A Utility Global Optimization Algorithm of Syntactical Parsing PCFG
Liang Xiong-you · Journal of Changchun University of Science and Technology · 2010
This article points out it isn't very feasible for disambiguation of sentence pattern by using Probabilistic Context Free Grammar or Stochastic Context Free Grammar,and solves some frequently asked sentence pattern ambiguity problems effectively based on Hownet and sentence structure semantic relevancy through contextual information.In order to correct the hidden mistakes when each stage gained most superior separately,founded on experience,this article proposes a global optimization algorithm that unifies rough word segmentation method of Nshortest-path,method of new words recognizing by using the dynamic programming,method of disambiguating interlink overlapping combinatorial and compound ambiguity fields,N-shortest path Viterbi method of rough part-of-speech tagging,method of word sense tagging based on Hownet,method of adjusted PCFG,so may realize the multi-stage optimal goal.It may solve the problems of each stage gained most superior separately.