An Associative Memory Artificial Neural Network System

This paper presents an artificial neural network system using Hopfield and Kosko associative memory algorithms. Both of algorithms are used for recognising patterns from partial/noisy input o r from other associated patterns. The recognition ra te of the system has been effectively improved by adjusting the network's parameters and by modifying the associative memory function for converging to a global minimum. The system developed has been tested using a set of data and the results obtained has proved its efficiency and robustness. The advantage of the Hopfield network can be