Design of an adaptive examination system based on artificial intelligence recognition model

Abstract

Nowadays, with the change of the times, the traditional paper-and-pencil examination can not satisfy the students' test. For students of different levels, it is difficult to say that the best test effect can be achieved by passing the fixed test. In order to test students' level accurately, this paper proposes an adaptive test system based on item response theory, which integrates AI recognition model. By extracting test questions dynamically, the test time is shortened and the effect of adaptive test for students at different levels is realized. While researching the determination of test parameters, the algorithm of ability evaluation and the rule of examination termination, the system designed in this paper improves the strategy of topic selection. Firstly, a strategy of topic selection for cognitive diagnosis is proposed, which can effectively improve the drawbacks of the existing system using the tested items to evaluate the ability. Secondly, the combination of cognitive diagnosis and artificial intelligence recognition model improves the efficiency and accuracy of the system topic selection. Finally, the Monte Carlo simulation experiment method is used to test the system designed in this paper. Experiments show that the system has good cognitive diagnostic ability, and achieves the efficiency and accuracy of topic selection, thus effectively improving the performance of students' ability level estimation.