Inferring Metaphoric Structure from Financial Articles Using Bayesian Sparse Models

Abstract

Drawing from a large corpus (17,000+ arti­cles) of financial news, we perform a Bayesian sparse model analysis of the argu­ment-distributions of the UP and DOWN-verbs, used to describe movements in indices, stocks and shares. Previous work, by Gerow and Keane (2011a, 2011b, 2011c), has shown, that metaphor hierarchies and antonymic relations can be found in this data. In the present paper, we re-analyzed their data using a Bayesian sparse model (Lake & Tenen­baum, 2010) in order to infer the metaphor space as a uniform representation, based on the argument distributions. Therefore, we treated arguments as fea­tures of metaphors. Our model learned three dimensional graphs in an unsu­pervised manner as sparse representations of the meta­phoric structure over all argument distributions, in parallel. Doing so, it also successfully indicates the metaphoric hierarchies and antonymy relations, that were found by the previous models. In conclusion, we discuss the benefits of this approach.


Back to Table of Contents