Statistical Analysis Of University Students’ Grades In School Of Mathematics

In an era that the whole people pay attention to the university entranceexamination, it is a focus problem that university entrance examination is astepping-stone leading to Higher Education or has a far-reaching influence onstudents’ academic records at university. So the education evaluation work isproduced along with the development of higher education. Many educators carry out avariety of quantitative and qualitative education evaluation work revolved aroundstudents’ university entrance exam scores, school performance, comprehensive quality,personal development after graduation and so on, systematically expound theproblems that university entrance exam scores’ influence on the subsequentdevelopment of students entering the higher education.Since university entrance exam scores is an important indicator to measurestudents’ learning in high school. According to students’ university entrance examscores and their undergraduate level of06class graduates in Hefei University ofTechnology, this paper draws a conclusion that there is a correlation between students’university entrance exam scores and their each semester grades, but this correlationweakens gradually as time goes on. At the same time, the correlation is different dueto different majors. And further study finds that there is a different correlationbetween students’ university entrance exam scores and their each semester grades dueto different type courses. This paper establishes a linear regression model andanalyzes based on some representative curriculums, university entrance exam scoresand courses results.In addition, this paper studies the mean change point estimation problem andgives the form estimates of change point cumulative sum. And then we buildrespectively a state space model about the overall SSE Composite Index data and thedata after change point with Eviews software. Through comparing and predictingabsolute percentage error (APE) of the results, a model based on segment data ofvariable structure is established, and it has a better prediction.