Design And Realization Of Analysis System Of Studenets’ Scores Based On Data Mining Technology

Big data become the focus of the network information industry vocabulary, theapplication of data mining technology in the field of education for the education revolutionbecome possible. Data mining technology in primary and middle school students and collegestudents’ learning behaviors of all kinds of performance, learning achievement and careerplans after graduation can provide valuable information, such as to improve the students’learning methods, to help students find in some of the homework or exams often ignored butvery important information, and it provides personalized service for students in specificsubjects, timely finds students potential risk of dropping out of school.Education data mining is an emerging field, EDM is prepared to meet the growingdemand for education universal evaluation. EDM is focused on collecting, archiving andanalyzing the data of student studies. There is much more obvious in the data in the school,such as the student’s enrollment, registration rate, class attendance rate, dropout rates,scholarship distribution ratio, etc., of course the student score data results in all the subjects isthe most important. There is a similar data specific to the classroom, such as the studentsanswer the questions of accuracy, submit homework correctly, a number of and interactionbetween teachers and students class time, the average time of answer questions. These datacan become the basis for students with all-round performance prediction after process throughprofessional collection, pretreatment, statistics, classification and analysis according to thedata mining,.Performance analysis system mainly uses the data mining algorithm design model toanalyze the education data. There are three goals:First, it predicts whether the new entrance students pass through the first semester exam.If it predicts a student tend to fail in the exam, and then you can suggest the students beforethe exam to take extra effort, to improve their performance and help him to pass the exam.Classification methods including similar decision tree, Bayesian network, this paper USES aCART and C4.5, ID3decision tree algorithm.Second, it selects students who have rebuilt risk with data mining technology. Collegestudents will rebuild the remedial class and test again when the examination score is lower than it needs to be rebuilt scores at the beginning of next term. This not only increases theteacher’s teaching burden, also gave the students bring unnecessary burden, with the help ofthe data mining technology; we can more accurately choose targeted students. In this article,will use a data mining based method to choose the class of students. Method is the keytechnology of association rules based on the score, the effect of this method is very good, isbetter than traditional methods.Third, it finds the relationship between student performance analysis of the students, therelationship between curriculum and influence. In education, students’ scores of quantitativeevaluation is a very important indicator, can objectively reflect the influence of the education,is an important scientific decision-making basis. Therefore, analysis and research on thestudents’ score is very important. The traditional query and simple statistics analysis can’tmeet the needs of analysis; can not capture useful information for the teaching. This paper, byusing clustering algorithm and decision tree mining students’ scores, through the analysis ofthe relationship between the students course-can get some teaching and management of thevaluable information.In this paper, the system test use SAS Enterprise Miner platform, it is designed forbeginners and experienced users, its GUI interface is driven data flow, and it is easy tounderstand and use. It allows an analysis, by constructing a use link nodes and deal with thedata of visual data flow graph to build a model. In addition, the interface allows the nodesdirectly inserted into the data stream processing.Finally, the paper use SAS Enterprise Miner platform testing data mining model isdesigned in this paper, including the input data source, data partition point inspection, variablegrouping missing value alternative, interaction, and the decision tree model, the clusteranalysis model, such as test results show that performance analysis functions can be achieved.