A comprehensive analysis of eye diseases and medical data classification

Abstract

Vision loss is a critical health issue that presents substantial challenges to both individuals and communities. For those affected, it can lead to difficulties in performing daily activities, hinder educational and employment opportunities, and significantly impact mental health and overall quality of life. The inability to see can also lead to increased dependence on others, creating emotional and financial strains on families and caregivers. This paper highlights the benefit of machine learning (ML) in exploring conditions that significantly affect vision loss. The goals that will be achieved in this paper are to determine the best classifier capable of dealing with medical datasets and to determine the best strategy for dealing with medical data. Determine which feature selection is most applicable to use for examining medical data. Two medical datasets, 4 strategies, 19 classifiers, and 2 feature selections were used. As for the best classifier, the stochastic gradient descent (SGD) model was the best in dataset 1 and 2. The function strategy showed the best performance, followed by the rules strategy. CorrelationAttributeEval was shown to be the best feature selection, while ClassifierAttributeEval was the second-best feature selection.