Analysis And Association Rule Mining Based On Campus One Card Data

With the improvement of the informatization level in college management and student service,the campus card has become a widely used infrastructure.It records the behavior data of students such as consumption,library visiting and book borrowing.However,many universities only conduct some simple statistical analysis,while neglecting the valuable information implied in these accumulated data.To this end,this paper mainly takes campus card data as the research content,then conducts in-depth analysis and mining on the behavior data of the student from two perspectives of group and individual.This paper aims at finding out association rules between the students’ behaviors and making personalized book recommendations based on the book borrowing data,thus improves refinement and personalization in the students management.When analyzing the students’ behavior as a group,this paper applies the FP-Growth algorithm in the filed of data mining to conduct association analysis on the data of student scores,book borrowing,library visiting and dining hall consumption,then mines the association rules among students’ behaviors and establish user portraits of the student group.The results show that there is different correlation between students’ academic performance and the factors mentioned below,and a good academic performance is related to a good reading habit.When analyzing the students’ behavior as a group individually,this paper achieves a personalized book recommendation by using book borrowing data based on the association rule study on the students’ behavior data.The model introduces time factor to improve the recommendation algorithm,adjusts the score data by the linear combination of the student interaction data and the borrowing time data,then uses Alternating Least Squares(ALS)algorithm to make book recommendation.Experimental results show that the mean square error of the original model is 0.66,and the improved model is 0.49,which can make the recommendation more accurate.The results of this study can provide a theoretical and experimental basis for school policy making and student management improvement.