Design And Implementation Of A Test Paper System Based On Question Difficulty Automatic Classification

With the continuous development of computer technology,education and learning methods have gradually become diversified.As a result,e-learning platforms have gradually emerged and are widely used by more and more people.The online examination system has therefore been developed and more and more courses are being tested online.For a set of test papers,the most basic requirement is that the difficulty level of the questions is reasonably allocated and the knowledge is clearly defined.Such papers can meet the basic teaching requirements.Whether it is the traditional manual volume or the intelligent volume that comes with the development of computer technology,the assembly process cannot be separated from the difficulty level of the questions.The difficulty level of the questions in the process of setting up a volume is too high or too low.,will affect the evaluation of student learning.Although some title questions are marked with difficulty,the process is also manually labeled by the issuer.In the process of manual labeling,not only a lot of manpower and time are wasted,the labeling process is interfered by various human factors,and the subjectivity is strong.Once the difficulty level is marked incorrectly,the test papers based on this group cannot be well reflected the students' learning situation and teachers' teaching effects.Therefore,based on the online database of test questions,this study designed an automatic test paper system based on the difficulty level of the questions.The difficulty level of the questions was judged by the computer,and the process of making the test paper was based on this.This study first selected the test question bank of the "Computer Composition Principles" online learning system developed by the embedded system and information security laboratory of Nankai University as the experimental object.It explored the features of the selected types of topics and analyzed the impact of the difficulty of the test questions.Degree factors,including the objective factors and subjective factors of the topic itself,then propose an automatic evaluation model(DAEM).In the design process,natural language processing techniques,word segmentation techniques,machine learning and other methods were applied.The feature set of the knowledge network model based on teaching materials was proposed as an important criterion for the classification of difficulty levels.The DEAM model mainly uses the process of matching attribute index sets and knowledge network models to determine the degree of difficulty.The data set of the test questions is obtained by the model,and the difficulty level of the test questions is finally obtained by the machine learning classification method.Finally,an automatic test-volume system based on the difficulty evaluation model was designed and implemented.That is to say,the input title was used,and the system automatically evaluated the difficulty level of the title.The diffulty level was used as an important index of the test paper.According to "Computer Composition Principles" in the online learning system,the difficulty level is automatically graded.After a scientific experiment evaluation system,the classification accuracy rate of the automatic difficulty level can reach 82.65%.The implementation of the assembly system is easy to use.Evaluation model,the final set of test papers can meet the basic needs.