Research On Multi-dimensional Student Academic Early Warning System Based On Hadoop Platform

With the continuous construction of the smart campus,academic early warning is one of the branches of the smart campus,and it is also developing rapidly.This article is based on the academic warning.In order to prevent students from being confused and slack in the process of learning,and even failing to graduate,it is necessary to evaluate the students' learning status through different stages of the student's semester.This thesis analyzes students' grades,credit score,attendance,and online logs to achieve early warning of students with poor status.Due to the large amount of students and the variety of data,the traditional relational database can only analyze and process structured data,and when the amount of data is large,the query speed is slow.However,Hadoop technology can solve these problems.This thesis uses the current mainstream Hadoop platform for processing.The system uses the distributed file system HDFS to store student data,uses the distributed column-oriented storage database HBase to process large-scale semi-structured data in real time,and uses the distributed data warehouse Hive to execute large-scale data sets stored in the database.The HiveQL query finally uses the Hue visualization tool to visually display the processed data,and obtains a list of students who need to be alerted through the analysis system.This thesis uses the Hadoop platform to process the students' data,and divides the data according to the college to provide early warning for students from different colleges.Most of the existing early warning systems are based on “post-processing”.Warnings for students who are already in the department or unable to graduate can not really serve as an early warning.This article is changed from “post-processing” to “prewarning”.Four-dimensional early warning analysis is carried out on student's performance data,credit score data,attendance data,and online log data.Analyze the attendance data,Internet time-length data and Internet preference data of each student in the current semester,and alert those students who are absent from class and play online for a long time.By analyzing the student's historical scores and credit score data,early warning is given to students who may not be able to graduate successfully.