| With the rapid development of computer technology and artificial intelligence,paperless examination has become an inevitable trend in the development of information technology.Automatic test paper can not only ensure the objectivity and impartiality of the examination,but also reduce the workload of compiling and marking the examination papers in the process of examination,and at the same time,it can save a lot of human and material resources,too.However,the existing test paper algorithm still needs to be improved in terms of the quality of test paper,and it rarely analyzes and evaluates the test paper quality.Aiming at the above problems,according to the classical test theory,the genetic algorithm is selected as the test paper algorithm,the test paper is generated through the simulation experiments,taking the difficulty,distinction,reliability and validity of the classical test theory as the evaluation index,the quality of the test paper is tested.The results show that the quality of the test papers is satisfied,and the results are also subject to normal distribution.The main work includes the following four aspects.Firstly,the relevant literature of automatic test papers is combined,the related theory of automatic test paper is analyzed,and the constraints of automatic test paper is determined;the mathematical model of automatic test paper is constructed.Through the analysis of the theory of automatic test paper,the classical measurement theory is selected as guidance.The number of test questions,the scores of test papers,the type of test questions,the chapters of test questions and the difficulty of test questions are determined as the constraints of the test paper,and a mathematical model of automatic test paper is constructed.Secondly,select and design the automatic test paper algorithm.Through comparing and analyzing the advantages,disadvantages and the application scope of the commonly used test paper algorithms,the genetic algorithm is selected as the test paper algorithm for this study.The genetic algorithm is designed.In the chromosome-coding scheme,the real number coding is adopted,the roulette selection method is used in the selection operator,and the group single point crossover method is adopted in the crossover operator.The segment gene mutation patterns are used in the mutation operator.Through the design of genetic algorithm,it lays a good foundation for the application of genetic algorithm in automatic test paper.Thirdly,the automatic test algorithm based on genetic algorithm is applied to the actual composition process of "Computer Basics" course to test the effect of the algorithm.According to the established "Computer Basics" database.Moreover,simulation tests were conducted using MATLAB to generate 10 test papers.The statistics and analysis of the chapters and difficulty of the 10 papers show that using the genetic algorithm has a good effect.The designed genetic algorithm is compared with the random extraction algorithm and the ant colony algorithm.The results show that the test paper generated by the genetic algorithm is more suitable for the actual test paper than the other two algorithms.Fourthly,the generated test papers are used in the actual examination process,and the quality of the test papers is tested according to test scores.Five papers is selected to be applied to the actual mock examinations.According to the test scores,the test paper quality is evaluated by the difficulty,differentiation,reliability and validity of the classic measurement theory.The normal distribution test is performed on the scores.The results showed that the results obey the normal distribution and can meet the requirements of the test paper quality. |
PDF Full Text Request