A Machine Learning Based Physical Exercise and Meditation Recommendation System for Patients with Chronic Health Conditions

Chronic health conditions are long-term medical issues that degrade a person's life quality. These disorders frequently require ongoing medical care and can severely limit daily activities. Managing chronic health issues usually requires regular physical exercise, meditation, and stress-minimization technique. Previous research has focused primarily on recommending physical exercise or meditation for patients with chronic health conditions, rather than combining the two. They also did not look into proper machine learning (ML) model selection, accuracy results in improvement, dataset development, or hyperparameter tuning techniques for the physical exercise and meditation recommendation systems. To eradicate these issues, this paper delivers an ML-based physical exercise and meditation recommendation system for patients with chronic health conditions based on 9 ML classifiers. This paper created a dataset based on patient data that includes characteristics such as person's habits, alcohol consumption information, exercise information, chronic health information, stress level, anxiety symptoms, sleeping issues, exercise, and medication recommendations. Among the ML models tested, the K-nearest neighbors (KNN) model had the highest accuracy of 96% and F1 score of 98% for predicting appropriate physical exercise and medication recommendations. According to the important performance metric value comparison results, the proposed KNN-based prediction scheme outperforms previous works by 6.6 percent in accuracy and 5 percent in recall value. This paper brings a mobile application that initiates features like physical exercise and medication recommendations, information about physical exercise and medication, and doctor appointments. The mobile app evaluation revealed that more than 83% of reviewers provided acceptable feedback on the app's design and necessity. Received: 1 November 2024 | Revised: 27 December 2024 | Accepted: 7 May 2025 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The dataset used in this paper was collected using Google Forms and is available from the corresponding author upon reasonable request. Author Contribution Statement Rakib Uddin Chowdhury: Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation. Mahfuzulhoq Chowdhury: Conceptualization, Methodology, Validation, Resources, Writing — original draft, Writing — review & editing, Visualization, Supervision, Project administration.