Scholarlink

Diabetes, one of the highest serious chronic diseases globally has increased danger to humans in recent years. Nonetheless, early finding of diabetes prevents the progress of the disease. This study applied the Lean Six Sigma Define-Measure-Analyze-Improve-Control (DMAIC) process development method to develop clinic workflow processes for obtaining measurements promptly to facilitate interventions to improve glycemic control. Our study proposes a new method based on the combined power of Lean Six Sigma (LSS), Bayesian optimization (BO), and Boosting Trees Machine Learning (ML) to enhance diabetes care quality and enable early diabetes detection. To complete this objective, we evaluated and compared the performance of six ML models (i.e., AdaBoost, XGBoost, CatBoost, LightGBM, Gradient Boosting, and SGD). Our experimental results demonstrate that the CatBoost model had the best classification performance with 99% accuracy and exceptional performance across various metrics. By combining LSS and BO with ML, healthcare associations can expand process efficiency, patient satisfaction, and quality improvement.

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

Boosting Tree Machine Learning Models Based on Lean Six Sigma to Predict Healthcare/Quality of Care Outcomes in Diabetes Patients Based on Clinical Factors

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Rabia Emhamed Al Mamlook,
  • Mohammed Almaiah,
  • Heba I. Elkhouly,
  • Qais Al-Na’amneh,
  • Ahmad Nasayreh,
  • Hasan Gharaibeh,
  • Ali Saeed Almuflih,
  • Hanin Bzizi

摘要

Diabetes, one of the highest serious chronic diseases globally has increased danger to humans in recent years. Nonetheless, early finding of diabetes prevents the progress of the disease. This study applied the Lean Six Sigma Define-Measure-Analyze-Improve-Control (DMAIC) process development method to develop clinic workflow processes for obtaining measurements promptly to facilitate interventions to improve glycemic control. Our study proposes a new method based on the combined power of Lean Six Sigma (LSS), Bayesian optimization (BO), and Boosting Trees Machine Learning (ML) to enhance diabetes care quality and enable early diabetes detection. To complete this objective, we evaluated and compared the performance of six ML models (i.e., AdaBoost, XGBoost, CatBoost, LightGBM, Gradient Boosting, and SGD). Our experimental results demonstrate that the CatBoost model had the best classification performance with 99% accuracy and exceptional performance across various metrics. By combining LSS and BO with ML, healthcare associations can expand process efficiency, patient satisfaction, and quality improvement.