UTD-HLT-CG

Abstract

In this paper we present a semantic architecture that was employed for processing two different SemEval 2007 tasks: Task 4 (Classification of Semantic Relations between Nominals) and Task 8 (Metonymy Resolution). The architecture uses multiple forms of syntactic, lexical, and semantic information to inform a classification-based approach that generates a different model for each machine learning algorithm that implements the classification. We used decision trees, decision rules, logistic regression and lazy classifiers. A voting module selects the best performing module for each task evaluated in SemEval 2007. The paper details the results obtained when using the semantic architecture.