Machine-learning and data mining techniques using hybrid system were used to accurately predict the development of diseases such as Chronic Kidney Disease (CKD) and Acute Renal Failure (ARF). In this study, Random Forests Decision Algorithm, Autoregressive Integrated Moving Average (ARIMA) Model and K-means Clustering Algorithm were used to predict the degree of urgency and progression of dialysis from patients’ electronic medical records. The use of such algorithms will provide a predictive model for forecasting the urgency level and CKD stages, clustering by gender, age, CKD stages and urgency level to anticipate adverse events that will help medical practitioners in the efficiency and accuracy of detecting the severity of the kidney disease. 20,000 instances were divided into training and testing data, wherein the data was able to label the urgency and progression of dialysis. Apart from this, the stages of CKD and urgency level were forecasted using ARIMA Model. The extracted pattern from the historical and current data predicted the urgency and progression of dialysis, thus a prototype software implementation was also proposed. The experimental results of the study show that 99 percent (%) of the prediction on the degree of urgency and progression of dialysis model deemed accurate, paving way to a better clinical decision-making process of nephrologists using a rule-based system from the important attributes of the patient’s electronic medical records which will also help improve a patient’s quality of life.
Read full abstract