynamics &<br>Structural Dependency for Portfolio Management"] --> B["Data Input:<br>Historic Cryptocurrency Price Data<br>e.g., Bitcoin, Litecoin"]
  B --> C["Methodology: Bayesian Framework<br>• Potential Field Theory<br>• Gaussian Process<br>• Dynamical Systems"]
  C --> D{"Computational Process"}
  D --> D1["Lyapunov Stability Analysis<br>Validate Non-linear Potential Function"]
  D --> D2["Bayesian Inference<br>Inferring Potential Function from Prices"]
  D --> D3["Identify Market Indicators<br>Attractors & Repellers"]
  D1 & D2 & D3 --> E["Key Findings & Outcomes<br>1. Structural Dependency Consistent with Wavelet Coherence<br>2. Attractors/Repellers Surpass Traditional Metrics<br>3. Captured Market Trend, Volatility, & Correlation<br>4. Improved Prediction (e.g., Litecoin up to 12 days)"]

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 a novel model that treats cryptocurrency price movements as trajectories of a dynamical system governed by a time‑varying non‑linear potential field, and a Bayesian framework to infer this potential from price data.

Source sentence
(i) Proposed a novel model for cryptocurrency price movements as a trajectory of a dynamical system governed by a time-varying non-linear potential field.

Assumes 3

  • The analysis is restricted to Bitcoin crash periods, assuming these periods are representative for studying market dynamics. market
    Source sentence
    To simplify analysis, we focus on crash periods of BTC, the cryptocurrency with the largest market cap, for the numerical results.
  • A valid non‑linear potential function exists for the cryptocurrency market, justified via Lyapunov stability analysis. model
    Source sentence
    the existence of a valid potential function, using Lyapunov stability analysis.
  • The inferred mean attractor follows a multivariate Gaussian distribution characterized by (µa, Σa). statistical
    Source sentence
    The uncertainty in the inferred mean attractor is characterized by the Gaussian cloud, N(µa, Σa), contours of which in state space are represented as concentric ellipses centred at the mean attractors

Data

  • Top ten cryptocurrencies historical price dataset · coinmarketcap.com market-prices public
    Bitcoin, Ethereum, XRP, Cardano, Algorand, Binance coin, Polkadot, Litecoin, Polygon, Sola
    Source sentence
    The historical data of top ten cryptocurrencies in terms of market shares, namely, Bitcoin (BTC), Ethereum (ETH), XRP, Cardano (ADA), Algorand (ALGO), Binance coin (BNB), Polkadot (DOT), Litecoin (LTC), Polygon (MATIC) and Solana (SOL), has been obtained from coinmarketcap.com, an open-source website that provides six

Recognised sources: Binance CoinMarketCap

Costs

Not applicable The paper reports no trading or execution result.

Finds 4

  • Attractors inferred from the potential field capture market trend, volatility, and correlation during Bitcoin crash periods. theoretical
    Source sentence
    ... shows that attractors captured the market trend, volatility, and correlation.
  • The structural dependence inferred by the proposed approach is consistent with wavelet coherence results. statistical
    Source sentence
    The structural dependence inferred by the proposed approach was found to be consistent with results obtained using the popular wavelet coherence approach.
  • Using the mean attractor as an input feature improves LSTM price‑prediction MSE by up to 25% for 12‑day horizons compared to the existing method. performance
    Source sentence
    It can also be observed from the Table VIII that, the mean attractor based approach consistently performs better in terms of mean MSE over all the time horizons of predictions, till 12 days. The mean MSE of 12 day prediction of the proposed approach is 25% better compared to the existing approach.
  • MAE performance of the proposed approach is lower than the existing approach across all tested horizons up to 14 days. performance
    Source sentence
    The MAE performance comparison of both existing and the proposed approaches is given in Table IX.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 20 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.