Analysis Of Academic Achievement And Research On Influencing Factors Based On Clustering

With the transformation of education from elite to popular,the enrollment of college students is increasing year by year,and the quality of education has gradually become the focus of colleges and universities.As an important index to measure the quality of education,student academic achievement not only represents students' learning achievements,but also reflects teachers' teaching achievements,school management achievements and the rationality of curriculum arrangement.Therefore,it is of great significance for students to improve their academic achievement and implement individualized management in schools through in-depth analysis of academic achievement and its influencing factors,meanwhile providing decision reference for schools to complete educational reform.My study mainly analyzes the students' academic achievements and its influencing factors from the following aspects.Firstly,the improved k-means algorithm improves the accuracy and stability of clustering results.The main improvements include: my study uses entropy method calculate attribute index weight to improve the accuracy of clustering results.I select centers according to the density of sample points and neighborhood radius to avoid the interference of outliers or noise points to clustering results,which overcomes the shortcoming of randomly selecting initial centers of the traditional k-means.A new clustering criterion function is formed by using the average intra-class distance and inter-class distance,and the optimal clustering number is determined according to the minimum of the criterion function,based on the above ideas,an improved k-means algorithm is proposed.The accuracy and stability of the improved algorithm are verified by the Iris data set,Wine data set and Glass data set in UCI database.Secondly,the improved k-means algorithm is applied to analyze academic achievement.The academic achievement of Logistics Majors in a university is taken as the research object.The improved k-means algorithm is used to cluster the students into six categories.The characteristics of each category of academic achievement are evaluated and analyzed to find out the weak subjects and provide reference for the implementation of classified management.Thirdly,the influence of different factors on academic achievement is simulated by using system dynamics.Constructing the index system of influencing factors of academic achievement,using Vensim PLE software to simulate the influencing factors of different types of students on the basis of clustering results,the result shows that students' individual and school factors have a greater impact on College Students' academic achievement.Next,a simulation model of student and school factors subsystem is established to calculate the actual effect rate of each factor on the subsystem.The result shows that learning initiative,professional interest,teachers' teaching attitude,class learning atmosphere and other factors are the key factors affecting each subsystem.Through the idea that combines the clustering analysis with system dynamics simulation,my study makes an in-depth analysis of academic achievement and puts forward research countermeasures to improve academic achievement from three aspects of students' individual,school and family,which is of great significance to improve teaching quality and school development in the future,It also broadens the research field of system dynamics and provides new research ideas for the analysis of academic achievement.