Combating Market Manipulation: A WassersteinGAN GP Augmented, Focal Loss TabNet for Cryptocurrency Pump and Dump Detection
DOI:
https://doi.org/10.53799/kg3xc638Keywords:
smoteGAN, cryptocurrency fraud detection, pump-and-dump schemes, interpretable deep learning, class imbalance, synthetic data augmentationAbstract
Pump-and-dump (P&D) manipulation is one of the most persistent forms of cyber-enabled fraud in cryptocurrency markets, and detecting it automatically is made difficult by the extreme scarcity of confirmed P&D events relative to normal trading activity. This study develops an augmentation strategy that integrates a fully specified Wasserstein GAN with gradient penalty, and an alpha-balanced focal-loss TabNet classifier for transaction-level P&D detection. A WGAN-GP generator is trained independently within each cross-validation fold on the training partition only, contributing 30,000 synthetic minority rows per fold, combined with a bounded thirty-fold replication of genuine minority transactions to reach a documented augmented imbalance ratio of approximately 8.9 majority rows to one minority row. On the primary 25-second resolution, the full proposed model attained a precision of 100.0%, recall of 85.7%, and F1-score of 92.3% on a stratified held-out test split, and a cross-validated mean F1-score of 0.875 (precision 0.878, recall 0.874) under 5-fold stratified cross-validation with fold-wise generator retraining. A five-configuration ablation isolates the marginal contribution of WGAN-GP synthesis (ΔF1 +0.019 over a class-weighted baseline), hard-negative mining (+0.006), and focal loss (+0.038), demonstrating that the generative component contributes meaningfully but that the largest single improvement is attributable to the loss function rather than to the generator. Applying the 25-second-tuned hyperparameter configuration without re-tuning to the 15 and 5 second aggregations yields cross-validated F1-scores of 0.840 and 0.766 respectively, quantifying for the first time on this corpus the cost of resolution under-tuning in a GAN-augmented pipeline.
References
[1] S. Foley, J. R. Karlsen, and T. J. Putniņš, “Sex, drugs, and bitcoin: How much illegal activity is financed through cryptocurrencies?,” The Review of Financial Studies, vol. 32, pp. 1798–1853, 2019. DOI: https://doi.org/10.1093/rfs/hhz015
[2] J. Kamps and B. Kleinberg, “To the moon: defining and detecting cryptocurrency pump-and-dumps,” Crime Science, vol. 7, pp. 1–18, 2018. DOI: https://doi.org/10.1186/s40163-018-0093-5
[3] J. T. Hamrick, F. Rouhi, A. Mukherjee, M. Vasek, T. Moore, and N. Gandal, “Analyzing target-based cryptocurrency pump and dump schemes,” in Proc. 2021 ACM CCS Workshop on Decentralized Finance and Security, 2021, pp. 21–27. DOI: https://doi.org/10.1145/3464967.3488591
[4] M. La Morgia, A. Mei, F. Sassi, and J. Stefa, “The doge of wall street: Analysis and detection of pump and dump cryptocurrency manipulations,” ACM Transactions on Internet Technology, vol. 23, pp. 1–28, 2023. DOI: https://doi.org/10.1145/3561300
[5] A. Dhawan and T. J. Putniņš, “A new wolf in town? Pump-and-dump manipulation in cryptocurrency markets,” Review of Finance, vol. 27, pp. 935–975, 2023. DOI: https://doi.org/10.1093/rof/rfac051
[6] F. Victor and T. Hagemann, “Cryptocurrency pump and dump schemes: Quantification and detection,” in Proc. 2019 International Conference on Data Mining Workshops (ICDMW), 2019, pp. 244–251. DOI: https://doi.org/10.1109/ICDMW.2019.00045
[7] D. Fantazzini and Y. Xiao, “Detecting pump-and-dumps with crypto-assets: Dealing with imbalanced datasets and insiders’ anticipated purchases,” Econometrics, vol. 11, p. 22, 2023. DOI: https://doi.org/10.3390/econometrics11030022
[8] Y. Li, Z. Wang, H. Wang, M. Wu, and L. Xie, “Identifying price bubble periods in the Bitcoin market-based on GSADF model,” Quality & Quantity, pp. 1–16, 2021. DOI: https://doi.org/10.1007/s11135-020-01077-4
[9] I. Goodfellow et al., “Generative adversarial nets,” in Advances in Neural Information Processing Systems, vol. 27, 2014.
[10] S. O. Arik and T. Pfister, “TabNet: Attentive interpretable tabular learning,” arXiv preprint arXiv:1908.07442, 2019.
[11] M. La Morgia, A. Mei, F. Sassi, and J. Stefa, “Pump and dumps in the bitcoin era: Real time detection of cryptocurrency market manipulations,” in Proc. 2020 29th International Conference on Computer Communications and Networks (ICCCN), 2020, pp. 1–9. DOI: https://doi.org/10.1109/ICCCN49398.2020.9209660
[12] SystemsLab-Sapienza, “Pump and dump dataset,” GitHub, 2021. [Online]. Available: https://github.com/SystemsLab-Sapienza/pump-and-dump-dataset
[13] R. K. Aggarwal and G. Wu, “Stock market manipulations,” The Journal of Business, vol. 79, pp. 1915–1953, 2006. DOI: https://doi.org/10.1086/503652
[14] M. J. Rajaei and Q. H. Mahmoud, “A survey on pump and dump detection in the cryptocurrency market using machine learning,” Future Internet, vol. 15, p. 267, 2023. DOI: https://doi.org/10.3390/fi15080267
[15] N. Gandal, J. T. Hamrick, T. Moore, and T. Oberman, “Price manipulation in the Bitcoin ecosystem,” Journal of Monetary Economics, vol. 95, pp. 86–96, 2018. DOI: https://doi.org/10.1016/j.jmoneco.2017.12.004
[16] R. Mitchell and R. Chen, “Behavior-rule based intrusion detection systems for safety critical smart grid applications,” IEEE Transactions on Smart Grid, vol. 4, pp. 1254–1263, 2013. DOI: https://doi.org/10.1109/TSG.2013.2258948
[17] P. Priyadarshi and P. Kumar, “A comprehensive review on insider trading detection using artificial intelligence,” Journal of Computational Social Science, pp. 1–20, 2024.
