A Stochastic Gradient Support Vector Optimization Algorithm for Predicting Chronic Kidney Diseases

Today, Chronic Kidney Disease has become one of the most crucial sicknesses and needs a serious diagnosis immediately. Past research outcomes have shown that machine-learning techniques are trustworthy enough for medical care. With the benefit of important results achieved from machine learning classifier algorithms, clinicians and medical staff can detect the disease on time. Besides, by employing unbalanced and small datasets of Chronic Kidney Disease, this work offers developers of medical systems insights to help in the early prediction of Chronic Kidney Disease to lessen the effects of late diagnosis, particularly in low-income and difficult-to-reach places. In this study, an effective prediction model based on machine learning methods is presented for Chronic Kidney Disease (CKD) using the Stochastic Gradient Support Vector Optimization Algorithm (SPegasos). Moreover, we use the benefit of the SMOTE technique during data pre-processing on a real dataset to remove all of the noisy and imbalanced data. Finally, the performance of the proposed prediction model using the SPegasos algorithm was evaluated by the WEKA tool. The experimental results show that the proposed model achieves the accuracy 99.9% to detect CKD with compare to the other machine learning algorithms.