Scholarlink

Sepsis is a dangerous condition that can lead to life-threatening consequences. It develops due to the over reactive toxic response of the human immune system to an infection that has already developed within the body and leads to damage of important organs as well as death in most of the affected patients. Due to its high fatality rate, it is vital for its detection to occur at the earliest moment. The paper has followed a stacked model approach where the multiple models have been used to analyse the data collected and train the model for detecting sepsis based on the patient’s vitals. The proposed stacked model has demonstrated remarkable performance metrics. It achieves an accuracy of 94.27%, showcasing its reliability in classifying cases correctly. Additionally, the model reports an Area under the Receiver Operating Characteristic Curve (AUROC) of 0.98, indicating its exceptional capability to distinguish between septic and non-septic cases. Furthermore, the model's F1 score of 0.940397 highlights its balanced effectiveness in precision and recall, ensuring minimal false positives and negatives.

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

Stacked Model Framework for Accurate and Early Sepsis Detection Using Ensemble Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Pavithra Sekar,
  • Aryan Kapoor,
  • Swapnil Sunil Baviskar,
  • Dhairya Gupta

摘要

Sepsis is a dangerous condition that can lead to life-threatening consequences. It develops due to the over reactive toxic response of the human immune system to an infection that has already developed within the body and leads to damage of important organs as well as death in most of the affected patients. Due to its high fatality rate, it is vital for its detection to occur at the earliest moment. The paper has followed a stacked model approach where the multiple models have been used to analyse the data collected and train the model for detecting sepsis based on the patient’s vitals. The proposed stacked model has demonstrated remarkable performance metrics. It achieves an accuracy of 94.27%, showcasing its reliability in classifying cases correctly. Additionally, the model reports an Area under the Receiver Operating Characteristic Curve (AUROC) of 0.98, indicating its exceptional capability to distinguish between septic and non-septic cases. Furthermore, the model's F1 score of 0.940397 highlights its balanced effectiveness in precision and recall, ensuring minimal false positives and negatives.