Scholarlink

The focus of this study is to explore the feasibility of using ensemble learning techniques for binary classification in the prediction of cardiovascular disease, with the aim of making it easier to understand. Ensemble models, like gradient boosting and neural networks, were able to generate moderate values to make predictions that were 73.78 and 74.03% accurate, which was better than traditional classifiers like logistic regression. Gradient Boosting and Neural Networks were chosen for superior accuracy, handling imbalanced data, and flexibility, outperforming simpler methods like AdaBoost and Random Forests in complex medical datasets. However, interpretability is challenging due to their complexity. The study highlights the compromise between transparency and performance in medical artificial intelligence, emphasizing the need for interpretable, high-performance models for consistent patient treatment.

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

A Study of Performance and Interpretability of Ensemble Learning in Binary Classification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Babalola Abayomi,
  • Faiyaz Ahamad,
  • Manish Madhava Tripathi

摘要

The focus of this study is to explore the feasibility of using ensemble learning techniques for binary classification in the prediction of cardiovascular disease, with the aim of making it easier to understand. Ensemble models, like gradient boosting and neural networks, were able to generate moderate values to make predictions that were 73.78 and 74.03% accurate, which was better than traditional classifiers like logistic regression. Gradient Boosting and Neural Networks were chosen for superior accuracy, handling imbalanced data, and flexibility, outperforming simpler methods like AdaBoost and Random Forests in complex medical datasets. However, interpretability is challenging due to their complexity. The study highlights the compromise between transparency and performance in medical artificial intelligence, emphasizing the need for interpretable, high-performance models for consistent patient treatment.