The Research And Implementation Of Intelligent Paper Testing System Based On IRT

In recent years,with the great change in students' learning style brought by the rapid development of computer and network technology,the Internet has broken the limits of time and space in the traditional ed-ucation,convenient students' autonomous learning anytime and anywhere.At the same time,the traditional tests based on paper and pencil have al-ready not satisfied the needs of diversified education nowadays.Comput-er adaptive test is a new breakthrough.It is the product by fully combin-ing the modern education measure theory and the computer technolo-gy and has received more and more application and attention with the development of computer technology.At present,the computer adaptive test still exists some problems,such as initial condition installation,abil-ity estimation,topic selection strategy,etc.Although computer adaptive test can generate test papers according to the conditions set by users,but some of them are of no help to improve students' grades.There are two primary reasons:one is because the value system has no update ability.When students began the first test,the test system generate paper based on students' current ability or the default value.Students have already grasped some difficult knowledge points by passing the test,but the next paper from the system is same as the first.Therefore,the examination papers make no sense.The other is because the system's generated paper is not always the best for multiple papers meeting requirements.To solve these problems,this article puts forward the test system which is based on the item re-sponse theory and combines maximum likelihood estimation with the largest amount of information.This article analyzes the basic principles,basic assumptions and item response models of the item response theo-ry first,analyzing the advantages of choosing the Logistic model.The maximum likelihood method is used in this paper to estimate the students'abilities and frequent tests to amend students' ability value.Adaptive testing system generates test papers according to the conditions set by us-ers.Because the question bank has multiple qualified questions,sys-tem generatesmultiple papers according to these conditions and finally uses the maximum amount of information method to calculate the amount of information of the test papers.It will choose the paper con-taining the largest amount of information as test paper,so as to achieve the aim of testing students.Students,response to the paper will be sent a feedback to the system,then the system will adjust the students' ability value,and prepare for the next exam.