Researchers have developed zScore-N, a neural network designed to replace traditional hand-written formulas for scoring wallet reputation in decentralized finance. This new model, trained on over 5 million wallets, significantly improves accuracy and handles incomplete data more effectively than previous methods. zScore-N achieves a 0.58 RMSE on a 1000-point scale, outperforming gradient-boosted trees and linear regression, and demonstrates a marked reduction in error when features are missing. AI
IMPACT Improves accuracy and robustness of reputation scoring in DeFi, potentially impacting access to financial services.
RANK_REASON The item is a research paper detailing a new neural network model for a specific task. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- Decentralised Finance's Unregulated Governance: Minority Rule in the Digital Wild West
- Dhanashekar Kandaswamy
- Gotit.pub
- Hugging Face
- Influence Flower
- ScienceCast
- zScore-N
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →