Interval Type-2 Dynamic Fuzzy Neural Network with Tensor Inverse

Fuzzy neural network is a research hotspot in the field of artificial intelligence, for the dynamic fuzzy neural network, it dynamically prunes the nodes according to the performance index, which proves that it can enhance the training model with dynamic parameters. However, due to over-fitting, the test error is large and the applicability in practical application is low. Therefore, this paper applies interval type 2 fuzzy sets to dynamic fuzzy neural networks with new consequent learning algorithm, and proposes an interval type 2 dynamic fuzzy neural network with tensor inversion, the test error is reduced, and the antiinterference ability of the network is improved with the hybrided strategy, comparable results are obtained among many intelligent algorithms.