The main challenge is to find the optimal architecture of this
type of MNNs, which means finding out the optimal number of
modules, layers and nodes of the neural network with a fuzzy gravitational
search algorithm for pattern recognition and we also present
a comparison with the original gravitational approach. We are
using a modular neural network in this paper and Fig. 5 shows the
architecture used for this work and is described as follows: we
have 2 modules, where each module has one layer for the input
of the data, 2 hidden layers are used and one layer in the output
of the modular neural network and finally, the recognition of the
echocardiograms in the last block of this architecture.