How to construct a Hopfield network

         A term is a word or expression that has a precise meaning in peculiar to a subject. Lin (Lin & Chen, 1996) refers terms extracting from a corpus as concept descriptors. The co-occurrence analysis module measures the similarities between the extracted terms. The extracted terms and their similarities are then modelled as the nodes and the weights between the nodes of the Hopfield network. The Hopfield network resembles an associate network of the extracted English and Chinese terms from the parallel corpus and the classifying behaviour of the Hopfield network converge the strongly associated neighbors to generate the concept space. A concept space is a cluster of related concept descriptors to represent the subject matter of a corpus (Lin & Chen, 1996).