Natural language generation method using automatically constructed lexical resources

In this paper, we propone a natural language generation method baaed on automatically constructed lexical resources. Many conventional approaches in sentence generation use manually constructed templates. Therefore, the variety of available sentences depends heavily on the quality and quantity of the templates, and the cost to construct these templates is very high. The proposed sentence generation method uses large-scale case frames and Google N-gram, which both are compiled automatically from Web documents. The proposed method uses words as an input. It generates a sentence from case frames, using Google N-gram as to consider co-occurrence frequency between words. Since we only use lexical resources which are constructed automatically, the proposed method has high coverage compared with the other methods using manually constructed templates. We carried out experiments to examine the quality of generated sentences and obtained satisfactory results.