National Matriculation Grade Analysis Based On OLAP And Data Mining Technologies

In China,national matriculation test(NMT)can be regarded as the largest and most influential examination that can select outstanding talents.As a test method,NMT can not only select professional talents but also measure the quality of modern basic course teaching.Thus,the division of grade and test paper of NMT are two controversial issues.Since Jiangsu Province is one of the most famous education provinces,the NMT of Jiangsu has aroused a great concern.In this thesis,we utilize the grade of students in Jiangsu as the experimental data.Spcecifically,the sum of Chinese score,Mathematics score,and English score,are included.The levels of batch and elective courses are also taken in consideration.Technically,we build a data warehouse based on the known information about students,create visual view through OLAP tools,andconstruct a mining model.At last,the final scores of NMT can be predicted according to the scores of simulation test.The main contributions of this paper are summarized as follows.(1)Deal with the scores of students.We first unify the format of these scores,and then import the scores into a database to creat a complete data warehouse.The scores of NMT,levels of batch and scores of elective courses are included.We apply the association rule to mine the relationship between simulation test and NMT.At first,we creat a data sourse view.Then we build the association rule to find all the links that satisfy the demand of confidence level and support level from the set of simulation test and NMT.As a result,the latent correlation between simulation test and NMT test can be found.(2)Use decision tree classification algorithms to predict the categories of admitted level.Specifically,the scores of NMT,levels of batch and elective courses are selected to creat a decision tree model.At last,we can predict the NMT scores according to the scores of simulation test.We also apply association rules to the decision tree classification algorithm.Then we can obtain the classification matrix of decision tree and compare the one without applying association rules.