Diabetes Classification Using ID3 and Naïve Bayes Algorithms

Abstract

Diabetes can be defined as a chronic disease identified by high levels of blood glucose that result from issues in the way insulin is generated, the way insulin works, or both those reasons. The aim of this research is to propose a technique using the Decision Tree (ID3) and Naive Bayes to categorize diabetes and reduce classification errors by increasing the accuracy of the classification. The results of the proposed method were evaluated by comparing them with other results through the application of the proposed system to Pima India Diabetes data set, obtained from the UCI database site. The experimental results show that the ID3 recorded a precision ratio of 91% and the naive class corrected it to 94% for the same number of the test group.