K- Nearest Neighbour-Based Chronic Kidney Disease Prediction System: A Case of Toxic Metals in Urine

Abstract

Loss of renal function is a symptom of chronic kidney disease (CKD). Numerous lives are lost each year as a result of the disease, which frequently has symptoms that are not obvious. Since kidney disease is a progressive condition, it spreads quickly and eventually causes a decline in kidney function. Using factors like age, bacteria, diabetes mellitus, and white blood cell count to name a few, previous studies predicted kidney illness, but they did not incorporate toxic metals in urine. Using machine learning methods, CKD may be predicted with a high degree of accuracy in medical investigations. Thus, this study uses three of the supervised classification methods learning algorithms, which include, Decision tree, Random Forest, and K Nearest Neighbor (KNN) algorithms for the prediction of CKD. A raw dataset of hazardous metals in urine that was gathered from two Nigerian hospitals was used to analyze the performance of the algorithms' predictions. Copper (Cu), Zinc (Zn), Iron (Fe), Lead (Pb), Magnesium (Mg), Chromium (Cr), Cadmium (Cd), and Selenium (Se) are some of these metals. Based on how accurate the results were, a performance analysis was carried out. According to the comparison of the algorithms, the KNN method, which has the highest accuracy, is ideal for predicting chronic renal disease. This work created a system to predict CKD along five classifications using the KNN algorithm.