Scholarlink

The global health crisis posed by COVID-19 and its variants has prompted an urgent need for alternative strategies to enhance recovery rates. Typically, diet is known to play a vital role in an individual’s recovery from infectious diseases. Hence, this study delves into the influence of dietary components on recovery of COVID-19 across 170 countries, recognizing the intricacies of the disease and the limited availability of active treatments. Utilizing a nutrient dataset and employing models such as Ridge Regression, Multi Linear Regression, Decision Tree, Random Forest, and clustering techniques like K-Means, along with SHapley Additive exPlanations (SHAP) analysis, this retrospective analysis reveals a favorable association with protein and optimal calorie intake. The study underscores the importance of dietary considerations, especially protein, fats, and calorie intake, in fostering resilience and well-being during the recovery from COVID-19. By using statistical inferences, this study aims to emphasize and highlight the individual feature contributions to the target variables. This aids in making the results more comprehensible to the non-stakeholders. In addition to this, clustering exhibited notably higher recovery rates (Mean: 2.37%) associated with diets high in protein, moderate in fats and calorie intake. By leveraging Explainable AI (XAI) frameworks like SHAP, the relation between accuracy and its interpretability for complex machine learning models is easily explainable to non-stakeholders, a crucial step in predictive healthcare analysis.

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

Unveiling Critical Insights Using Predictive Analytics and Explainable AI: A Case Study on COVID-19 Through Statistical Inference

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Shalini Kammalam Srinivasan,
  • Murari B. Deshpande,
  • Onkar Suhasrao Pampattiwar,
  • Sudeepa Roy Dey

摘要

The global health crisis posed by COVID-19 and its variants has prompted an urgent need for alternative strategies to enhance recovery rates. Typically, diet is known to play a vital role in an individual’s recovery from infectious diseases. Hence, this study delves into the influence of dietary components on recovery of COVID-19 across 170 countries, recognizing the intricacies of the disease and the limited availability of active treatments. Utilizing a nutrient dataset and employing models such as Ridge Regression, Multi Linear Regression, Decision Tree, Random Forest, and clustering techniques like K-Means, along with SHapley Additive exPlanations (SHAP) analysis, this retrospective analysis reveals a favorable association with protein and optimal calorie intake. The study underscores the importance of dietary considerations, especially protein, fats, and calorie intake, in fostering resilience and well-being during the recovery from COVID-19. By using statistical inferences, this study aims to emphasize and highlight the individual feature contributions to the target variables. This aids in making the results more comprehensible to the non-stakeholders. In addition to this, clustering exhibited notably higher recovery rates (Mean: 2.37%) associated with diets high in protein, moderate in fats and calorie intake. By leveraging Explainable AI (XAI) frameworks like SHAP, the relation between accuracy and its interpretability for complex machine learning models is easily explainable to non-stakeholders, a crucial step in predictive healthcare analysis.