DEPTH LIMITED SEARCH AND PRUNING IN SYNTAX ANALYSIS OF NATURAL LANGUAGE PROCESSING
自然语言句法分析的有界深度控制和早期剪枝

Because of large vocabulary and high complexity, it is difficult and almost impossible to make use of LL, LR and Well Formed Substring Table techniques to speed up syntactic analysis in Natural Language Processing. Based on the traditional Top-Down analysis of rules and ATN network, a depth limited search and early pruning technique is proposed and discussed in this paper, which greatly improves the efficiency of natural language parser, by restricting the search scope and reducing backtracking.