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{202225,
        author = {Shivanshu Sharma and Mrs. Reshma and Shukthija and S Ramviswa},
        title = {A Self-Evolving Neural Network-Based Predictive Model for Assessing Risk of Jaundice During Pregnancy},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {6599-6611},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=202225},
        abstract = {Jaundice arising during pregnancy can develop rapidly and may lead to significant risks for both the mother and the fetus if it is not identified at an early stage. Traditional screening practices depend on periodic laboratory examinations and clinician-driven assessments, which offer limited predictive strength and do not adapt to trimester-specific physiological variations. This study proposes a Self-Evolving Neural Network (SENN)–based diagnostic framework that modifies its internal structure through continual learning mechanisms to enhance reliability in medical risk prediction.
The dataset comprises 2,300 organized maternal clinical pro-files that include liver function indicators (bilirubin, ALT, AST, ALP), symptom observations, medical background information, hereditary factors, and trimester-aware physiological attributes. Model development followed a two-stage workflow: an initial baseline classification using a multilayer perceptron and XG-Boost, followed by an adaptive learning phase utilizing Elastic Weight Consolidation (EWC) implemented with the Avalanche toolkit. The final model achieved an AUC-ROC of 0.92 and an overall accuracy of 90.2%, showing clear improvements over conventional static models.
The complete system integrates a secure FastAPI backend, a React-based patient portal, a clinician-oriented dashboard, ONNX Runtime for optimized inference, and data governance protocols compliant with HIPAA and GDPR standards. The adaptive and interpretable design enables the framework to ad-just continuously as new clinical information becomes available. Overall, this work demonstrates how self-evolving deep learning systems can play a transformative role in maternal healthcare and early jaundice risk prediction.},
        keywords = {Pregnancy Jaundice, Self-Evolving Neural Net-works, Continual Learning, Explainable AI, ONNX Runtime, FastAPI, Maternal Health Analytics.},
        month = {May},
        }