Comprehensive Analytical Study of Existing Malnutrition Prediction Models

Malnutrition continues to be a critical global health concern, profoundly affecting morbidity and death rates, especially among at-risk communities. Progress in machine learning (ML) and deep learning (DL) offers a revolutionary possibility for the early identification and forecasting of malnutrition, enabling prompt and efficient therapies. This paper analyzes modern machine learning and deep learning models developed for predicting malnutrition, focusing on their methodologies, data sources, feature selection techniques, and performance results. It identifies key trends and challenges in the field, including limitations related to data quality, model generalizability, and ethical considerations surrounding the use of sensitive health data. The review also explores critical gaps in current research, such as the lack of standardized evaluation metrics and the limited application of ML in diverse demographic contexts. The existing studies works on several kind of dataset such as IDHS (Indian Demographic Health Survey), BDHS (Bangladesh DHS) and MICS (Multiple Indicator Cluster Survey) dataset and also customized images dataset.