Semantic analysis in the automation of ER modelling through natural language processing

This paper deals with the problem of extracting semantic knowledge in the production of ER models from natural language specifications. The application of semantic heuristics is proposed as the strategy to obtain the relevant ER elements such as entities, attributes and relationships from the specifications. Earlier research has shown that syntactic heuristics produced good results in identifying the relevant and correct results of the ER elements in terms of recall and precision. The inclusion of the semantic lexical knowledge is hoped to further improve the results. The semantic heuristics may later be implemented as part of a natural language tool in the generation of the ER models.