A HYBRID MACHINE LEARNING APPROACH FOR SMART NUTRITION RECOMMENDATION IN PREGNANCY USING MULTI-SOURCE HEALTH DATA
DOI: https://doi.org/10.65725/JCISE/2/1/001
JOURNAL OF COMPUTATIONAL INTELLIGENCE SCIENCE AND ENGINEERING (JCISE)
ISSN: 3107-8168
Volume 2 Issue 1, Jan-Mar 2026

Abstract: Maternal nutrition is a critical determinant of healthy pregnancy outcomes, yet traditional dietary guidelines often fail to address individual variability and dynamic physiological changes. This study presents a hybrid machine learning–based smart nutrition recommendation system for pregnancy that leverages multi-source health data to deliver personalized and adaptive nutritional guidance. The proposed framework integrates IoT derived physiological sensor data, dietary intake records, ultrasound imaging features, and epigenetic markers to comprehensively model maternal–fetal health interactions. A hybrid learning architecture combining deep learning for multi-modal feature extraction, classical machine learning for predictive modelling, and rule-based clinical nutrition constraints is employed to enhance accuracy, interpretability, and clinical relevance. Advanced data fusion techniques address data heterogeneity, temporal dynamics, and missing values. Experimental results demonstrate that the proposed approach outperforms single-source and standalone models in terms of prediction accuracy and personalization. The framework enables real-time monitoring and adaptive nutrition recommendations across different stages of pregnancy, supporting precision maternal care and improved pregnancy outcomes. This research underscores the potential of hybrid machine learning and multi-modal data integration in advancing intelligent prenatal nutrition systems

Authors: Mrs.K.Rajasankari and Dr.S.Silvia Priscila

Keywords: Smart nutrition; Pregnancy; Hybrid machine learning; Multi-source data; IoT-based monitoring; Precision nutrition.