A SNoW based learning approach to shallow parsing tasks is presented andstudied experimentally. The approach learns to identify syntactic patterns bycombining simple predictors to produce a coherent inference. Two instantiationsof this approach are studied and experimental results for Noun-Phrases (NP) andSubject-Verb (SV) phrases that compare favorably with the best publishedresults are presented. In doing that, we compare two ways of modeling theproblem of learning to recognize patterns and suggest that shallow parsingpatterns are better learned using open/close predictors than usinginside/outside predictors.
展开▼
