Analysis Of Mathematical Natural Language Structure And Its Application Based On Parser

In recent years,online education has received more and more attention,but the cur-rent online education is basically still centered on tutors.In terms of technology,it is more reflected in the competitive development of education platforms.Therefore,the au-tomatic inference question system that combines artificial intelligence and education has great development and application prospects,and it is of great benefit to the development of society.As an important subject in education,mathematics has high requirements for logical reasoning,and for artificial intelligence,mathematics is also the foundation of its development.The two methods are complementary.The elementary mathematics problem-solving system involved in this project takes mathematics as a breakthrough point,and strives to achieve automatic reasoning and problem-solving of elementary mathematics.In order to automatically reason about math-ematics,you must first be able to fully understand the meaning of the mathematics prob-lem.Therefore,the natural language processing for the mathematical language is essen-tial.This article mainly uses Parser to process mathematical natural language from the perspective of syntactic structure.This article first analyzes the situations that easily lead to confusion in mathemat-ical sentences,and combines the structural characteristics of mathematical sentences to carry out mathematical semantic disambiguation from two aspects:vocabulary and part of speech.After mathematical semantic disambiguation,the syntactic parser Stanford Parser is used to parse the mathematical sentence,and the corresponding dependent structure of the mathematical sentence is obtained,and then the dependent structure is processed to extract the first-order predicate logic information and parallel information.The extracted information is applied to the named entity recognition module in the natural language processing of mathematical sentences to process sentences that require progressive entity types,which improves the accuracy of the naming of mathematical sentence entities,and further improves the accuracy of mathematics questions.Semantic understanding.After organizing the research content and constructing a complete named entity pro-gressive system,this article selected 1000 high school mathematics questions from the elementary mathematics self-built library for comprehensive testing,and finally extracted the first-order predicate logic and parallel information on this system.The accuracy rate has reached 96.51%,and the whole system has good applicability and reliability.