CHRONIC KIDNEY DISEASE PREDICTIONS USING MACHINE LEARNING MODELS

Abstract

When it comes to clinical disorders, chronic kidney disease
(CKD) is an umbrella term that refers to a wide range of illnesses
that deteriorate as kidney function degrades over time. It refers to
a wide variety of medical conditions. The term "chronic renal
failure" is sometimes used to describe this illness in some circles.
Various factors, including genetic abnormalities in the k idneys
and systemic illnesses that damage the kidneys, can contribute to
chronic kidney disease. Depending on the underlying rea son, i t
might express itself in a variety of ways. Worldwide, the number
of people suffering from chronic kidney disease (CKD) is
growing year after year, according to the World Health
Organization. As defined by the World Health Organization,
chronic kidney disease (CKD) is a worldwide public health
concern with an increasing incidence and a vast geographic
reach that affects individuals all over the world. GFR rises in the
presence of renal failure needing dialysis, and it is widely
regarded to be the most reliable overall indicator of kidney
function in the general population. Heart disease (including high
blood pressure and anaemia) and a variety of metabolic
problems, to mention a few, are among the additional risk factors
for kidney failure. Because of a statistical approach known as
10-fold cross-validation, the algorithms of logistic regression,
support vector machines, random forest, and gradient b oo st ing
have all been trained and tested on real-world data. According on
the F1measure gathered by the classifier after training, the
accuracy of the Gradient Boosting classifier is 99.1 percent
correct. In addition, we discovered that haemoglobin has a bigger
significance for both random forest and gradient boosting in the
diagnosis of chronic renal sickness than was previously believed
to be the case, which is in direct opposition to previous notions