Method for Constructing an Asset-Conditioned Index of News Direction of Impact for Crypto Assets using Zero-Shot NLI and GDELT
Keywords:
Keywords: cryptocurrency, news analytics, sentiment analysis, zero-shot NLI, cross-lingual NLP, GDELT, event-driven markets, volatility, decision-support systems, sentiment index.
Abstract
Abstract. This paper presents a reproducible pipeline for constructing asset-conditioned, direction-of-impact news sentiment for cryptocurrency markets. A multilingual news corpus is assembled using the GDELT DOC 2.0 API, and a cross-lingual Natural Language Inference encoder is applied to judge whether a headline indicates a positive or negative effect on a given asset’s price direction. The approach is zero-shot, meaning it does not require task-specific labeled data. It is language-agnostic and computationally efficient. The output includes per-article sentiment scores and daily indices per asset, suitable for descriptive analysis and for integration into decision-support contexts. The method is framed within research on cross-lingual inference and zero-shot classification and is designed for clarity, portability, and reproducibility.
References
1. Kulbhaskar A. K., Anamika, Subramaniam S. Breaking news headlines: Impact on trading activity in the cryptocurrency market. Economic Modelling, 2023, 126. DOI: 10.1016/j.econmod.2023.106397. Available at: https://www.sciencedirect.com/science/article/abs/pii/S0264999323002092 (accessed 25.08.2025).
2. Brauneis A., et al. Crypto volatility forecasting: Mounting a HAR, sentiment, and ML challenge. Computational Economics, 2024. Available at: https://link.springer.com/article/10.1007/s10690-024-09510-6 (accessed 25.08.2025).
3. Long S. C., et al. From whales to waves: Social media sentiment, volatility and crypto market dynamics. Journal of Economic Behavior & Organization, 2025. Available at: https://www.sciencedirect.com/science/article/pii/S0890838925001325 (accessed 25.08.2025).
4. Canayaz M., Cao C., Nguyen G., Wang Q. An anatomy of cryptocurrency sentiment. SSRN Electronic Journal, 2023. DOI: 10.2139/ssrn.4495683. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4495683 (accessed 25.08.2025).
5. Conneau A., Lample G., Rinott R., Williams A., Bowman S. R., Schwenk H., Stoyanov V. XNLI: Evaluating cross-lingual sentence representations. In: Proceedings of EMNLP, 2018. Available at: https://aclanthology.org/D18-1269/ (accessed 25.08.2025).
6. He P., Gao J., Chen W. DeBERTaV3: Im-proving DeBERTa using ELECTRA-style pre-training with gradient-disentangled embedding sharing. arXiv:2111.09543, 2021. Available at: https://
arxiv.org/abs/2111.09543 (accessed 25.08.2025).
7. Laurer M. mDeBERTa-v3-base-XNLI-multilingual-nli-2mil7: Model card. Hugging Face, 2022–2025. Available at: https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 (accessed 25.08.2025).
8. GDELT Project. GDELT DOC 2.0 API debuts. 2017. Available at: https://blog.gdeltproject.org/gdelt-doc-2-0-api-debuts/ (accessed 25.08.2025).
9. Yin W., Hay J., Roth D. Benchmarking zero-shot text classification: Datasets, evaluation and entailment approach. In: Proceedings of EMNLP, 2019. Available at: https://aclanthology.org/D19-1404/ (accessed 25.08.2025).
10. Wang S., Fang H., Khabsa M., Mao H., Ma H. Entailment as few-shot learner. arXiv:2104.14690, 2021. Available at: https://arxiv.org/abs/2104.14690 (accessed 25.08.2025).
11. Nguyen N. How event studies can be ap-plied to crypto markets. The Brattle Group, 2023. Available at: https://www.brattle.com/wp-content/uploads/2023/07/How-Event-Studies-Can-Be-Applied-to-Crypto-Markets.pdf (accessed 25.08.2025).
12. Zhou F. Application of event study methodology in the analysis of cryptocurrency returns. Emerging Markets Finance and Trade, 2024. DOI: 10.1080/1540496X.2024.2404173. Available at: https://www.tandfonline.com/doi/full/10.1080/1540496X.2024.2404173 (accessed 25.08.2025).
13. Yue W., et al. Asymmetric news effects on cryptocurrency liquidity. International Review of Economics & Finance, 2021. Available at: https://www.sciencedirect.com/science/article/abs/pii/S1544612320316135 (accessed 25.08.2025).
14. Haider Ali H., et al. Return and volatility spillover between cryptocurrencies, oil price and stock market in GCC countries. Cogent Economics & Finance, 2025. Available at: https://www.tandfonline.com/doi/full/10.1080/23322039.2025.2453584 (accessed 25.08.2025).
