Abstract:
The number of diabetic patients around the world is increasing alarmingly day by day. It
has now become a threat to our human society. This disease is usually caused by eating
heavy sugary foods and not following a proper diet. However, nowadays machine learning
algorithms can be used to easily and accurately predict for diabetics by checking and
sorting out different types of symptoms. This can greatly reduce our mortality rate and
make us more aware of diabetes. The purpose of my work is to make patients aware of and
predict diabetes in advance using machine learning algorithms. I have used three
algorithms for this task- Logistic Regression, Gaussian Naive Bayes, Random Forest. The
overall performance of the three algorithms is evaluated in exceptional steps which
includes accuracy, precision, F1 score, ROC accuracy, Recall, Standard deviation and KFold mean accuracy. Analyzing the all algorithms, it is seen that the random forest
algorithm gave the best result of 86%.
