Abstract:
Diabetes is a prevalent chronic disease with significant health implications worldwide. It
is usually prolonged in a patient for their entire vitality. Early detection and intervention
are vital for successfully managing and preventing any complications. Diabetes can lead to
complications if not recognized and diagnosed early enough. In this dissertation, I will be
talking about how machine-learning methods are crucial for predictive modeling. Aimed
at early detection of diabetes. These models will be based on different factors, including
demographic, clinical, and lifestyle, among others, with large datasets being used to come
up with them. Therefore, the author prefers using machine learning methods such as SVM,
KNN, ANN, Naive Bayes, logistic regression, XGB Classifier, and Decision Tree. The
results are evaluated using performance measures including recall, accuracy, precision, and
the F-measure, which are computed from the confusion matrix. I designed a predictive
model to identify whether a patient will develop diabetes, utilizing specific diagnosis
measurements in the dataset. This project tries to find a way to improve healthcare
outcomes by enabling early intervention and enhanced disease management.