15. Yang H., Liu X., Wang C. D. FinGPT: Open-source financial large language models. arXiv:2306.06031, 2023. Available at: https://arxiv.org/abs/2306.06031 (accessed 25.08.2025).
16. Bhatia G., Nagoudi E. M. B., Cavusoglu H., Abdul-Mageed M. FinTral: A family of GPT-4-level multimodal financial large language models. Findings of ACL, 2024. DOI: 10.18653/v1/2024.findings-acl.774. Available at: https://aclanthology.org/2024.findings-acl.774.pdf (accessed 25.08.2025).
2. Brauneis A., et al. Crypto volatility forecasting: Mounting a HAR, sentiment, and ML challenge. Computational Economics, 2024. Available at: https://link.springer.com/article/10.1007/s10690-024-09510-6 (accessed 25.08.2025).
3. Long S. C., et al. From whales to waves: Social media sentiment, volatility and crypto market dynamics. Journal of Economic Behavior & Organization, 2025. Available at: https://www.sciencedirect.com/science/article/pii/S0890838925001325 (accessed 25.08.2025).
4. Canayaz M., Cao C., Nguyen G., Wang Q. An anatomy of cryptocurrency sentiment. SSRN Electronic Journal, 2023. DOI: 10.2139/ssrn.4495683. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4495683 (accessed 25.08.2025).
5. Conneau A., Lample G., Rinott R., Williams A., Bowman S. R., Schwenk H., Stoyanov V. XNLI: Evaluating cross-lingual sentence representations. In: Proceedings of EMNLP, 2018. Available at: https://aclanthology.org/D18-1269/ (accessed 25.08.2025).
6. He P., Gao J., Chen W. DeBERTaV3: Im-proving DeBERTa using ELECTRA-style pre-training with gradient-disentangled embedding sharing. arXiv:2111.09543, 2021. Available at: https://
arxiv.org/abs/2111.09543 (accessed 25.08.2025).
7. Laurer M. mDeBERTa-v3-base-XNLI-multilingual-nli-2mil7: Model card. Hugging Face, 2022–2025. Available at: https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 (accessed 25.08.2025).
8. GDELT Project. GDELT DOC 2.0 API debuts. 2017. Available at: https://blog.gdeltproject.org/gdelt-doc-2-0-api-debuts/ (accessed 25.08.2025).
9. Yin W., Hay J., Roth D. Benchmarking zero-shot text classification: Datasets, evaluation and entailment approach. In: Proceedings of EMNLP, 2019. Available at: https://aclanthology.org/D19-1404/ (accessed 25.08.2025).
10. Wang S., Fang H., Khabsa M., Mao H., Ma H. Entailment as few-shot learner. arXiv:2104.14690, 2021. Available at: https://arxiv.org/abs/2104.14690 (accessed 25.08.2025).
11. Nguyen N. How event studies can be ap-plied to crypto markets. The Brattle Group, 2023. Available at: https://www.brattle.com/wp-content/uploads/2023/07/How-Event-Studies-Can-Be-Applied-to-Crypto-Markets.pdf (accessed 25.08.2025).
12. Zhou F. Application of event study methodology in the analysis of cryptocurrency returns. Emerging Markets Finance and Trade, 2024. DOI: 10.1080/1540496X.2024.2404173. Available at: https://www.tandfonline.com/doi/full/10.1080/1540496X.2024.2404173 (accessed 25.08.2025).
13. Yue W., et al. Asymmetric news effects on cryptocurrency liquidity. International Review of Economics & Finance, 2021. Available at: https://www.sciencedirect.com/science/article/abs/pii/S1544612320316135 (accessed 25.08.2025).
14. Haider Ali H., et al. Return and volatility spillover between cryptocurrencies, oil price and stock market in GCC countries. Cogent Economics & Finance, 2025. Available at: https://www.tandfonline.com/doi/full/10.1080/23322039.2025.2453584 (accessed 25.08.2025).
15. Yang H., Liu X., Wang C. D. FinGPT: Open-source financial large language models. arXiv:2306.06031, 2023. Available at: https://arxiv.org/abs/2306.06031 (accessed 25.08.2025).
16. Bhatia G., Nagoudi E. M. B., Cavusoglu H., Abdul-Mageed M. FinTral: A family of GPT-4-level multimodal financial large language models. Findings of ACL, 2024. DOI: 10.18653/v1/2024.findings-acl.774. Available at: https://aclanthology.org/2024.findings-acl.774.pdf (accessed 25.08.2025).
Published
2025-10-03
Section
Data Science & Big Data
This work is licensed under a Creative Commons Attribution 4.0 International License.
