Scholarlink

<p>The ability of human body to process glucose is affected by diabetes, and this condition can lead to various complications. If left untreated, it can cause various health issues, such as blindness, kidney failure, nerve damage, and amputation of the legs and feet. People with diabetes typically suffer from a reduced quality of life due to their condition and strict diet restrictions. Early diagnosis can help prevent or delay any complications, and it can facilitate treatment adjustments and lifestyle modifications. In addition, it can lower health costs and improve a person’s life expectancy. One of the most common methods of diagnosing diabetes is by measuring blood sugar levels. However, this method can be very invasive and uncomfortable. Some other methods only detect the condition after the symptoms have already developed, which can prevent early intervention. Deep learning and machine learning algorithms can help predict the likelihood of a person developing diabetes before its symptoms appear. They can also help improve the accuracy of the diagnosis and provide more comfortable and personalized care for patients. The present study proposed a mathematical model to detect the diabetics by understanding the disease and normal conditions using machine learning and deep learning algorithms. The study proposed a hybrid mathematical model that describes the step-by-step process to perform the classification of diabetic and no-diabetic data classes through feature extraction method. The classification process is conducted through experimentation using Pima Indian Diabetic dataset. The results achieved has shown that the proposed model has produced an efficient classification with nearly and more than 90% accuracy using machine learning and deep learning techniques respectively. The algorithms have shown an improvement of 20% when the classification is performed with augmented dataset. The future direction of the research is to improve the reliability and accuracy of identifying the diabetic disease using diverge datasets.</p>

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 hybrid mathematical model for diabetic data classification using machine learning and deep learning algorithms

  • 期刊论文

  • 发表日期：

    2026年3月9日

查看全文
  • Pothumarthi Sridevi,
  • K. Padmanaban

摘要

The ability of human body to process glucose is affected by diabetes, and this condition can lead to various complications. If left untreated, it can cause various health issues, such as blindness, kidney failure, nerve damage, and amputation of the legs and feet. People with diabetes typically suffer from a reduced quality of life due to their condition and strict diet restrictions. Early diagnosis can help prevent or delay any complications, and it can facilitate treatment adjustments and lifestyle modifications. In addition, it can lower health costs and improve a person’s life expectancy. One of the most common methods of diagnosing diabetes is by measuring blood sugar levels. However, this method can be very invasive and uncomfortable. Some other methods only detect the condition after the symptoms have already developed, which can prevent early intervention. Deep learning and machine learning algorithms can help predict the likelihood of a person developing diabetes before its symptoms appear. They can also help improve the accuracy of the diagnosis and provide more comfortable and personalized care for patients. The present study proposed a mathematical model to detect the diabetics by understanding the disease and normal conditions using machine learning and deep learning algorithms. The study proposed a hybrid mathematical model that describes the step-by-step process to perform the classification of diabetic and no-diabetic data classes through feature extraction method. The classification process is conducted through experimentation using Pima Indian Diabetic dataset. The results achieved has shown that the proposed model has produced an efficient classification with nearly and more than 90% accuracy using machine learning and deep learning techniques respectively. The algorithms have shown an improvement of 20% when the classification is performed with augmented dataset. The future direction of the research is to improve the reliability and accuracy of identifying the diabetic disease using diverge datasets.