Using machine learning models with elephant herd feature selection method for diagnosing chronic kidney disease

Natarajan, Senthil Selvan; Balasubramanian, Elangovan; Raghupathy, Bala Krishnan; Ganesan, Manikandan

doi:10.21608/cjmss.2025.374590.1150

Using machine learning models with elephant herd feature selection method for diagnosing chronic kidney disease

Document Type : Original Article

Authors

¹ School of Computing, SASTRA Deemed To Be University, Thanjavur 613401, India

² Department of CSE, Koneru Lakshmaiah Education Foundation, Guntur 522302, India

³ Department of CSE, SASTRA Deemed to be University, Kumbakonam 612001, India

10.21608/cjmss.2025.374590.1150

Abstract

The kidneys filter waste, toxins, and excess water from the bloodstream, promoting body balance. Chronic kidney disease (CKD) is defined as a progressive deterioration of kidney function. Chronic kidney disease is a serious health condition that affects people all over the world. Early detection of kidney disease is crucial due to the lack of visible symptoms. The timely diagnosis of chronic kidney disease (CKD) significantly impacts the patient’s health development and allows for prompt treatment. The primary goal of this study is to detect the presence or absence of CKD in the human body by analyzing various features obtained from medical tests. So, to accurately detect and diagnose chronic kidney disease (CKD), we use machine learning techniques such as classifiers K-Nearest Neighbor (KNN), Naive Bayesian (NB), Logistic Regression (LR), and Decision Tree (DT). The study used the 400 instances and 25 features of the chronic kidney disease dataset obtained from Kaggle. The elephant herd optimization feature selection algorithm identifies eleven key features. Decision tree outperformed other classification algorithms, achieving an accuracy of 98.75%.

Keywords