减少回溯的自然语言理解

Reducing Backtracking with Natural Language Understanding

摘要: 采用DCG(Define Clause Grammar)方法实现的自然语言理解,出现了大量回溯。本文提出的一种方法可使回溯减少到几乎最少的情况,并且可以推广使用到诸如MT(MechineTranslator)等许多领域。该方法的结果已经在IBM-PC微机上得以验证。

Abstract: There is a lot of backtracking by using DCG(Define Clause Grammar) method with NLU (Natural Language Understanding). In this paper, a new method for NLU will be introduced, by which, backtracking can be almost reduced to the least case. And this method can be also used in many fields, such as MT (Machine Trans- lator), etc.. The result has been proved on IBM-PC microcomputer.