Classifiers differences method for experimental modeling of intraocular pressure altitudes versus ultrasonic corneal thickness variables

Abstract

The amplitude of intraocular pressure (IOP) variation was defined via experimental modeling, conditioned by age, sex, complementary diseases, pharmaceuticals, exposure to allergens, etc. The subject of our study was to provide additional numerical analysis for fixing the reliability of the measured IOP magnitudes with Goldmann applanation tonometer (GAT) and ultrasonically measured central cornea thickness (CCT), while employing the experimental data of created correlation matrix and artificial neural networks via classifiers differences method. The clinical experimental data were restricted by 99 eyes of 53 subjects (65 ocular hypertension cases and 34 somatically healthy eyes). Experimental data in terms of IOP, CCT, age matrix indicated that variation of ultrasonically measured central corneal thickness is a positively correlated source of variation in IOP measurements among ocular hypertension subjects (R=0.648, p=0.073). Optimizing the distribution among classes, i.e. dominance of IOP error and the absence of IOP error, the fitting of radial basis function (RBF) network and multilayer perceptron (MLP) was provided and minimal error of the networks was obtained.