Researches On Behavior Analysis System Based On Big Data Technology

The emergence of the campus network has promoted the overall process of university construction,making the management of colleges and universities more convenient and intelligent.Colleges and universities do not need to record information on paper.Instead,they record the information in a computer and store it.As time goes by,more and more data are stored in various departments of colleges and universities,and these data can reflect students'living habits and learning status.Therefore,the analysis of campus network data is impotent to observe students' learning status and the construction of colleges and universities.However,the lack of interoperability between departments and the large amount of data accumulated by various departments has also become the focus and difficulty in the analysis of campus network data.To solve the two problems of excessive data volume and data intercommunication between departments,this topic uses big data technology to analyze the behavior of users.Based on the daily information of students in the network center and the academic affairs office,the mainstream big data processing platform Spark is used to analyze students'behaviors in the daily surfing and the students' learning in peacetime.The teacher has a more comprehensive understanding of the student's learning situation,and also provides a decision-making basis for the analysis of the daily behavior of the student.On the Spark platform,the K-Means++ clustering algorithm is used to classify users,and students who seriously rely on campus networks are selected for in-depth mining,and students' online behaviors are analyzed.Based on the students' specific conditions,reasonable thresholds are put forward to judge whether users are overly dependent on the Internet.we combine the test scores with academic achievements of students in the Academic Affairs Office to prevent students from relying too much on the Internet;By analyzing students' use of the Internet,a reasonable campus network tariff strategy is proposed to assist campuses works.The final experiment proves that this system can analyze students' behavioral data from multiple dimensions and help teachers to understand the students' learning situation;and the student's behavior data can also assist the management of the campus network.