Scholarlink

<p>Physical activity plays a crucial role in maintaining physical health and enhancing quality of life. With advancements in technology, healthcare wearable devices have become popular tools for monitoring physical activity and predicting health status. This study aims to develop a model for classifying physical activity types using healthcare wearable devices and optimized machine learning (ML) methods. The developed model has potential applications in continuous activity monitoring for rehabilitation and remote patient care in clinical settings. The dataset comprised sensor data collected from multiple sessions with annotated samples across six activity classes. The activities considered in the study include Lying, Self Pace Walk, Sitting, Running 3 METs, Running 5 METs, and Running 7 METs. To classify the data, four primary ML algorithms HistgradientBoosting, Random Forest Classifier, Extreme Gradient Boosting (XGBoost) Classifier, and Decision Tree were used. Evaluation results showed that XGBoost performed the best in terms of accuracy. To further enhance classification performance, XGBoost hyperparameters (n_estimators and max_depth) were optimized using six meta-heuristic algorithms: Joint Adaptive Differential Evolution (JADE), Levy Flight Evolutionary Programming (Levy EP), Original Cultural Algorithm (CRO), Original Differential Evolution (DE), Self-Adaptive Differential Evolution (SADE), and Self-Adaptive Particle Swarm Optimization-based Differential Evolution (SAP_DE). The study found that the SADE + XGBoost hybrid model outperformed the others, achieving the highest accuracy. The proposed optimization resulted in approximately 1% accuracy improvement over the baseline XGBoost model. The optimized model’s results Accuracy: 0.915644, Recall: 0.916451, Precision: 0.914528, and F1 Score: 0.91534 demonstrated a significant improvement in classification performance, surpassing other models in all evaluation metrics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction of physical activity types based on healthcare wearable devices using optimized machine learning methods

  • 期刊论文

  • 发表日期：

    2025年9月11日

查看全文
  • Wei Si

摘要

Physical activity plays a crucial role in maintaining physical health and enhancing quality of life. With advancements in technology, healthcare wearable devices have become popular tools for monitoring physical activity and predicting health status. This study aims to develop a model for classifying physical activity types using healthcare wearable devices and optimized machine learning (ML) methods. The developed model has potential applications in continuous activity monitoring for rehabilitation and remote patient care in clinical settings. The dataset comprised sensor data collected from multiple sessions with annotated samples across six activity classes. The activities considered in the study include Lying, Self Pace Walk, Sitting, Running 3 METs, Running 5 METs, and Running 7 METs. To classify the data, four primary ML algorithms HistgradientBoosting, Random Forest Classifier, Extreme Gradient Boosting (XGBoost) Classifier, and Decision Tree were used. Evaluation results showed that XGBoost performed the best in terms of accuracy. To further enhance classification performance, XGBoost hyperparameters (n_estimators and max_depth) were optimized using six meta-heuristic algorithms: Joint Adaptive Differential Evolution (JADE), Levy Flight Evolutionary Programming (Levy EP), Original Cultural Algorithm (CRO), Original Differential Evolution (DE), Self-Adaptive Differential Evolution (SADE), and Self-Adaptive Particle Swarm Optimization-based Differential Evolution (SAP_DE). The study found that the SADE + XGBoost hybrid model outperformed the others, achieving the highest accuracy. The proposed optimization resulted in approximately 1% accuracy improvement over the baseline XGBoost model. The optimized model’s results Accuracy: 0.915644, Recall: 0.916451, Precision: 0.914528, and F1 Score: 0.91534 demonstrated a significant improvement in classification performance, surpassing other models in all evaluation metrics.