Application Of Machine Learning In The Prediction Of Chronic Kidney Disease

Abstract

Chronic kidney disease also known as CKD for short is a condition characterized by progressive loss of kidney function over time, affecting the body’s overall health due to the loss of ability to filter waste. An incomplete or late diagnosis and treatment will eventually cause a patient to develop endstage renal disease, which is fatal. CKD piqued interest of many researchers that have decided to use machine learning to predict CKD in an early stage thus minimizing fatality. This paper also aims to assess and compare machine learning models including Logistic regression and LightGBM’s role in improving the prognosis of Chronic Kidney Disease by analyzing health conditions using a dataset that consists of 1659 patients information and 54 attributes providing a full picture of patients’ health. This study also tackles the challenge of having imbalanced data through sampling techniques like Synthetic Minority Oversampling Technique (SMOTE), this approach ensured that the model predicts both negative and positive cases of CKD, with an accuracy of $98 \%$.

Ähnliche Arbeiten