Data Analysis and Prediction of Chronic Kidney Disease Using Machine and Deep Learning Techniques

Abstract

The timely detection and effective management of chronic kidney disease (CKD) are essential for decelerating its progression and minimizing associated complications. Regular monitoring and strict adherence to medical guidance are vital for individuals diagnosed with CKD. Machine learning is a field dedicated to enabling computers to learn autonomously, thereby eliminating the need for explicit programming. Through this process, machines independently acquire knowledge by analyzing relevant data, reducing reliance on external inputs. As a cornerstone of deep learning, and neural networks, machine learning is pivotal for advanced pattern recognition and predictive modeling. This paper examines CKD-related datasets comprising attributes such as blood pressure (Bp), specific gravity (Sg), albumin (Al), sugar (Su), red blood cells (Rbc), blood urea (Bu), serum creatinine (Sc), sodium (Sod), potassium (Pot), hemoglobin (Hemo), white blood cell count (Wbcc), red blood cell count (Rbcc), hypertension (Htn), and disease classification (Class). Machine learning techniques, including Linear Regression, Multilayer Perceptron, SMOreg, M5P, Random Forest, REP Tree, and proposed deep learning approaches, are employed to analyze and predict this dataset. Numerical results, supported by statistical tests and accuracy metrics, are presented to validate the proposed methodologies.