Design And Implement Of An Educational Resource Oriented Mining System Based On User Behavior Feature

The appearance of open education platform coupled with the way to access education resources and changing of teaching mode, makes online education become the important forms of modern education. More and more resources appear on the network which have brought the problem of information overload and the trouble to users who retrieved the resources. Referring the appropriate education resources to users and providing personalized service have become the developing direction of the online education. The construction of educational resources data mining system has a very important role for individual learning and makes learning more efficient and convenient.Based on the above background, we research the network users’ behavior features and the related technologies of data mining. We design and implement an education resource oriented data mining system, focusing on providing users with personalized education resources. Our system includes four function modules:(1) Resource management function module is mainly used as storage, retrieval, browse, download and other processing in large-scale education resources in system.(2) User management function module is mainly used as users’ role permissions and information management.(3) System news management module is used to show the organizations which have a wealth of educational resources.(4) The data mining module provides personalized services to users based on the analysis of the characteristics of users’ behavior and on users’ preference model. This module is the core of the thesis. In this part, four key technologies are introduced. The first data mining technology is based on topic extraction with gray prediction, the second one is based on Web log, the third one is based on association rule and the last one is based on item’s features using collaborative filtering. Our system also adopts the popular Web development framework SSH, that is, struts2, spring and hibernate, which clearly displays the results of data mining.At the last of the thesis, we design a number of test cases to evaluate the system. The results shows that our system reaches a desirable performance and meets all the goals we analysis before.