Analysis And Statistics Of Examination System Based On Data Mining Technology

Examination is an indispensable part of education,Examination Analysis is also an important topic of research of Education Testing.Examination Analysis can be divided into three parts: Enrollment Analysis,Achievement Analysis and Examination Quality Analysis.Those three parts have their theories and methods respectively.Mathematical statistics method have been used for Examination Analysis with statistical models and indexes.Classical Test Theory(CTT theory)makes pertinent measurement of the examination by seting up a series of concepts and principles such as reliability,validity,project analysis,norm,standardization and so on.Item Response Theory(IRT theory)measure examinee's ability and quality by using probability model(such as Rasch model)to construct nonlinear function to represent examinee's response to test paper.The Data Mining technology can excavate the information in the examination data deeply.C4.5,Aprior and K-means are classic algorithms in the field of data mining.The C4.5 algorithm generate decision tree of test data,.The Apriori algorithm mining association rules between achievement and various factors.The K-means algorithm introduces the concept of distance by measuring the similarity of data for facilitating quantitative research.The paper tries to utilize the three algorithms for empirical analysis in the examination analysis system,and compares it with Educational Statistics and Educational Surveying,tying to find out the problem in the application,achievement and quality of the examination paper and the Complementation of data mining technology and the traditional examination analysis system.I will continue trying to combine the data mining technology with the traditional test analysis system for meeting the requirements of the modern paperless examination and the network examination for the examination analysis,and promoting the reform of the examination analysis in the big-data background.