•  
  •  
 

Corresponding Author ORCID Identifer

K. Revathi https://orcid.org/0000-0003-3411-9264

Section

Regular Issue

Abstract

Global shifts toward sedentary habits and refined dietary choices are magnifying public-health pressures from obesity, metabolic dysregulation, and cardiovascular disease, positioning targeted, personalized fitness coaching tools at the forefront of preventive-care strategy. By embedding tailored nutritional and physical-activity plans into daily life, these tools reinforce adherence, enhance intrinsic motivation, and catalyze sustained wellness achievements—each moderated by real-time analytics, feedback loops, and progressive adaptation to evolving behavior and physiology. Their leading benefits include prolonged engagement, diminished risk of relapse, and deep integration of healthful routines, culminating in lowered lifetime healthcare costs and improved individual life satisfaction. Current machine-learning approaches deliver commendable dietary and activity personalization yet often fragment the upstream clustering, forward-looking forecasting, and downstream monitoring into separate silos. This manuscript proposes a comprehensive intelligent coaching architecture that combines probabilistic Gaussian Mixture Models to cluster users by real-time activity and baseline health profile, fulfills dietary and regimen prescription by context-sensitive Extreme Gradient Boosted Trees, and pairs MediaPipe biosignal capture for on the-spot performance diagnosis and postural adjustment. User-background information—age, weight, and unique preferences—pours into the platform and sparks tailored exercise plans that refresh automatically via the linked mobile app, ensuring routines stay fresh and relevant. Progress is recorded continuously, and instant feedback is given to maintain safety during workouts. When tested against varied datasets, the system delivers strong results: Gaussian Mixture Model clustering reaches 92% in segmenting users, XGBoost-based recommendations report 95% precision, 93% recall, and 94% F1-score for personalizing plans, and posture tracking powered by MediaPipe achieves 96% accuracy with minimal latency. Collectively, the metrics confirm the platform’s effectiveness in fostering sustained and meaningful wellness engagement.

Download

Share

COinS