Scholarlink

Patient classification within a hospital’s Emergency Department demands immediate attention from medical staff. Manual classification creates constant alerts for healthcare professionals in emergency cases. Only a few efficient approaches were present to classify the patient’s condition in triage. Our research centers on categorizing patients based on risk levels, utilizing their basic vitals as input features. This work gave an adequate classification of patients according to their risk level. We used Random Forest, Extra Tree, Gradient Boosting, Adaboost, Extreme Gradient Boosting, and Light Gradient Boosting Machine to classify patients in triage. Evaluation criteria encompass F1 score, recall, precision, sensitivity and specificity for patient classification within the triage system. Results indicated that Gradient Boost outperformed other algorithms, achieving an impressive F1 score (78.5%), sensitivity (57.31%) and recall value (76.67%) for the patient classification. Random Forest gave the best specificity (96.02%) and precision (87.69%) among all models in this work.

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

Patient Classification in Emergency Department Triage Using Ensemble Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Mehanas Shahul,
  • K. P. Pushpalatha

摘要

Patient classification within a hospital’s Emergency Department demands immediate attention from medical staff. Manual classification creates constant alerts for healthcare professionals in emergency cases. Only a few efficient approaches were present to classify the patient’s condition in triage. Our research centers on categorizing patients based on risk levels, utilizing their basic vitals as input features. This work gave an adequate classification of patients according to their risk level. We used Random Forest, Extra Tree, Gradient Boosting, Adaboost, Extreme Gradient Boosting, and Light Gradient Boosting Machine to classify patients in triage. Evaluation criteria encompass F1 score, recall, precision, sensitivity and specificity for patient classification within the triage system. Results indicated that Gradient Boost outperformed other algorithms, achieving an impressive F1 score (78.5%), sensitivity (57.31%) and recall value (76.67%) for the patient classification. Random Forest gave the best specificity (96.02%) and precision (87.69%) among all models in this work.