A Comparison of Different Prediction Models in the ?Progression of Ocular hypertension to Primary Open ?Angle Glaucoma

The issue of risk assessment in glaucoma has ‎received increasing attention in the past few ‎years. Predictive models are in order to ‎estimate the risk that patients with ocular ‎hypertension will develop to primary open ‎angle glaucoma (POAG) if left untreated. ‎These models are based on classification ‎techniques on the risk factors. Classification is ‎accomplished using conventional risk factors ‎besides retinal nerve fiber layer (RNFL) ‎thickness. It was found that RNFL is sensitive ‎to glaucomatous damage by using different ‎classification algorithms in order to reach to ‎best prediction model.‎ We have applied the Decision tree (DT), Fuzzy ‎logic and Neural Network to the glaucoma ‎problem. The performances of the various ‎classifiers are compared by the area under the ‎receiver operating characteristics curve ‎(AUROC) and the accuracy. The decision tree ‎classifier gives the best result with accuracy 8o% for the training dataset, ‎68.7% for testing data set with AUROC 0.868. ‎