Comparative Analysis for Prediction of Kidney Disease Using Intelligent Machine Learning Methods.

Area of Science:

• Nephrology
• Medical Informatics
• Data Science

Background:

• Chronic kidney disease (CKD) presents a significant global health challenge due to rising prevalence and poor patient outcomes.
• Dietary habits and hydration significantly impact kidney health, necessitating early detection methods.
• Current methods for CKD prediction require enhancement for improved reliability and accuracy.

Purpose of the Study:

• To develop and evaluate machine learning models for accurate prediction of chronic kidney disease (CKD) status.
• To identify the most effective machine learning algorithm for early CKD detection using clinical data.
• To establish a reliable methodology for CKD prediction incorporating data preprocessing and feature extraction.

Main Methods:

• Utilized a publicly available CKD dataset for model training and validation.
• Applied data preprocessing techniques, including missing value imputation and data aggregation.
• Trained and compared three distinct machine learning models: Logistic Regression (LR), Decision Tree (DT), and K-Nearest Neighbors (KNN).

Main Results:

• Logistic Regression (LR) demonstrated the highest prediction accuracy, achieving approximately 97%.
• The developed models showed superior accuracy compared to previous research findings.
• Extensive model comparisons confirmed the resilience and trustworthiness of the proposed prediction scheme.

Conclusions:

• Machine learning, specifically logistic regression, offers a highly accurate and reliable approach for early CKD prediction.
• The implemented methodology, including data handling and feature extraction, enhances prediction capabilities.
• This study provides a robust framework for improving CKD detection rates and patient management.