METHOD FOR SYNTACTIC PARSING OF NATURAL LANGUAGE

摘要

Disclosed is a method for syntactic parsing of natural language. The present invention addresses technical drawbacks of the Berkeley Parser and the Stanford Parser, two leading natural language syntactic parsers that are internationally recognized in the field of computer science, and discloses a technical solution to resolve these drawbacks. The present invention establishes a novel mathematical model for describing sentences, and proposes a computer-based syntactic parsing method based thereon. The present invention integrates, by using technical means, three aspects of computer processing of natural language, namely, lexical analysis, syntactic parsing, and semantic analysis, in an organic manner, and strengthens mutual constraints of these three aspects, thereby improving the ability of a computer to resolve structural ambiguity. The present invention involves high technical complexity, achieves a high level of integration, has a wide range of applications, involves a large amount of computation, and complies with natural principles of mathematics and computer science, thereby enhancing accuracy of computer syntactic parsing.