Estudo sobre redes neurais artificiais.

Abstract

Conventional computers are efficient in several areas. However, they do not yet perform so satisfactorily when compared to real biological system (human visual system, bats’ sonar, etc). The artificial neural network was developed by the idea of gathering in common characteristics of the human being nervous system. The fact of the neurons be connected to each other, receiving and sending stimuli, creates a network. The great trump of this computational method is efficiency. Such differential is due to the fact that this model provides real time learning, featured by the neurons’ connections. All those facts results in a reason for studying this network. In this context, two artificial neural networking models are discussed, perceptron and Adaline, which are applied in different areas. Therefore, MATLAB software was used as simulation tool.