Intuitionistic fuzzy neural network with filtering functions. An index matrix interpretation

Biological neurons and their connection in neural networks have motivated the creation of the architecture of artificial neural networks. In the previously considered cases, the description of the neural networks and their connections are described with standard matrices where the values for the weighting coefficients and biases are placed. By recalculating them, the artificial neural network is trained. The paper presents an approach for describing multilayer neural networks with Intuitionistic Fuzzy Index Matrix (IFIM). The neural network input was described in IFIM form, then the weight coefficients of the connections between the nodes of the input vector, and then activation functions of the neurons. The use of IFIM extends the understanding and description as well as the structure and use of multilayer neural networks.