Resolution of ellipses in coordinate noun phrases

Abstract

A coordinate noun phrases connects two words or phrases together via a coordinate conjunction. The duplicate words of conjunctions are mainly omitted. This structure often occurs in the science literature. And this makes it difficult to understand of the sentence. Our research is motivated by the need to reduce the costs of misunderstandings that can occur during NER. We propose a method for resolving coordinate noun phrases with simple or complex ellipses using rules and dataset. And we describe a method to automatically build dataset. This method is applicable to a general-purpose in various fields. Our dataset effectively is used to distinguish between high and low modifier attachment. We reported on a set of experimental results to evaluate the performance of our approach. The results show that our system can efficiently resolve coordinate noun phrases. And we are sure that the method can resolve ellipses in various domains.