Abstract
This paper presents the development of a data-driven personalized fitness web application aimed at combating obesity and sedentary behavior.
The web application leverages machine learning algorithms, particularly
Gradient Boosting, to provide individualized fitness and dietary recommendations based on user input, such as BMI, activity levels, and health
data. Built on the Django framework, the application ensures scalability, security, and ease of use. Key features include real-time adaptive
recommendations, a user-friendly interface, and a feedback loop that personalizes fitness plans according to user progress. The machine learning
models were trained on a large dataset and tested against several models, with Gradient Boosting achieving the highest prediction accuracy (R²
= 0.9975). Initial user feedback indicated high satisfaction, particularly
due to the system’s adaptability to evolving health conditions. Future
research directions include enhancing algorithm performance, expanding
data sources, and incorporating wearable devices for more precise realtime recommendations.