Learning ensemble classifiers for diabetic retinopathy assessment
作者：
Highlights：
• Two ensemble classifiers are proposed for the diagnosis of diabetic retinopathy.
• Classifiers are based on medical attributes available in the health care record.
• Methods generate linguistic rules with two approaches: fuzzy and rough sets.
• The best achieved accuracy is 84%.
• Using these classifiers for decision support may avoid unnecessary medical tests.
摘要
•Two ensemble classifiers are proposed for the diagnosis of diabetic retinopathy.•Classifiers are based on medical attributes available in the health care record.•Methods generate linguistic rules with two approaches: fuzzy and rough sets.•The best achieved accuracy is 84%.•Using these classifiers for decision support may avoid unnecessary medical tests.
论文关键词：Diabetic retinopathy,Decision support systems,Rule-based models,Fuzzy decision trees,Random forest,Ensemble classifiers,Dominance-based rough set approach,Class imbalance
论文评审过程：Received 8 February 2017, Revised 29 August 2017, Accepted 13 September 2017, Available online 6 October 2017, Version of Record 16 March 2018.
论文官网地址：https://doi.org/10.1016/j.artmed.2017.09.006
