Data mining method of evaluating classifier prediction accuracy in retinal data

• Published in: 2012 IEEE International Conference on Computational Intelligence and Computing Research pp. 1 - 4
• Main Authors: Ramani, R. Geetha, Lakshmi, B., Jacob, Shomona Gracia
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2012
• Subjects: Classifier prediction accuracy; Data mining; Diabetic Retinopathy
• ISBN: 1467313424; 9781467313421
• DOI: 10.1109/ICCIC.2012.6510290

The research in recent years emphasizes the application of computational techniques in the field of ophthalmology. Diabetic Retinopathy, a retinal disease is the major cause of blindness. Early detection can help in treatment but regular screening for early detection has been a highly labor - and resource-intensive task. Hence automatic detection of the diseases through computational techniques would be a great social cause. In this paper, the classifiers used for the automatic detection of the disease are evaluated using the data mining methods. The prediction accuracy of all the classifiers, evaluated using various evaluation methods is presented. Our results show that a training accuracy of 100% can be achieved by a few classifiers and a prediction accuracy of 76.67%.