A WordNet based lexicon model for controlled natural language

The lexicon is an integral part of controlled natural language. In this paper, a WordNet based CNL lexicon model is proposed. First of all, a CNL lexicon model is defined according to English word´s part of speech classification and characteristics. Then the method of constructing CNL lexicon based on WordNet is proposed by utilizing WordNet features. Thirdly, an algorithm for the CNL lexicon parser is designed using lexicon-based context-free grammar. Finally, a lexicon for Attempto Controlled English has been build with the algorithm proposed in this paper. Experiment results show that our approach has two advantages: one is that it can significantly improve recognition rate of CNL, while reducing difficulty of construction and maintenance of the lexicon of a CNL; the other is that the approach proposed is so portable that can be transplanted into other CNL system.