Chronic Kidney Disease Dataset

Published: 3 November 2025| Version 1 | DOI: 10.17632/pthckzzh49.1
Contributor:
Sara Fanatirashidi

Description

The original Chronic Kidney Disease (CKD) dataset was obtained from Kaggle. It contains 400 patient records. After data cleaning and removal of incomplete or inconsistent entries, 158 valid samples were retained for analysis. To ensure compatibility with the principles of Data Envelopment Analysis (DEA), the class labels were recoded into binary form: 0 = CKD (diseased) and 1 = non-CKD (healthy). All categorical attributes (e.g., yes/no, present/absent) were converted into binary numerical variables. Additionally, the patient identifier (ID), which has no diagnostic value, was excluded from the feature set.

Files

Categories

Machine Learning, Classification (Machine Learning)

Licence