Machine learning models for childhood obesity risk assessment: design, validation, and reporting practices. A systematic review protocol

Systematic review, reported with PRISMA 2020 and PRISMA-S, of machine-learning models that classify or predict overweight or obesity in children and adolescents. It describes designs (prospective or cross-sectional), the use of anthropometric predictors, validation, calibration, and risk of bias with PROBAST+AI. The protocol was registered after the searches were run (15 October 2026; PubMed, Scopus, and Web of Science) and before screening of the results, data extraction, and risk-of-bias assessment.