Using multiple methods including Naïve Bayes, K- Nearest Neighbours, and Decision Tree Algorithms with Ensemble Learning to diagnose diabetes

Paper ID: EIJTEM_2022_9_3_19-27

Author's Name: Mahek Tikedar, Rallapalli Lakshmi Chandana, Mrs. Beulah J Karthikeyan and Dr. Sankara Sarma KVSSRS

Volume: 9

Issue: 3

Year: 2022

Page No: 19-27

Abstract:

Diabetes is a medical condition characterized by high blood sugar levels caused by insufficient insulin production or the body's inability to respond to insulin effectively. This condition increases the risk of heart disease, stroke, and damage to vital organs such as the kidneys, eyes, nerves, heart, and blood vessels. Various classification techniques are used in medical, business, and industrial applications to diagnose and manage diabetes. Three well- known algorithms - naïve Bayes, k-nearest neighbours, and decision tree - were utilized to construct classification models based on selected features. Naïve Bayes is a statistical classifier that employs Bayes' theorem, while k-nearest neighbours is suitable for large training sets, as it identifies the k closest training points to the unknown object in pattern space. The most popular algorithm, decision tree, is easy to understand and selects the best split attribute as the root node. Finally, popular ensemble learning techniques such as bagging and boosting were applied to the three base classifiers.

Keywords: Machine Learning, Classification Techniques, Naïve Bayes, Regression, Artificial Neural networks, K-nearest neighbours medical Applications, Diabetes detection, Decision tree.

View PDF