Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{197137,
        author = {Mrs. N. Jeevana Deepa and R. Vaishnavi and m.N.V.D. Surya Sri and N.B.N. Surekha and V. Samson Raj and Prof. Y. Venkat},
        title = {Design and Implementation of a Data Reconciliation System Using Deterministic Fuzzy Matching with AI-Assisted Explainability and Confidence Analysis},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {5719-5725},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=197137},
        abstract = {Data reconciliation is a critical requirement in enterprise systems where records from multiple sources must be accurately matched. Traditional fuzzy matching techniques often fail in such scenarios because they ignore or underweight numeric identifiers embedded within textual descriptions, leading to false positives and unreliable mappings. This project presents the design and implementation of a Deterministic Data Reconciliation System using numeric-aware fuzzy matching, augmented with AI-assisted explainability and ML-based confidence analysis. The proposed system combines textual similarity with explicit numeric comparison to enforce consistency and eliminate incorrect matches caused by numeric discrepancies. The system is implemented using an enterprise-grade N-Tier architecture comprising a React.js with TypeScript frontend, a Flask-based RESTful backend, and a PostgreSQL relational database. Secure access is enforced using JWT-based authentication with Role-Based Access Control (RBAC). While the deterministic matching engine remains the authoritative source of truth, OpenAI GPT-4o-mini is used exclusively to generate human-readable explanations, and Machine Learning (scikit-learn) provides independent confidence indicators (HIGH/MEDIUM/REVIEW) for reconciliation decisions. Experimental observations demonstrate that the deterministic, numeric-aware approach significantly reduces false positives compared to conventional token-based fuzzy matching. The integration of AI/ML as supportive augmentation layers enhances explainability and decision confidence without compromising determinism.},
        keywords = {Data Reconciliation, Deterministic Fuzzy Matching, Numeric-Aware Matching, Explainable AI (XAI), GPT-4o-mini, N-Tier Architecture, JWT Authentication, Role-Based Access Control (RBAC), Flask, React, PostgreSQL},
        month = {April},
        }