Campus Card Consumption Records Used To Assist Student Management Work

With the rapid development of science and technology,the prosperity of the information age laid the foundation for the arrival of the era of big data.The word ‘era of big data' is increasingly mentioned from 2012.Nobody would doubt the importance of data now,data mining technology has been actively used by all industries to explore the potential value of the data that they hold.The field of education is also actively applying data mining techniques to analyze the various data which is generated during the education and teaching process.So there formed a subject—educational data mining(EDM).EDM is an interdisciplinary discipline.It includes pedagogy,psychology,statistics,and computer science knowledge.Predicting Students' Performance(PSP)is one of the earliest and most popular studies of EDM.In fact,educational data mining is not limited to student performance prediction.It's widely used in many other field.As long as the school provides educational and teaching data,it belongs to its research scope.Though universities and colleges have very rich and well-developed management systems and tools,the depth of their management work(such as award-assistance loans,etc.)is still superficial.It has not reached a certain depth.It only stays at a stage of solving trouble when questions emerge.For example,although colleges and universities have an information management system of student,the teaching managers can't get the student's real-time information.If they want to get those information,they will have to communicate with students online or offline.In addition,if a student has a mental problem,the managers of teaching can only provide some help after the student's problems being revealed.This is a typical way of resolving problems after emerging.Obviously,such management is not efficient enough,so an efficient management method should be to find problems as soon as possible and solve the problems before they become serious.The development of educational data mining offers lay the foundation for more efficient student management.In light of the problems existing in traditional teaching management,educational mining technology can provide a good solution.For example,the research in this paper is to assist student management work through analyzing the student's campus card consumption records with educational data mining techniques.From the research results,education data mining does provide an assistance for student management.In the study of this paper,we preprocess the campus card consumption records of students,and analyze the consumption situation in canteens,supermarkets,and networks.Through the establishment of multiple models for students' canteen consumption,the students' financial situation was forecasted.The best prediction result of the model reached 76.4% with a precision of 91.1%.In addition,we also conducted a comparative analysis of the student's campus card consumption records and historical grade records.We first preprocess scores and consumption data,and then put forward a reasonable conjecture that the habits have a greater impact on the grade based on the counselor's feedback and related research results.According to these conjectures from the campus card consumption records,we calculate the corresponding characteristics.Features are selected through principal component analysis and Boruta feature selection algorithm,and then the predictive effects of these features are tested through scholarship prediction.After the selected features were determined,the selected features were used for student's academic risk prediction.The experimental results show that the Recall of predicting academic risk was up to 79.38% on Naive Bayesian model,through using the six selected features.Comparing the above characteristics with students' academic risks,we find that the correlation coefficient of these characteristics of academic risk is relatively high,which also shows that the selected characteristics do have a greater correlation with student performance to some extent.It also shows that the campus card consumption records have certain research value for assessing students' academic risk.