Association Of Students' Consumption Behavior And Performance Data Based On Campus Card

With the development of high education in China,the number of students in colleges and universities increased every year,and it leads to a dramatically promotion of management pressure.The campus card has become a necessary item for teachers and students,which records the data of students eating,borrowing books,supermarket consumption,access control,and online payment.This problem seriously affects the healthy development of students' academics.This paper uses data mining technology to study the student card data and performance data,analyzes the daily consumption behavior of students,and mines the behaviors and achievements.The results of analysis and mining can help school managers to solve problems in school in time,and provide guidance and assistance as well as support and reference for rational allocation of resources and scientific decision making for students and promote intelligent management of schools.The main work of this paper is as follows:(1)Pre-processing campus card data and course grade data.The data acquisition,data integration,data transformation and other pre-processing techniques are used to process the original card data and performance data,eliminating the problems of missing,error and redundancy in the data,and providing effective input data for later analysis and mining.(2)Using K-means and statistics methods to study the different aspects of students' card using behavior,consumption level,breakfast,library utilization and bathing,and obtained some consumption characteristics and behavior habits of students.The student's daily activities are segmented into time series according to the time period,and the sequence-based PrefixSpan mining algorithm is used to mine and obtain the student's activity track information.(3)Aiming at the problem that the traditional support-confidence framework has defects that result in a large number of users not interested in the rules,this paper combines a correlation-based interest degree with the AprioriTid algorithm for screening rules.Experiments show that the improved AprioriTid algorithm effectively reduce the number of meaningless rules and improve the quality of mining rules.The improved AprioriTid algorithm is applied to the analysis of student course scores,and the relationship between different courses is obtained.Using the improved AprioriTid algorithm to explore the relationship between student consumption level,borrowing books,library learning,bathing,internet fees and other achievements,the main conclusions are as follows: 1)Students who often go to the library to study and spend fewer time on the Internet.Grades are generally medium or excellent;2)students with low Internet consumption and moderate bathing levels are more likely to achieve excellent results;3)Students with high Internet fees and low utilization of libraries generally get poor academic result;4)Students with poor grades have three main characteristics: less library self-study,low bath frequency,and high Internet consumption.