Semantic Role Labeling Using Support Vector Machines

摘要

In this paper, we describe our systems forthe CoNLL-2005 shared task. The aim ofthe task is semantic role labeling using amachine-learning algorithm. We apply theSupport Vector Machines to the task. Weadded new features based on full parsesand manually categorized words. We alsoreport on system performance and whateffect the newly added features had.