[Class II malocclusions studied by neural networks]

The aim of this paper is to study how a population is turning when the facial growth regulation is optimal. For this purpose we are using an artificial neural network. Two samples are used: one comprising 100 individuals with ideal occlusion, and another one comprising 500 patients with basal Class II. The first one is used, during the training phase, to teach the network the rules of an optimal facial growth regulation. This network is then applied to the second sample to study how this sample is turning when the facial growth is perfectly regulated. Two samples are obtained after calculation, one measured and one virtual. Both have the same growth potential, the difference lies in how this potential is used. In the virtual sample, some groups of patients disappear and large differences appear between these two samples. But still, one third of the patients in the virtual sample remain their basal Class II.