Detection of the Tonometrical Measurements Error Adapting the Radial Basis Function Method Versus Multilayer Perceptron

Abstract

The values of eye pressure as well as the values of other parameters of the biological environment vary within a wide range, wheeas t rhe amplitude of variation depends on the age, the sex, adjacent diseases, medicines, exposure to allergens, and etc. The computation of the accurately measured eye pressure is the positive perspective of the system for diagnosing the autonomous disease. This research targets to create the system, which models the detection of tonometric measurment error. Modeling the target relations (IOP vs CCT) thresholding classifier and radial-basis function classifier versus multi-layer perceptron classifier was applied.. By numerical experiments, we have proved that the discrete model applied is adequate to the modeling of the expression being analyzed and that the optimal synergy of artificial neural networks and the biosubject is realized with the purpose of minimizing the frequency of diagnostical mistakes in the everyday activities of the clinician. Ill. 5, bibl. 3 (in English; summaries in English, Russian and Lithuanian).