Glaucoma Classification Model Based on GDx VCC Measured Parameters by Decision Tree

This study is to propose and evaluate the diagnostic accuracy of decision tree classifiers using the full set of standard GDx VCC measurements for classifying glaucoma in a Taiwan Chinese population. The classifiers were trained and tested using standard GDx VCC parameters from examinations of 74 subjects with glaucoma and 72 normal subjects. Six promising decision rules were generated from decision tree methods and the overall accuracy from tenfold cross validation was 0.801. Classification tree based on GDx VCC data promises to be a diagnostic tool in glaucoma disease. However, its exact clinical application in glaucoma practice should be retested. Further longitudinal study should address this issue.