Open AccessData Mining Technique for Diabetes Diagnosis using Classification Algorithms
Author(s) -
M. Priya*,
M. Karthikeyan
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.d4429.118419
Subject(s) - naive bayes classifier , logistic regression , decision tree , diabetes mellitus , statistical classification , receiver operating characteristic , algorithm , decision tree learning , computer science , id3 , artificial intelligence , bayes' theorem , data mining , machine learning , medicine , mathematics , pattern recognition (psychology) , statistics , support vector machine , bayesian probability , endocrinology
Diabetes mellitus is defined as a one of the chronic and deadliest diseases which combined with abnormally high level of sugar (glucose) in the blood. The classification technique helps in diagnosis the symptoms at starting stages. This paper focused to prognosticate the chance of diabetes in patients with extremely correct classification of Diabetes. The classification algorithms viz., Naïve Bayes, Logistic Regression, and Decision Tree can be used to detect diabetes at an early stage. The algorithm performances are evaluated based on various measures like Recall, Precision, and F-Measure. Experiments are conducted where the time complexity of each of the algorithm is measured. Accuracy is also measured over correct classification and misclassification instances, observed that a Logistic Regression algorithm has much better performance when compared to the other type classifications. Using Receiver Operating Characteristic curves the results are verified in a systematic manner.