Research On College Students' Academic Achievement Based On Trajectory Data Mining

Academic achievement is the ultimate result of college students' learning activities and is a concentrated expression of college students,quality.Studying the academic achievement of college students is an inevitable requirement for the development of "knowledge economy" and"talent strategy" in China.As a major measure of academic achievement,scholarships are an important incentive system for mobilizing students' enthusiasm.At present,the main methods of college scholarship selection are still highly dependent on manual review and processing.The personal emotional factors of scholarship reviewers may have strong interference.The automation of the process is an inevitable requirement for the improvement of students'academic achievement.School precaution is another system to improve the quality of students'learning.However,the current school precaution system still has a single content and relies on manual processing.How to enhance the automatic processing of the academic early warning system is also the focus of improving students' enthusiasm and learning quality.The development of educational data mining provides the basis for school audit automation.The way to study academic achievement varies according to research questions and research interests.And the main branch is the empirical research based on multiple regression,structural equation modeling and multi-level analysis.With the development of educational informationization and e-learning in recent years,data mining is also increasingly used in education.The term educational data mining has become an exploration paradigm for educational data.This paper uses Wosk-means algorithm and trajectory data mining,taking college students as the research object,and proposes TOCSS method to explore college students' academic achievement.Firstly,the feature weights are generated from the automatically generated objective weights and the subjective weights.The feature weights are introduced into the k-means algorithm to obtain different student groups.Then,the student trajectory is modeled,and the frequent trajectory pattern is found by finding matching patterns at different levels.Based on the frequency of different places,the similarity between the target students and the cluster can be calculated according to this.The degree of degree reflects the degree of deviation between the target student and the traj ectory of the cluster.Calculating the distance between the target student and other clusters,the moving direction can be observed.Combining the typicaltrajectory patterns and clustering results of students,it can effectively provide the basis and guidance for the evaluation of scholarships,and the calculation of trajectory patterns can automatically analyze and identify students with academic risks.Therefore,this paper reduces the interference caused by the individual's emotional factors in the manual evaluation,which is conducive to the promotion of scholarship assessment and automatic processing of school precaution,and enhance campus fairness.