Scholarlink

<p>In the domain of healthcare analytics, accurate and efficient prediction of chronic diseases such as diabetes remains a significant challenge due to the complexity of medical data and the necessity for high precision in classification. Traditional machine learning models struggle with scalability and feature selection, leading to suboptimal performance in large datasets. To address these challenges, we propose the Dual-Phase Regressive Deep Neural MapReduce Classifier (DPRDNMR), an advanced predictive framework that integrates the Enhanced Feature Discriminative Preprocessing Model (EFDPM) for optimal feature selection, bivariant regression for mapping, and a deep neural network for classification, all optimized through a distributed MapReduce framework. Our model demonstrates superior performance across key evaluation metrics compared to conventional classification approaches such as Support Vector Machines (SVM), Deep Neural Networks (DNN), Random Forest, Logistic Regression, and XGBoost. Specifically, using the PIMA Indian Diabetic Dataset, the proposed framework achieves an accuracy of 90.3% for 250 patients, surpassing DNN (88.5%), SVM (86.6%), and other baseline models. Additionally, it exhibits a precision of 88.9%, recall of 99.6%, specificity of 87.1%, and an F1-score of 89.2%, significantly outperforming competing methods by an average margin of 2–5%. These results affirm the robustness of our approach in handling complex medical data with enhanced predictive accuracy and scalability. By leveraging EFDPM for feature selection, distributed processing, and deep learning-based classification, our framework facilitates early and accurate diabetic prediction, thereby assisting healthcare professionals in informed decision-making and improving patient outcomes.</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

Dual-Phase Regressive Deep Neural MapReduce Classifier for Scalable and Accurate Diabetic Prediction

  • 期刊论文

  • 发表日期：

    2025年5月5日

查看全文
  • D. Dhinakaran,
  • S. Edwin Raja,
  • M. Thiyagarajan,
  • D. Selvaraj,
  • R. Ramani

摘要

In the domain of healthcare analytics, accurate and efficient prediction of chronic diseases such as diabetes remains a significant challenge due to the complexity of medical data and the necessity for high precision in classification. Traditional machine learning models struggle with scalability and feature selection, leading to suboptimal performance in large datasets. To address these challenges, we propose the Dual-Phase Regressive Deep Neural MapReduce Classifier (DPRDNMR), an advanced predictive framework that integrates the Enhanced Feature Discriminative Preprocessing Model (EFDPM) for optimal feature selection, bivariant regression for mapping, and a deep neural network for classification, all optimized through a distributed MapReduce framework. Our model demonstrates superior performance across key evaluation metrics compared to conventional classification approaches such as Support Vector Machines (SVM), Deep Neural Networks (DNN), Random Forest, Logistic Regression, and XGBoost. Specifically, using the PIMA Indian Diabetic Dataset, the proposed framework achieves an accuracy of 90.3% for 250 patients, surpassing DNN (88.5%), SVM (86.6%), and other baseline models. Additionally, it exhibits a precision of 88.9%, recall of 99.6%, specificity of 87.1%, and an F1-score of 89.2%, significantly outperforming competing methods by an average margin of 2–5%. These results affirm the robustness of our approach in handling complex medical data with enhanced predictive accuracy and scalability. By leveraging EFDPM for feature selection, distributed processing, and deep learning-based classification, our framework facilitates early and accurate diabetic prediction, thereby assisting healthcare professionals in informed decision-making and improving patient outcomes.