Comparative Study of Machine Learning Techniques on Chronic Kidney Disease Prediction
Soumyarashmi Panigrahi, Dibya Ranjan Das Adhikary, Binod Kumar Pattanayak, Bibhuprasad Mohanty, Bibhuti Bhusan Dash, Sudhansu Shekhar Patra
Abstract
A malfunctioning kidney can cause the increase of waste in the blood, adversely affecting various bodily processes and systems. ML techniques can assist in classifying patients as having chronic kidney disease (CKD) or not. Although previous researchers have conducted several studies, they have often neglected important steps like data preprocessing and feature engineering. In this paper, we have addressed these gaps by implementing various data preprocessing and feature engineering techniques such as Principal Component Analysis (PCA) on the CKD dataset. By carefully selecting features from our dataset, we identified the top-performing classification models which are RFC, SVC (Polynomial Kernel), XG Boost Classifier (XGBC) and Gradient Boost Classifier (GBC) gave the best performance in terms of accuracy. To evaluate the effectiveness of these models, we have utilized various performance metrics. The results obtained have significant contribution to the understanding and management of CKD, ultimately helping in the early detection and treatment of this widespread and critical health issue.