[18] Z. Chen, S. Wang, D. Yan, and Y. Li, “Research and implementation of bank credit card fraud detection system based on reinforcement learning and LSTM,” in Proc. 2023 3rd International Conference on Mobile Networks and Wireless Communications (ICMNWC), 2023, pp. 1–8. DOI: https://doi.org/10.1109/ICMNWC60182.2023.10435890
[19] J. Xu and B. Livshits, “The anatomy of a cryptocurrency pump-and-dump scheme,” in Proc. USENIX Security Symposium, Santa Clara, CA, USA, Aug. 2019, pp. 1609–1625.
[20] T. M. Mitchell, Machine Learning, 1st ed. New York, NY, USA: McGraw-Hill, 1997.
[21] T. Hastie, R. Tibshirani, J. H. Friedman, and J. H. Friedman, The Elements of Statistical Learning: Data Mining, Inference, and Prediction. New York, NY, USA: Springer, 2009. DOI: https://doi.org/10.1007/978-0-387-84858-7
[22] S. Shao, “The effectiveness of supervised learning models in detection of pump and dump activity in Dogecoin,” in Proc. Second IYSF Academic Symposium on Artificial Intelligence and Computer Engineering, 2021, pp. 356–363. DOI: https://doi.org/10.1117/12.2622877
[23] V. Chadalapaka, K. Chang, G. Mahajan, and A. Vasil, “Crypto pump and dump detection via deep learning techniques,” arXiv preprint arXiv:2205.04646, 2022.
[24] D. Alfano, R. Abbruzzese, and D. Parente, “Pump and dump cryptocurrency detection using social media,” in Proc. International Conference on Data (DATA), 2023, pp. 235–240. DOI: https://doi.org/10.5220/0012059300003541
[25] N. V. Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer, “SMOTE: synthetic minority over-sampling technique,” Journal of Artificial Intelligence Research, vol. 16, pp. 321–357, 2002. DOI: https://doi.org/10.1613/jair.953
[26] G. Menardi and N. Torelli, “Training and assessing classification rules with imbalanced data,” Data Mining and Knowledge Discovery, vol. 28, pp. 92–122, 2014. DOI: https://doi.org/10.1007/s10618-012-0295-5
[27] C. Tantithamthavorn, A. E. Hassan, and K. Matsumoto, “The impact of class rebalancing techniques on the performance and interpretation of defect prediction models,” IEEE Transactions on Software Engineering, vol. 46, pp. 1200–1219, 2018. DOI: https://doi.org/10.1109/TSE.2018.2876537
[28] T. Wongvorachan, S. He, and O. Bulut, “A comparison of undersampling, oversampling, and SMOTE methods for dealing with imbalanced classification in educational data mining,” Information, vol. 14, p. 54, 2023. DOI: https://doi.org/10.3390/info14010054
[29] H. Nghiem, G. Muric, F. Morstatter, and E. Ferrara, “Detecting cryptocurrency pump-and-dump frauds using market and social signals,” Expert Systems with Applications, vol. 182, p. 115284, 2021. DOI: https://doi.org/10.1016/j.eswa.2021.115284
[30] R. Sauber-Cole and T. M. Khoshgoftaar, “The use of generative adversarial networks to alleviate class imbalance in tabular data: a survey,” Journal of Big Data, vol. 9, p. 98, 2022. DOI: https://doi.org/10.1186/s40537-022-00648-6
[31] A. Sharma, P. K. Singh, and R. Chandra, “SMOTified-GAN for class imbalanced pattern classification problems,” IEEE Access, vol. 10, pp. 30655–30665, 2022. DOI: https://doi.org/10.1109/ACCESS.2022.3158977
[32] U. Abdulrazaq, and M. N. Musa, “SmoteGAN and TabNet: a hybrid framework for detecting pump-and-dump schemes in cryptocurrency markets,” Proceedings of the Nigerian Society of Physical Sciences, vol. 3, pp. 257-257, 2026 DOI: https://doi.org/10.61298/pnspsc.2026.3.257
[33] I. Gulrajani, F. Ahmed, M. Arjovsky, V. Dumoulin, and A. C. Courville, "Improved training of Wasserstein GANs," in Advances in Neural Information Processing Systems, vol. 30, 2017.
[34] T.-Y. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár, "Focal loss for dense object detection," in Proc. IEEE International Conference on Computer Vision (ICCV), 2017, pp. 2980–2988. DOI: https://doi.org/10.1109/ICCV.2017.324
Downloads
Published
Issue
Section
License
Copyright (c) 2026 AIUB Journal of Science and Engineering (AJSE)
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
AJSE contents are under the terms of the Creative Commons Attribution License. This permits anyone to copy, distribute, transmit and adapt the work non-commercially provided the original work and source is appropriately cited.
