Abstract:
Kidney disease is a chronic condition because it causes a progressive loss of kidney function over time. When kidney function is impaired, the body can no longer perform all of these essential functions, leading to a range of symptoms, including swelling, fatigue, and shortness of breath. That's why early detection of chronic kidney disease is important because it allows for early intervention, which can slow the progression of the disease and prevent or delay the development of end-stage renal failure. Early detection can also help identify treatments to reduce or prevent further damage to the kidneys. The study aims to develop an automated system to predict CKD, identify the stage of CKD, find out the probability of CKD, predict CKD progression, and manage chronic kidney disease. The study proposes a machine-learning-based system using supervised learning algorithms to predict the onset of CKD accurately. The system is designed to use a combination of patient data, such as age, gender, ethnicity, family history, lifestyle, and laboratory test results, to predict and determine the probability of the onset of CKD. In addition, the system is designed to provide personalized recommendations for lifestyle modifications and all the relevant information related to CKD, such as symptoms, risk factors, causes, nutrition, diagnosis and testing, and treatments, and to create awareness about CKD. The study also proposes using interactive visualizations to help patients understand their condition and make informed decisions about their health. The study also outlines a procedure for validating the accuracy of the system and provides recommendations for further development. To predict the CKD, we have applied different machine learning algorithms, including Decision Tree, Ada Boost, KNN, Random Forest, Gradient Boosting, Stochastic Gradient Boosting, XGBoost, and Extra Tree. Of these, Decision Tree and AdaBoost have higher accuracy, at 96.25% each.