Abstract
The new model of a growing fuzzy inference neural network is based on the main principle of growing neural structures. The weights between hidden neurons and the output of the network can be found very rapidly.
A new technique for the selection of significant input features also supports extraction of a small set of representative fuzzy rules.