Key points are not available for this paper at this time.
Current approaches to supervised learning of metaphor tend to use sophisticated fea-tures and restrict their attention to con-structions and contexts where these fea-tures apply. In this paper, we describe the development of a supervised learning sys-tem to classify all content words in a run-ning text as either being used metaphori-cally or not. We start by examining the performance of a simple unigram baseline that achieves surprisingly good results for some of the datasets. We then show how the recall of the system can be improved over this strong baseline. 1
Klebanov et al. (Wed,) studied this question.