|
Abstract : |
A SNoW based learning approach to shallow parsing tasks is presented and studied experimentally. The shallow parsing method suggested learns to identify syntactic patterns by combining simple predictors to produce a coherent inference. Learned predictors are cascaded and their outcome is used as an input to an inference algorithm that produces the nal phrases. We discuss the requirements from a learning system for this approach to be applicable and present a system based on the SNoW learning architecture that satises these conditions. Two instantiations of this approach are studied and experimental results for Noun-Phrases (NP) and Subject-Verb (SV) phrases that compare favorably with the best published results are presented. In doing that, we compare two ways of modeling the problem of learning to recognize patterns and suggest that shallow parsing patterns are better learned using open/close predictors than using inside/outside predictors. 1, |
