K-pruning algorithm for semantic relevancy calculating model of natural language
Liang Yan-jun · Jisuanji gongcheng yu sheji · 2013
In order to use semantics more effectively in natural language processing,a semantic relevancy calculating model of natural language is proposed,and the k-pruning algorithm is proposed for solving the model.In the model,the best parsing process for a clause could be determined by the value of semantic relevancy of the clause.The two-level semantic structure of a clause are analyzed,and two grammar rules are used to describe the two-level semantic structure.In the process of solving the model,a state tree would be generated;the k-pruning algorithm could be used to delete the states with less semantic relevancy when searching the state tree,and the computational complexity is effectively reduced and the approximate solution could be accurately acquired.In the experiments,the results demonstrate that the algorithm is effective in solving the model.