Plane Geometry Machine Judgement Based On Symbolic Computation

With the development of technology,the traditional domain is undergoing a transformation to intelligence,and the rapid development of the Internet and artificial intelligence makes this change faster.“Internet” + AI has become the trend of future development,especially in the domain of education.As the strategic foundation of the country,utilizing the advantages of the two can promote the development of education.In recent years,online education has become increasingly popular,people began to use computers and other tools to carry out teaching,tests,etc.In traditional scoring work,the process of manual review and correction is burdensome,which consumes a lot of time,and is easy to cause correction errors.For the correction of objective questions,there are many automated correction systems,which are widely used in the school’s marking process,but the main reason for the complexity of scoring – the judgment of subjective questions,still relies on manual inspection.With the help of “Internet” + AI technology,the automatic scoring of subjective questions can greatly reduce the heavy labor and promote the development of intelligence education.On the other hand,thanks to the fast and convenient automatic scoring,students can get Feedback immediately to better assist self-examination and improve learning efficiency.This thesis is dedicated to the research of machine-based scoring of plane geometry in junior high school.The purpose of the research is to design an automatic scoring system of plane geometry,so as to achieve the correctness and logical judgement of students’ answers.The research involves the processing of mathematical concepts,algebra-based correctness judgement,logical judgement based on AOV network,algebraic logic verification based on variable relaxation,and revers reasoning based on theorem instantiation.The research is as follows:1.Mathematical concept processing.Use natural language processing(NLP)technology to achieve the extraction of mathematical entities and relationships in the answer,and convert them into computer-understandable representations through coordinatization and algebraization of geometric concepts.2.Correctness judgment based on algebraization.For the algebraized students’ answers,based on the coordinate information,use the symbolic calculation tool Maple to make a correct judgement on a single student’s answer.3.Logical judgement based on AOV network.Perform equivalent preprocessing on students’ correct answers and map them to the standard answers.According to the key steps in the standard answer and the AOV network generated by the derivation information,get the logical judgement result.4.Verification of algebraic logic based on variable relaxation.Algebraic logic verification based on variable relaxation is used to verify the students’ answers that cannot be determined in the previous logical judgement.The verification process belongs to forward logic reasoning,with symbolic calculation as the core,constructing pre-steps as input,calculating whether the verification conclusion is valid,and obtaining the forward logic judgments result.5.Reversing reasoning based on instantiation of theorem.Pre-construct the instantiated theorem knowledge base,start from the students’ conclusion,reverse search whether the premise of the conclusion is complete,and obtain the reverse logic judgement result.