A Choice Questions Generator Tool For Engineering Conceptions Based On Knowledge Graph

The traditional question generation relying on human resources has the disadvantages of large consumption of human resources and insufficient diversity of questions,which is difficult to meet the needs of high efficiency and diversity of actual question generation.As a kind of structured semantic knowledge base,knowledge graph has important application in the field of automatic generation of multiple choice questions.The existing multiple choice questions based on knowledge graph often focus on the objective statement of facts of a single knowledge point,but engineering fields pay more attention to the comparison,understanding and discrimination of relevant knowledge points,so it is difficult to directly use the existing multiple choice questions generation technology.Therefore,it is necessary to combine the characteristics of engineering specialty field,based on the design of knowledge graph,to study the technology of multiple choice questions generation suitable for engineering specialty field.This thesis designs and implements a tool system based on knowledge map to generate multiple choice questions for engineering majors.In order to design multiple choice questions generation technology suitable for engineering field,the basic data structure of knowledge graph must be designed to express and store the complex knowledge system in engineering specialty field.Taking the course "computer network" as an example,this thesis analyzes the basic strategies of engineering concept questions,and summarizes three types of common multiple choice questions,including knowledge point description discrimination,multi knowledge point discrimination and logic discrimination.Firstly,a mounting scheme of extended sentences is proposed to design the basic knowledge graph,and a disturbance item generation strategy of pseudo proposition construction and concept misappropriation is designed to realize the generation of the first two types of multiple choice questions.Then,based on the basic knowledge graph,a scheme of knowledge point attribute and relation representation is proposed to design the extended knowledge graph,and a strategy of generating disturbing items by keyword substitution is designed to realize the generation of three types of multiple choice questions.Finally,in order to assist teachers to choose appropriate questions,the thesis puts forward the question difficulty evaluation strategy based on the knowledge difficulty and other factors,designs the evaluation model and the revision scheme of the knowledge difficulty value,and realizes the difficulty evaluation of various types of questions.In this thesis,the computer network course is selected as the experimental object,the basic knowledge map and the extended knowledge map of the computer network course are constructed respectively after sorting out the course knowledge,and the supported question types are tested respectively.The experimental results show that the knowledge graph designed in this thesis and the corresponding question generation strategy can generate the expected randomized questions.These questions are used in the actual teaching activities,and the actual difficulty of the questions is calculated by counting the results of students,which verifies the feasibility of the difficulty assessment and revision model.The research work of this thesis can provide automatic means for college professional course teachers to generate questions,and provide the evaluation index for the difficulty of questions to assist teachers to choose appropriate questions.