Restricted-domain ontology-based analysis for Chinese interrogative sentences
Xiuli Wei · Journal of Computer Applications · 2013
There are lots of studies on Chinese interrogative sentences employing restricted-domain ontology-based approaches which utilize a structured ontology base as the knowledge base to store the knowledge.They focus on calculating the semantic similarity of interrogative sentences or matching the models of sentence pattern,then they use querying and reasoning of the ontology to return the answers.However,most of the previous studies do not consider the semantics of interrogative sentences and the precision of ontology queries enough.To overcome the above inadequacies,this article emphasized on Part-Of-Speech(POS) tagging and middle term to enhance the structural description of interrogative sentences and the semantic frame of formalization to depict their semantics.After rewriting querying module of Jena technology according to our analyzing description rules about property and equivalent class of ontology in OWL(Web Ontoloby Language) files,the result show the extension and intension of knowledge to user.The experiment was carried out on the Eclipse platform.It turns out that the combination of above methods returns answers that are more precise and reasonable for restricted domains.What is more,it has strong expansibility and does not need expensive corpus to train the model.It can be used effectively in online teaching and other fields.