Digital technologies have grown tremendously from the past two decades. This has resulted in the increased usage of electronic devices by the people, regardless of age. The most vulnerable among the entire population are the children, who develop ocular defects at a very young age due to the prolonged usage of electronic devices. Myopia is one of the ocular defects that is common among children. There are several factors that contribute to the vision impairment. Using machine learning models, the major factors that cause myopia can be identified. People who are likely to get myopia can be classified using machine learning models. In this paper, myopic data set is classified using various supervised machine learning techniques like logistic regression, decision tree, support vector machine, Naïve Bayes, K-nearest neighbor, random forest and neural network and the best model is determined. Multi-layer perceptron achieved highest accuracy among all the machine learning models. The proposed methodology optimizes the prediction accuracy further by selecting important features through recursive feature elimination with tree-based classifier.
Classification of Myopia in Children using Machine Learning Models with Tree Based Feature Selection
2020-11-05
457607 byte
Conference paper
Electronic Resource
English
Effective feature selection for image steganalysis using extreme learning machine
British Library Online Contents | 2014
|MACHINE LEARNING MODEL DEVELOPMENT WITH INTERACTIVE FEATURE CONSTRUCTION AND SELECTION
European Patent Office | 2020
|