College Students' Behavioral Correlation Analysis Based On Campus Multi-source Fusion Data

In recent years,universities at home and abroad have generally carried out the construction of "smart campuses" based on big data and the Internet of Things,and use a variety of application service systems as carriers to realize an integrated environment of work,study,and life information on campus.The campus infrastructure represented by the campus card system and campus network not only provides convenient services for teachers and studentson campus,but also records the massive behavior data generated by students in their life and learning activities.A large amount of data has been in a state of accumulation and waste since it was generated.How to apply data mining technology to maximize the potential value of student behavior data,accurately carry out student management,and improve the quality of education and teaching is an important issue faced by the innovation work of relevant management departments in universities.Based on the topic of "Correlation Analysis of College Students' Behavior Based on Multi-Source Fusion Data on Campus",we mainly completed the following research work in response to the needs of data analysis in the field of college student management:(1)The investigation and analysis of the status quo of data analysis technology of college students' campus behavior.A total of 380,045 student data for the 2013 and 2016 grades of a major were selected from the card data,student performance data,and book borrowing data of a college for cleaning,integration,and conversion.The work of data preprocessing is completed.(2)A student behavior classification model based on cluster analysis is constructed and the indicators of meal consumption level,regularity and diligence are proposed.After determining the optimal number of classes based on the sum of squared errors within the clusters,the k-means clustering algorithm is used to divide the students under each indicator.The clustering results can help the students management workers to further grasp the students' situation,and provide decision support for improving the efficiency of education management and guiding the students.(3)Behavior indicators and exam grades data,learning preparation and result data in campus behavior data are integrated.Based on the Apriori algorithm,association rules between behavior and exam grades data,learning preparation and results are obtained.The results show that students who regularly eat breakfast,regular living,bathing and fetching water frequently get better grades.The students who study on the online platform have a high school degree or less before the course,and the result is a high probability of failing or dropping out.Through the correlation analysis of online learning interactive data by Eclat algorithm,the results show that the learning mode of students on this platform is relatively single and the utilization rate of some functions is low.Student management workers can use this to accurately grasp the students' situation,guide the students to establish efficient and healthy study and living habits,and improve the work and learning efficiency.The relevant parties of online learning platform can optimize the system platform,carry out personalized service and design teaching content based on the obtained rules.