Research On The Prediction Method Of Test Paper Difficulty Based On Association Analysis And Neural Network

With the continuous development of modern information technology,educational informatization has become a necessary measure in the educational management of colleges and universities.The popularization of computers and the improvement of storage performance have accumulated a large amount of examination information data in the field of education,most of the potential information of the data has not been mined,resulting in a waste of data resources.As the main evaluation index to judge students' academic level,the quality of test papers in curriculum examination is especially important,and the difficulty is the key to measuring the quality of the examination paper,if it can provide a method to predict the difficulty of test paper,it can provide a wealth of prior feedback and suggestions to the teaching provider.The main research contents of this paper are as follows:(1)Research on Knowledge point association mining method based on association analysis.Aiming at the phenomenon that the potential information in a large amount of data has not been mined,this study constructs a knowledge point association mining method based on association analysis.First,the algorithm suitable for knowledge point association mining is selected by comparison;then the steps of the knowledge point correlation mining method in this study are discussed in detail,the original data is discretized and the knowledge points are associated analyzed from two aspects,including the classified mining of knowledge points samples with high scoring rate and low scoring rate,and hierarchical mining based on the proportion of knowledge point samples in the total amount of data;finally,using the mining method constructed in this paper,the association between knowledge points in the database course is mined,and the association rule table among the knowledge points with high reliability is obtained to measure the strength of the relevance of knowledge points.In addition,the results can provide data input support for the method of predicting the difficulty of test papers in this study.(2)Research on the prediction method and optimization of test paper difficulty based on neural network.Because the influencing factors of the difficulty of the test paper are more complicated and difficult to extract,it is difficult to predict the difficulty of the test paper directly,therefore,this paper proposes a method of obtaining the difficulty of the test paper prediction after obtaining the difficulty of prediction of each test question in the test paper and then constructs a method based on neural network to predict the difficulty of the test paper,and optimizes the method based on the association of knowledge points.Firstly,this study predicts the difficulty of the test by constructing the RBF neural network prediction model,and uses the nearest neighbor clustering algorithm to optimize the center selection of the RBF model;then,based on the traditional test paper difficulty calculation method of the weighted cumulative test difficulty,combined with the association of each knowledge point in the test paper,the results of the optimized test paper prediction difficulty is obtained;finally,the experiment shows that 1)the accuracy of the optimized RBF Prediction model reached 87%,and it is superior to the RBF neural network,BP Neural Network and LM Neural network in terms of precision,recall and other indexes;2)Compared with the traditional method of calculating the difficulty of the test paper,the average absolute error of the test prediction difficulty and the real difficulty is reduced by 13%,which indicates the test paper difficulty prediction method combined with the difficulty of test questions and the knowledge points association is more accurate.(3)Design and implementation of teaching assistant system.The system uses Spring Boot combined with MyBatis as the development framework,uses Echarts graphics library to realize front-end interface visualization,combines POI to realize data set table uploading,applies the mining knowledge point relevance and test paper difficulty prediction method to the system,and finally displays the mining analysis and prediction results in the form of a report form to provide suggestions for teachers' teaching.