Paper: arXiv 2509.03260

Authors: Minjung Park, Gyuyeon Na, Soyoun Kim, Sunyoung Moon, HyeonJeong Cha, Sangmi Chai

Abstract

Abnormal cryptocurrency transactions - such as mixing services, fraudulent transfers, and pump-and-dump operations – pose escalating risks to financial integrity but remain notoriously difficult to detect due to class imbalance, temporal volatility, and complex network dependencies. Existing approaches are predominantly model-centric and post hoc, flagging anomalies only after they occur and thus offering limited preventive value. This paper introduces HyPV-LEAD (Hyperbolic Peak-Valley Lead-time Enabled Anomaly Detection), a data-driven early-warning framework that explicitly incorporates lead time into anomaly detection. Unlike prior methods, HyPV-LEAD integrates three innovations: (1) window-horizon modeling to guarantee actionable lead-time alerts, (2) Peak-Valley (PV) sampling to mitigate class imbalance while preserving temporal continuity, and (3) hyperbolic embedding to capture the hierarchical and scale-free properties of blockchain transaction networks. Empirical evaluation on large-scale Bitcoin transaction data demonstrates that HyPV-LEAD consistently outperforms state-of-the-art baselines, achieving a PR-AUC of 0.9624 with significant gains in precision and recall. Ablation studies further confirm that each component - PV sampling, hyperbolic embedding, and structural-temporal modeling - provides complementary benefits, with the full framework delivering the highest performance. By shifting anomaly detection from reactive classification to proactive early-warning, HyPV-LEAD establishes a robust foundation for real-time risk management, anti-money laundering (AML) compliance, and financial security in dynamic blockchain environments.

Complexity vs Empirical Score

  • Math Complexity: 6.0/10
  • Empirical Rigor: 8.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper introduces a novel framework (HyPV-LEAD) combining hyperbolic geometry and peak-valley sampling to shift cryptocurrency anomaly detection from reactive to proactive lead-time forecasting. The empirical evaluation is strong, demonstrating high PR-AUC (0.9624) on large-scale Bitcoin datasets alongside thorough ablation studies.

Research Flowchart

  flowchart TD
    A["Research Goal: Proactive Early-Warning<br/>for Crypto Anomalies (HyPV-LEAD)"] --> B["Data Input: Large-Scale Bitcoin<br/>Transaction Networks"]
    
    B --> C["1. Window-Horizon Modeling<br/>(Guarantees Actionable Lead-Time)"]
    B --> D["2. Peak-Valley (PV) Sampling<br/>(Fixes Imbalance & Preserves Continuity)"]
    B --> E["3. Hyperbolic Embedding<br/>(Captures Hierarchical Network Topology)"]
    
    C & D & E --> F["Structural-Temporal<br/>Anomaly Detection Model"]
    
    F --> G["Key Outcomes: Outperforms SOTA (PR-AUC 0.9624)<br/>Enables Real-Time AML Risk Management"]

Cited References (41)

  1. Kernel bandwidth selection for SVDD: The sampling peak criterion method for large data (2017)
  2. Classifying time series using local descriptors with hybrid sampling (2015)
  3. Optimal modification of peak-valley period under multiple time-of-use schemes based on dynamic load point method considering reliability (2021)
  4. Automatic topography of high-dimensional data sets by non-parametric density peak clustering (2021)
  5. Evaluation of performance enhancement in Ethereum fraud detection using oversampling techniques (2024)
  6. Kernel density estimation based sampling for imbalanced class distribution (2020)
  7. Introducing DeepBalance: Random deep belief network ensembles to address class imbalance (2017)
  8. Ensemble synthesized minority oversampling-based generative adversarial networks and random forest algorithm for credit card fraud detection (2023)
and 33 more
  1. Enhancing the random forest model via synthetic minority oversampling technique for credit-card fraud detection (2024)
  2. Enhancing fraud detection in imbalanced motor insurance datasets using CP-SMOTE and Random Under-Sampling (2025)
  3. Vulnerability of scale-free cryptocurrency networks to double-spending attacks (2021)
  4. Scaling properties of extreme price fluctuations in Bitcoin markets (2018)
  5. Adaptive attention-based graph representation learning to detect phishing accounts on the ethereum blockchain (2024)
  6. CoinLayering: an efficient coin mixing scheme for large scale bitcoin transactions (2020)
  7. Bitcoin data analytics: Scalable techniques for transaction clustering and embedding generation (2021)
  8. Hyperbolic self-supervised contrastive learning based network anomaly detection (2022)
  9. Hyperbolic graph convolutional neural networks (2019)
  10. Hyperbolic Graph Embeddings: A Survey and an Evaluation on Anomaly Detection (2024)
  11. A comprehensive look at the empirical performance of equity premium prediction (2008)
  12. Enhancing fraud detection in the Ethereum blockchain using ensemble learning (2025)
  13. Machine Learning for Anomaly Detection in Blockchain: A Critical Analysis, Empirical Validation, and Future Outlook (2025)
  14. LSTM-based anomaly detection in blockchain transactions (2020)
  15. Graph convolutional networks for fraud detection in financial transaction networks (2019)
  16. A novel hybrid GCN-LSTM algorithm for energy stock price prediction: leveraging temporal dynamics and inter-stock relationships (2025)
  17. Hybrid temporal-structural models for blockchain anomaly detection (2022)
  18. Attention-based LSTM for imbalanced financial fraud detection (2021)
  19. CoinJoin in the Wild: An Empirical Analysis in Dash (2021)
  20. Tracing Transactions Across Cryptocurrency Ledgers (2017)
  21. Live graph lab: Towards open, dynamic and real transaction graphs with NFT (2023)
  22. Anatomy of a Pump-and-Dump Scheme: Evidence from the Cryptocurrency Market (2019)
  23. Anomaly Detection in the Bitcoin System–A Network Perspective (2016)
  24. Unsupervised Learning for Robust Bitcoin Fraud Detection (2016)
  25. Toward Detecting Illegal Transactions on Bitcoin Using Machine-Learning Methods (2019)
  26. Detecting Illicit Bitcoin Transactions with Machine Learning (2020)
  27. Why do tree-based models still outperform deep learning on tabular data? (2022) — link
  28. Understanding the limitations of self-supervised learning for tabular anomaly detection (2023) — link
  29. A Novel Time Series Approach to Anomaly Detection and Correction for Complex Blockchain Transaction Networks (2024)
  30. Comparing hyperbolic graph embedding models on anomaly detection for cybersecurity (2024)
  31. GCN-based user representation learning for unifying robust recommendation and fraudster detection (2020)
  32. Blockchain as a disruptive force transforming industries, redefining markets, and shaping the future of global economies (2025)
  33. Cryptocurrencies and AI -Enabled Organised Fraud: Emerging Risks and Countermeasures (2025)

0 of 41 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.