| The development of education informatization has generated large volumes of data concerning the teaching management and students' campus activities in colleges and universities.A large but scattered and complex data set is available now and at the same time the number of college students is constantly increasing,which brings unprecedented challenges to the educational management of schools.The lack of scientific data analysis for this growing group can result in overlook of some potential problems,such as mental health,consumer lending,games addition,etc.,which may lead to major issues.A consensus of large data application is to extract valuable information from it and the user portrait -- defined as the labeling of user attributes,behaviors and characteristics,is the core foundation of mass data application in colleges and universities.It provides the precondition of fine management and makes up an important aspect of smart campus construction.Compared with the previous data analysis based on small samples,using large data can obtain information of college students from more dimensions,which is closer to the full sample and can thus more accurately outline the appearance of user data.By collecting the data of college students' basic attributes,academic attributes,consumption habits,book borrowing,online behavior and other dimensions,and by going on a series of preprocessing to select the feature of user portrait,including data cleaning,integration,transformation and specification,this paper uses data mining technology to extract the hidden information,thus constructing objective,rich and detailed student user portrait.The establishment of college student user portrait system is not only conducive in comprehensively display of the digital appearance of college students,but also can assist the scientific management and decision-making,dynamically monitor student anomalies for timely warning,and provide personalized customized services for students. |
PDF Full Text Request