The Design And Implementation Of Online Study Behavior Assessment System Based On Data Mining Technology

Lifelong learning is the trend of current education development. Modern distance education is the basis of lifelong learning system. The Distant Education System (DES) supported by computers plays an important role in distance learning. It bulids a platform based on computer networks for teachers and students. Today, the Web-based DES has become the main pattern of distance learning.The assessment of learning is an important portion of teaching and learning. In distant education environment, students are physically separated from instructors by distance or time and arrange their learning by themselves. To assess students' online study is almost unachievable, because most of the Web-based DESes are lack of the monitoring of learning process and the assessing of the online study behavior.Motivated by this problem, this paper proposes a solution to assess students' online study behavior based on Data Mining technology. It implements the Online Study Behavior Assessment System above the distant learning website of Hunan Radio and TV University. The system is able to grade the students according to their online study behaviors. More than 90% of teachers and students agree with the assessment results. This system provides an objective, reasonable method for assessing online study and also directs students to plan their learning effectively. About 99% of users express that the system is helpful to improve online study.This paper firstly introduces the principle of Data Mining technology and the decision tree method. Then it proposes a new C4.5 rule generation algorithm based on attribute correlation which takes full advantage of the correlation between attributes when simplifing the rules. Experimental results show that the new algorithm can speed up the computation efficiently while keeping the original precision of the classification, reduce the size of rule set and decrease the number of rules. Finally, this paper describes the process of constructing "Online Study Behavior - Effect" model using the improved C4.5 algorithm.