Abstract



This paper presents a model based on neural networks that permits to build a conceptual hierarchy in order to approximate functions over a given interval. A new kind of artificial neural networks using bioinspired axoaxonic connections. These connections are based on the idea that the signal weight between two neurons is computed by the output of other neuron. Such model can generate polynomial expressions with lineal activation functions and the degree n of the output depends on the number n ¿ 2 of hidden layers. This network can approximate any pattern set with a polynomial equation, similar to Taylor series approximation. Results concerning function approximation using artificial neural networks based on multilayer perceptrons with axoaxonic connections are shown. This neural system classifies an input pattern as an element belonging to a category or subcategory that the system has, until an exhaustive classification is obtained, that is, a hierarchical neural model. The proposed neural system is not a hierarchy of neural networks; this model establishes relationships among all the different neural networks in order to propagate the neural activation when an external stimulus is presented to the system. Each neural network is in charge of the input pattern recognition to any prototyped class or category, and also in charge of transmitting the activation to other neural networks to be able to continue with the approximation. Therefore, the communication of the neural activation in the system depends on the output of each one of the neural networks, so as the functional links established among the different networks to represent the underlying conceptual hierarchy.  
International

Si 
JCR

No 
Title

Journal of Software Science and Computational Intelligence 
ISBN

19429045 
Impact factor JCR

0 
Impact info


Volume




Journal number

0 
From page

67 
To page

80 
Month

ENERO 
Ranking
