構文的予測の分析とその構文解析への応用

概要

論文の詳細を見る
We focus our attention on syntax of natural languages,particularly English,and predictions or expectations that can be made solely with syntactic information during the sentence recognition process are analyzed in detail. We also present machine executable mechanisms that enable proper handling of analyzed aspects and a description method of the mechanisms as grammar rules. We reinforce Kuno's predictive analyzer by introducing five types of predictions. For each type of prediction,we discuss and present its necessity,its description method,and recognition mechanism. We make use of three kinds of stacks whose behavior is specified by grammar rules in an extended version of Greibach normal form. The extended mechanism provides us with a simple yet still powerfull means for recognition of sentence structures. The recognition method can be seen as a deterministic one if we permit looking some words ahead. We also investigate other factors that affect the predictive recongnition process,i.e.,preferences among syntactic ambiguities and necessary amount of look ahead. These factors as well as the proposed handling mechanisms of predictions are tested by analyzing four kinds of articles. In our experiment,about seventy percent of sentences are recognized and looking two words ahead seems to be the critical length for the predictive recognition.
社団法人人工知能学会の論文
1986-12-01