A Machine Learning Framework for Chronic Kidney Disease Analysis Using ORANGE Tool
Author Name: Vaibhav Bhatnagar, Shilpa Sharma, Swami Nisha Bhagirath and Divya Sharma
Paper id: 24101
Abstract:
Chronic Kidney Disease (CKD) is a critical worldwide health concern that needs early detection and accurate prediction to ensure timely intervention and treatment. This study explores the use of the ORANGE data mining tool for CKD prediction, using machine learning algorithms and visualization techniques. Key models employed include Neural Networks, Regression Analysis, and Random Trees, to determine their predictive performance. The evaluation metrics utilized include the Confusion Matrix, ROC Curve, and other statistical measures to ensure a comprehensive assessment of model accuracy and reliability. Results indicate that the Neural Network model achieved the highest predictive accuracy, while Regression Analysis provided significant insights into feature importance. The Random Tree model demonstrated robustness and interpretability in decision-making processes. ROC curve analysis revealed that all models achieved high Area Under the Curve (AUC) values, signifying strong classification capabilities. This research underscores the potential of using the ORANGE tool as a user-friendly platform for CKD prediction and highlights the comparative strengths of various machine learning techniques in diagnosing chronic conditions. These findings aim to aid clinicians and researchers in implementing efficient, data-driven approaches for early CKD detection.