![[2602.14009] Named Entity Recognition for Payment Data Using NLP](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.14009] Named Entity Recognition for Payment Data Using NLP
Summary
This paper explores Named Entity Recognition (NER) techniques for payment data, presenting advanced models like PaymentBERT that enhance entity extraction accuracy in financial transactions.
Why It Matters
As financial institutions increasingly rely on automated systems for transaction processing, effective NER is crucial for compliance and efficiency. This research highlights cutting-edge methods that improve data extraction from various payment formats, which can significantly impact anti-money laundering efforts and operational efficiency.
Key Takeaways
- BERT-based models outperform traditional NER methods in payment data extraction.
- PaymentBERT achieves a 95.7% F1-score, setting a new standard for accuracy.
- The study includes extensive experiments on diverse payment formats, enhancing generalizability.
- Real-time processing capabilities are maintained with advanced NER architectures.
- Insights from this research can aid financial institutions in compliance and transaction processing.
Computer Science > Computation and Language arXiv:2602.14009 (cs) [Submitted on 15 Feb 2026] Title:Named Entity Recognition for Payment Data Using NLP Authors:Srikumar Nayak View a PDF of the paper titled Named Entity Recognition for Payment Data Using NLP, by Srikumar Nayak View PDF HTML (experimental) Abstract:Named Entity Recognition (NER) has emerged as a critical component in automating financial transaction processing, particularly in extracting structured information from unstructured payment data. This paper presents a comprehensive analysis of state-of-the-art NER algorithms specifically designed for payment data extraction, including Conditional Random Fields (CRF), Bidirectional Long Short-Term Memory with CRF (BiLSTM-CRF), and transformer-based models such as BERT and FinBERT. We conduct extensive experiments on a dataset of 50,000 annotated payment transactions across multiple payment formats including SWIFT MT103, ISO 20022, and domestic payment systems. Our experimental results demonstrate that fine-tuned BERT models achieve an F1-score of 94.2% for entity extraction, outperforming traditional CRF-based approaches by 12.8 percentage points. Furthermore, we introduce PaymentBERT, a novel hybrid architecture combining domain-specific financial embeddings with contextual representations, achieving state-of-the-art performance with 95.7% F1-score while maintaining real-time processing capabilities. We provide detailed analysis of cross-format generalization, abl...