This paper presents a novel machine learning-based system for predicting safe food consumption windows. By integrating environmental factors, cooking methods, and storage conditions, our system dynamically estimates food safety durations. Using a Gradient Boosting Regressor model, the system achieves robust performance (with a mean absolute error of approximately ±2.3 hours and an R² score of 0.89) across diverse storage scenarios. In addition, the full-stack implementation—featuring a Next.js frontend and a Flask API backend—facilitates real-time predictions and user-friendly data entry. This approach has significant potential to reduce foodborne illness risks while optimizing storage practices.
