An approach to recognize temporal relations between Chinese events

Abstract

The research on temporal relations between events plays an important role in natural language processing tasks, such as information extraction, question answering and text summarization. In this paper, we first annotate a document-level corpus to be used for the recognition of temporal relations between Chinese events. We then introduce several effective features according to the characteristics of Chinese, such as trigger semantics, special words, event arguments, event co-reference relation, etc. to improve system performance over the baseline. The experimental results on our annotated corpus show that our system outperforms the baseline by 3.55% of relative improvement in F1.