Opinion Target Extraction in Chinese News Comments

News Comments on the web express readers’ attitudes or opinions about an event or object in the corresponding news article. And opinion target extraction from news comments is very important for many useful Web applications. However, many sentences in the comments are irregular and informal, and sometimes the opinion targets are implicit. Thus the task is very challenging and it has not been investigated yet. In this paper, we propose a new approach to uniformly extracting explicit and implicit opinion targets from news comments by using Centering Theory. The approach uses global information in news articles as well as contextual information in adjacent sentences of comments. Our experimental results verify the effectiveness of the proposed approach.