Machine Learning-based Multi-Class Classification of Human Fitness Activities for Personalized Wellness Solutions

Abstract

The growth of sedentary lifestyles and lifestyle-related health conditions has highlighted the need for individualized wellness solutions to promote physical activity and health. Basic exercises like jumping jacks and squats to complex routines like pull-ups promote physical health and personalized wellness. General advice or manual tracking methods could not provide enough precision and customisation to meet individual fitness needs. Manual fitness diaries, basic workout regimens, and standardized fitness exams gave limited user performance and progress insights, sometimes resulting in unsatisfactory results. These labour-intensive, error-prone, and unadaptable methods made it hard to achieve goals or discover areas for improvement.