Diabetes diagnosis system using modified Naive Bayes classifier

<span>In today’s world, Diabetes is one of these diseases and now a big growing health problem. The techniques data mining have been widely applied to extract knowledge from medical databases. In this work, Medical Diagnosis system proposed for the ‎diagnosis diabetes in manner ‎that rapid cost-effective. three stages are ‎involved diagnosis (DDS) including: dataset constructing, preprocessing classification algorithm using traditional Naïve Bayesian ‎‎(TNB) modified (MNB)). MNB Classifier NB that used ‎enhance accuracy ‎diagnosis, by adding modest model help separate ‎the overlapping classes. outcome‎ ‎showed classifier generally higher than ‎TNB ‎classifier all sets features. An about (63%) was achieved TNB ‎model, whereas (100%). experimental results showed better ‎NB both two cases constructed ‎datasets; first case filling missing values experiences second ‎missing K-nearest neighbor (KNN) algorithm.</span>