Scholarlink

Cardiovascular diseases pose a significant public health challenge, particularly among young populations where early risk detection remains crucial. This study proposes using interpretable machine learning models for assessing cardiovascular risk in young Mexican adults using anthropometric, biochemical, and hematological biomarkers, focusing on homocysteine as a key metabolic indicator. We propose a hierarchical diagnostic approach comparing four decision tree algorithms (J48, CART, CHAID*, and SPAARC), prioritizing model explainability for clinical decision support. Our analysis identifies SPAARC as the best-balanced algorithm, maintaining diagnostic accuracy while preserving interpretability, critical for screening asymptomatic youth where traditional tools underperform. These findings provide a framework for developing adapted preventive tools in underrepresented populations.

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

Decision Trees as a Support Tool for the Early Diagnosis of Cardiovascular Diseases in Young Individuals

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Abraham Carrasco Barradas,
  • Marisol Cruz Ramírez,
  • Handry Serrano Fabela,
  • Abraham Barradas Cortés,
  • Daniela Ortega Juárez,
  • Rafael Rivera López,
  • Nayeli Salomé Barrios Cruz,
  • Erick de la Cruz Hernández

摘要

Cardiovascular diseases pose a significant public health challenge, particularly among young populations where early risk detection remains crucial. This study proposes using interpretable machine learning models for assessing cardiovascular risk in young Mexican adults using anthropometric, biochemical, and hematological biomarkers, focusing on homocysteine as a key metabolic indicator. We propose a hierarchical diagnostic approach comparing four decision tree algorithms (J48, CART, CHAID*, and SPAARC), prioritizing model explainability for clinical decision support. Our analysis identifies SPAARC as the best-balanced algorithm, maintaining diagnostic accuracy while preserving interpretability, critical for screening asymptomatic youth where traditional tools underperform. These findings provide a framework for developing adapted preventive tools in underrepresented populations.