Scholarlink

Diabetes, a global health challenge characterized by insufficient insulin production or utilization, has experienced an alarming surge, reaching 422 million cases in 2014 from 108 million in 1980. This epidemic disproportionately affects low- and middle-income nations, contributing significantly to adverse health outcomes. Between 2000 and 2019, diabetes-related mortality rates saw a concerning 3% rise, particularly pronounced with a 13% increase in lower-middle-income countries. Addressing the urgency of effective diabetes management, this study employs supervised machine-learning algorithms (ML Classifiers)—“Logistic Regression (LR), Naïve Bayes (NB), Decision Tree (DT), Random Forest (RF), Artificial Neural Network (ANN) and Light Gradient Boosting Machine (LightGBM)”. Notably, LightGBM emerges as the preeminent performer, surpassing other classifiers in predicting diabetes with unparalleled accuracy. These results underscore the potential of advanced ML techniques in enhancing our understanding and prediction of diabetes, providing valuable insights for proactive intervention and management strategies.

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

Machine Learning Advancements for Diabetes Prediction with LightGBM

  • 章节

  • 发表日期：

    2026年

查看全文
  • K. V. V. B. Durga Prasad,
  • Elakkiya Sekar,
  • M. Janaki Ramudu,
  • K. Bhanurekha

摘要

Diabetes, a global health challenge characterized by insufficient insulin production or utilization, has experienced an alarming surge, reaching 422 million cases in 2014 from 108 million in 1980. This epidemic disproportionately affects low- and middle-income nations, contributing significantly to adverse health outcomes. Between 2000 and 2019, diabetes-related mortality rates saw a concerning 3% rise, particularly pronounced with a 13% increase in lower-middle-income countries. Addressing the urgency of effective diabetes management, this study employs supervised machine-learning algorithms (ML Classifiers)—“Logistic Regression (LR), Naïve Bayes (NB), Decision Tree (DT), Random Forest (RF), Artificial Neural Network (ANN) and Light Gradient Boosting Machine (LightGBM)”. Notably, LightGBM emerges as the preeminent performer, surpassing other classifiers in predicting diabetes with unparalleled accuracy. These results underscore the potential of advanced ML techniques in enhancing our understanding and prediction of diabetes, providing valuable insights for proactive intervention and management strategies.