The aim of the study was to predict the permeability of the Eustachian tube by means of neural computing-based networks in patients with sekelary otic pathology, in view of their selection for subsequent surgical functional operation, once known that the success of such procedures is assured, among others, by a normal functioning of tuba auditiva. To this end, neural networks with simple topologies have been designed, that is feed-forward type networks with one or two hidden layers. Highly encouraging results have been obtained in the validation stage, the probability of obtaining correct responses being of 90 – 100%. Special attention was paid to the quality of the resulting models and to their possible ability of reproducing, through simulation, new behaviour types, unknown in the development stage (i.e., the generalization capacity)
Keywords:- artificial neural networks
- neural computing-based
- tuba auditiva