FitCluster AI: A Semi-Personalized Fitness and Diet Recommendation System Using Rule-Based and Lightweight Machine Learning Techniques

Maintaining a consistent fitness and dietary routine has become increasingly difficult in modern lifestyles due to lack of time, improper guidance, and absence of personalization. Many individuals rely on generic plans that fail to consider their unique body conditions and goals. This leads to ineffective results and lack of motivation. To address this issue, this paper presents FitCluster AI, a semi-personalized recommendation system that combines rule-based logic with lightweight machine learning techniques. The system utilizes user inputs such as age, height, weight, activity level, and fitness goals to categorize users into predefined clusters. Based on these clusters, rule-based logic is applied to generate tailored recommendations. Unlike deep learning approaches, the proposed system does not require large datasets or high computational resources, ensuring efficiency and privacy. Experimental results demonstrate that the system produces consistent and meaningful outputs across different user profiles. The proposed approach provides a practical solution for real-world fitness applications.