Identification and characterisation of choroidal neovascularisation using e-Health data through an optimal classifier. (12th February 2020)

• Record Type: Journal Article
• Authors: Anitha, G.; Ismail, Mohamed; Lakshmanaprabu, S.K.
• Is Part Of: Electronic government. Volume 16:Number 1/2(2020)
• Page Start: 39
• Page End: 55
• Publication Date: 2020-02-12
• Subjects: choroidal neovascularisation -- optical coherence tomography -- OCT -- machine learning classifiers -- support vector machine -- SVM -- kNN classifier -- naive Bayes classifier; Internet in public administration -- Periodicals; 352.380285467805
• Languages: English
• ISSNs: 1740-7494

Abstract

Over the years, health informatics and eHealth gained more popularity in health care application. The collection of eHealth data becomes easier due to the advancement of digital technology. In this paper, the e-Health based supporting system is developed for the classification of a retinal disease called CNV. CNV is a retinal disease caused due to the growth of abnormal blood vessels in the choroidal layer. A good classifier for CNV data makes the process of identifying the disease easier and it will help the medical practitioners to give the treatment at the right time. A comparison has been done among different machine learning classifiers such as support vector machine (SVM), k-nearest neighbours (kNN), neural network (NN), ensemble and naive Bayes classifiers and they are tested and evaluated based on accuracy and training time. From the results, it is observed that kNN classifier outperforms the other classifiers in all aspects.