Comprehensive Analysis Of College Student Behavior In Big Data Environment

With the development of informatization construction,most universities in China have begun to build digitized campuses,having brought great convenience to all teachers and students and school administrators.School administrators can use the digital campus office system to deal with affairs,improving the efficiency of handling affairs;students can use the digital campus system to handle admission,registration,and school leaving to facilitate students’ daily life.With the construction of digital campus,colleges and universities will generate a considerable number of data every day.The data generated by students during school period account for a large proportion,including the data of score,borrowing books,dormitory access card,campus network recharge and campus card consumption that includes canteen consumption and campus supermarket consumption records.Record and campus card consumption data,where the consumption data includes canteen consumption and campus supermarket consumption records.After analyzing the behavior data generated by students,the remaining study finds some behavior characteristics during student school period and uses these characteristics to achieve more scientific management of students.The research in this paper is based on students’ data from Lanzhou University of Technology.These data contains the records of book borrowing,campus card consumption,score.Firstly,the data preprocessing operation is performed on the original data,so that the processed data can meet the basic demands of data analysis,and then the data discretization technology is used to discretize the preprocessed data so that the data can meet the demands of the association rule Apriori algorithm.Finally,the Apriori algorithm is used to mine the student’s academic achievement,which finds the correlation between the number of book borrowing and the campus card consumption by the relationship between the students’ behaviors is found.Using strong correlation to predict and warn students’ behaviors provides scientific decisions for student managers.Secondly,using the K-means clustering algorithm based on partitioning,the students’ behaviors are clustered and analyzed.The relationship between student achievement and borrowing quantity,grades and campus card spending quota,borrowing quantity and campus card spending quota is used.In the process of means clustering algorithm analysis,there are obvious differences between students and excellent grades and those with unsuccessful grades.In order to further discover the difference between the two kinds of data,density-based DBSCAN clustering algorithm is used to achieve excellent results.Students with unsuccessful grades are analyzed by cluster analysis,and the behavior characteristics of students with excellent grades and those with unqualified grades are found.The clustering results show that campus card spending quota of students with excellent grades are more stable.They often eatin the school canteen and borrow more books.The students with unsuccessful grades are not stable.They often do not eat in the school canteen and the number of borrowings is low.Finally,using the results of the clustering,students who have achieved excellent results and low consumption quotas will be able to successfully complete the bursary after they understand the actual situation,and warn the students who have the characteristics of unqualified students.The study provides scientific decision-making basis and provides good guidance for the healthy development of the students.