作者: Constantin Orasan , Dalila Mekhaldi , Laura Hasler , Shiyan Ou
DOI:
关键词:
摘要: This paper presents an automatic question pattern generation method for ontology-based answering with the use of textual entailment. In this method, a set patterns, called predictive questions, which are predicted to be asked by users in domain, were generated on basis domain ontology. Their corresponding query templates, can used extract answers questions from knowledge base, as well. The process producing these patterns and templates is described discussed context “Movies & Cinemas” domain. An evaluation was carried out assess quality help entailment engine.