Application And Research On Text Clustering Of Students' Behavior Data In Post Bar Of The Colleges And Universities

Since the advent of the Internet,people's way of life and work has undergone tremendous changes.Under the new environment "Internet plus" mode,the Internet has penetrated into various fields.The continuous development of Internet applications,has accumulated a large amount of network behavior data,resulting in "big data".With the "Digital Campus" to "smart campus" concept being put forward,almost all domestic universities have gradually launched its own campus website,WeChat public and intelligent office system etc the education and teaching service system.Universities have become a big data application position.The internet application in the college students also occupy the first place,inside and outside the school of the each application platform also recorded a large number of student behavior data.Students' behavior has an effect on their academic achievement and learning state.Effective guidance of students' behavior is the necessary condition to promote the all-round development of students.Therefore,college education managers often analyze the behavior of students in order to find problems and correct them in time.The traditional analysis of students' behavior is mainly based on the collection of student behavior data and the use of mathematical statistics.In the "Internet + education" under the new situation,student behavior data is already not limited to the campus,has long been through the network around the world.Students in the campus inside and outside the platform application will leave a record,student behavior has been very direct performance in the campus inside and outside the platform.Therefore,the education manager can directly through the platform to collect student behavior data for analysis,and then make management decisions.Due to the diversity of student behavior data recorded in each application platform,it is more in the form of unstructured data such as pictures,text,audio and video.The education managers of the analysis work has brought challenges and opportunities,modern education,teaching management development of a new trend is to consider a wide range of evaluation indicators to measure the level of education,teaching management.For student behavior data outside applications platform of campus,this topic chooses a more representative platform in the online community----the baidu post bar,the world's largest Chinese community.Baidu Post Bar operating for many years,accumulated a considerable amount big data.In the college paste bar category students are often on the same topic to express their views,the information on the school management decision-making very reference.Posted Bar on the students are used to elaborate with the text data,the resulting data are semi-structured.This article for Baidu Post Bar data of the real,objective,comprehensive and other characteristics,Using Weka open source tools,The K-Means and EM method of data mining in large data is applied in text clustering,Successfully the student behavior done data text clustering mining collected on the website of Baidu Post Bar,In order to assist the educational managers of student behavior analysisThe vision of this paper is to hope for the research of this subject,which can provide a new way for the educational managers to analyze the students' behavior analysis.