Exploratory Data Analysis and Machine Learning Classification of Chronic Kidney Disease: Insights into Causes and Complications

• Published in: International Conference on Signal Processing and Communication (Online) pp. 393 - 401
• Main Authors: Gavade, Priyanka, Halli, Akash, Shahapurkar, Swapnil Laxman, Patil, Abhay
• Format: Conference Proceeding
• Language: English
• Published: IEEE 20.02.2025
• Subjects: Blood; Blood pressure; Chronic kidney disease; Correlation; Data analysis; Decision Tree; Decision trees; Diabetes; K Nearest Neighbor; Linear regression; Logistic regression; Machine learning; Naïve Bayes
• ISSN: 2643-444X
• DOI: 10.1109/ICSC64553.2025.10967962

As a component of the United Nations' 2030 Agenda, Sustainable Development Goal 3 (SDG 3) aims to improve access to healthcare and increase global health by tackling a variety of health issues, including both infectious and non-communicable diseases. Roughly 9.5% of people worldwide suffer from chronic kidney failure(CKD) which is a serious health disease, with prevalence rates differing across regions. CKD progressively reduces kidney efficiency in filtering blood, leading to the accumulation of toxins and fluids in the body. These complications can result in severe health outcomes, such as cardiovascular disease, elevated blood pressure, stroke, and reduced life expectancy. To deepen insights into CKD, we applied exploratory data analysis to explore correlations between key variables and class-specific outcomes. Using five machine learning models, Logistic Regression demonstrated superior performance, attaining 98% accuracy, 98% precision, and 97% recall. Logistic Regression outperformed Linear Regression, K-Nearest Neighbor, Naive Bayes, and Decision Trees, according to 10-fold cross-validation used to validate these models.