Scholarlink

Diabetes is a disease that gradually begins to affect the physical condition of an individual. It is slowly starting to rise among the majority of people with increasing time. It can be caused by hereditary or dietary reasons and lead to an impact on different other parts of the body and create an unhealthy and disastrous life. To detect this incurable and fatal disease, this study aims to predict the occurrence of diabetes in a person using machine learning algorithms. The study employs ten supervised and unsupervised learning techniques to analyze a large dataset of medical records with three different approaches. The Explainable Artificial Intelligence (XAI) approach with LIME is used to interpret the model’s predictions and provide explanations for the results. We have achieved 99% accuracy and F1-score with our best-performing models.

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

Comparison of Machine Learning Models for Early Prediction of Diabetes with LIME Interpretability

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Abanti Chakraborty Shruti,
  • Marufa Kamal,
  • Rakib Hossain Rifat,
  • Ehsanur Rahman Rhythm,
  • Md. Humaion Kabir Mehedi,
  • Annajiat Alim Rasel

摘要

Diabetes is a disease that gradually begins to affect the physical condition of an individual. It is slowly starting to rise among the majority of people with increasing time. It can be caused by hereditary or dietary reasons and lead to an impact on different other parts of the body and create an unhealthy and disastrous life. To detect this incurable and fatal disease, this study aims to predict the occurrence of diabetes in a person using machine learning algorithms. The study employs ten supervised and unsupervised learning techniques to analyze a large dataset of medical records with three different approaches. The Explainable Artificial Intelligence (XAI) approach with LIME is used to interpret the model’s predictions and provide explanations for the results. We have achieved 99% accuracy and F1-score with our best-performing models.