Year: 2025
Type: conference-paper
Abstract: Millions of people throughout the world suffer from CKD (Chronic Kidney Disease), which significantly affects the rates of mortality and morbidity.To slow progression of the CKD while preventing respective complications, prompt identification and therapy are essential.Acknowledgement to the improved measures of variables that can define via predictive modeling, we are able to create a set of predictive models with the use of machine learning and prediction analytics.In fact, Classifiers based on machine learning algorithms have been studied thoroughly showing that they are thebest performance indicators which reveal that innovative logics and better machine learning models produce superior outcomes.The most crucial variables for the prediction model were chosen after pre-processing the information and selecting the features.The models' performance was assessed using their accuracy, specificity, precision, and sensitivity and F1-score metrics.The findings suggest that Decision Tree algorithms produced more precise and effective detection outcomes.Specific gravity, hypertension, hemoglobin, diabetes mellitus, albumin, appetite, and red blood cell count were among the most crucial factors for predicting CKD.In summary, decision tree algorithms, in particular Random Forest can be useful for predicting CKD and can help with early detection and management of the condition.According to the methodology applied in the study, it can be said that finding newer results that can be useful in accessing the mandatory predictions in the area of renal disease and beyond is an exciting way to take advantage of current advancements in machine learning and predictive modeling.
Authors: Ramjee Dixit, Vivek Kumar Pandey, Avdhesh Kumar Tiwari
Institution: Dr. A.P.J. Abdul Kalam Technical University
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: gold
OpenAlex ID: W4411094143