Neuron networks on the basis of the metric methods of recognition In application to tasks with fuzzy inference

In the work the transformation algorithm of the task of fuzzy inference into the task of recognition with the application of neuron networks of those realizing the metric methods of recognition is proposed. The transformation of the collection of fuzzy sets and fuzzy rules into the collection of sample curves is carried out for this. On the basis of the obtained collection of curves the creation of neuron networks on the basis of the metric methods of recognition is proposed in application to tasks with the fuzzy inference. Are shown the possibilities of the optimization of the obtained networks in application to tasks with the fuzzy inference and the possibility of the multitask use of united architecture of neuron network. Is given the algorithm for learning that provides the selection of expert opinion.