Optimizing Neuron Function Based on Entropy Clustering in Functional Networks

Functional networks are the extension of neural networks which have been studied recently. Like neural networks, there is no systematic method for designing approximation functional network structures. In this paper, a new entropy clustering method designed for functional networks is presented, which combines each neuron function and functional parameters by performing the optimal search to achieve the learning between functional network structures and the functional parameters. The simulation results indicate that the proposed method can produce more rational structure and greatly improve convergent precision of functional networks.