Intelligent Insurance Policy Compliance and Fraud Detection System Using Natural Language Processing and Multi-Layer Validation Architecture
International Journal of Advanced Research in Science Communication and Technology · 11.05.2026 · 626
DOI: 10.48175/ijarsct-35168 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 11.05.2026. Verfasst von Nikhil Kumar und 2 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben.
Abstract (Originalsprache)
Health insurance fraud and policy non-compliance represent two of the most significant financial challenges facing the insurance industry today. Manual claim verification processes not only fail to detect fraudulent submissions consistently but also introduce delays of 15 to 20 days per claim, creating substantial operational inefficiencies. This paper presents an Intelligent Insurance Policy Compliance and Fraud Detection System that addresses these challenges through a novel multi-layer validation architecture combined with Natural Language Processing techniques. The proposed system automatically detects policy violations, identifies potential fraudulent claim submissions such as inflated amounts exceeding actual billed values, and enforces general exclusion criteria using a predefined semantic keyword matching layer. A Retrieval Augmented Generation pipeline powered by FAISS vector indexing and Sentence Transformer embeddings retrieves contextually relevant policy clauses in real time, which are subsequently evaluated by the Groq LLaMA 3.3 70B large language model to generate structured compliance reports. Optical Character Recognition using EasyOCR extracts billing information from uploaded medical documents, enabling cross-verification of claimed amounts against actual billed totals — a key fraud prevention mechanism. The system is implemented using Python and Flask and deployed on Hugging Face Spaces at zero infrastructure cost. Experimental evaluation across eight fraud and compliance scenarios demonstrates 100% detection accuracy. Results confirm that the proposed multi-layer architecture reduces average processing time from 15 to 20 days to under 5 seconds while simultaneously improving fraud detection capability beyond what manual review achieves.
Zitieren
APA
Kumar, N., Swami, V., & Singh, R. (2026). Intelligent Insurance Policy Compliance and Fraud Detection System Using Natural Language Processing and Multi-Layer Validation Architecture. International Journal of Advanced Research in Science Communication and Technology, 626. https://doi.org/10.48175/ijarsct-35168
Vancouver
Kumar N, Swami V, Singh R. Intelligent Insurance Policy Compliance and Fraud Detection System Using Natural Language Processing and Multi-Layer Validation Architecture. International Journal of Advanced Research in Science Communication and Technology. 2026:626. doi:10.48175/ijarsct-35168
BibTeX
@article{Kumar2026,
author = {Kumar, Nikhil and Swami, Vandana and Singh, Rajendra},
title = {Intelligent Insurance Policy Compliance and Fraud Detection System Using Natural Language Processing and Multi-Layer Validation Architecture},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {626},
doi = {10.48175/ijarsct-35168}
}
EUVIMED