A Machine Learning Approach to Assessing Packaged Food Healthiness and Recommending Healthier Alternatives via Mobile Application

“A Machine Learning Approach to Assessing Packaged Food Healthiness and Recommending Healthier Alternatives via Mobile Application”

Sarthak Nakhate1, Guruprasad Murme2, Sarthak Kulkarni3 , Vishal Patil4, Prof. Balaji Chaugule5

1,2,3,4UG Scholars, Department of Information Technology, Zeal College of Engineering and Research, Pune

5 Proffesor, Department of Information Technology, Zeal College of Engineering and Research, Pune

Abstract - This research presents an AI-powered Android application designed to assess the healthiness of packaged food products and offer healthier alternatives based on nutritional data. Users can scan barcodes to retrieve product information via the OpenFoodFacts API, including details such as sugar, fat, and salt content. The application evaluates these values using a machine learning model, predicting a health score out of 100. A key innovation of this system is its ability to recommend alternative products that score higher in health value, aligning with the user's dietary preferences and wellness goals. The backend is powered by Firebase for real-time data storage and user authentication, while barcode scanning is implemented using Google ML Kit. A custom ML model hosted via Flask provides score predictions dynamically. This mobile solution bridges the gap between consumer convenience and nutritional awareness, aiming to promote informed food choices and healthier consumption habits through intuitive technology integration.

Key Words:  machine learning, barcode scanner, packaged food, food health assessment