Abstract:

Data mining technology is concerned with the discovery of hidden and unexpected patterns from large databases. Recently, Educational Data Mining has become an emerging research field used to extract knowledge and discover patterns from E-learning systems. The educational system in Albania is currently facing a number of issues such as identifying students’ needs, personalization of training and predicting the quality of student interactions. Educational Data Mining provides a set of techniques, which can help the educational system to overcome these issues. The objective of this work is to describe a Educational Data Mining analysis, by describing a step-by-step process using a variety of techniques such as Attribute Weighting (Weighting by Information Gain, Relief, Hi-Squared, Uncertainty), Clustering (K-Means), Classification(Tree Induction), Association Mining (Apriori, FPGrowth, Create Association Rule, GSP) in order to achieve the goal to discover useful knowledge from the Moodle LMS. Analyzing mining results enables educational institutions to better allocate resources and organize the learning process in order to improve the learning experience of students as well as increase their profits. In this study, some EDM techniques were applied in the "Informatics" course taken by 27 Physical Education students at Sports University of Tirana, during Spring Semester of 2009-2010. Blended Learning was applied for this course. Rapid Miner (v5.0) and Weka (v3.6.2) data mining tools were used to mine data from the Moodle system. Data for the initial set of experiments in this study was derived from the Moodle logs. These data were analyzed from various levels and perspectives, in order to provide more insight in the overall educational system. From the level of an individual course user activity is considered, accomplishing assignments, quizzes project assignments, participation in forum and chat. By applying clustering methods, the goal was to split data set in groups of data points that naturally group together. By using prediction techniques, the goal was to develop a model that can infer predicted variables from predictors’ variables. By applying Association Mining techniques, the goal was to discover relationships between variables. The mining results can be used by the course instructor to investigate the impact of a number of e-learning activities on the students’ learning development. This paper concentrated on the overall LMS performance at Sports University of Tirana and the mining process of Moodle data. Mining the Moodle data allowed identifying the most effective ways to the teaching process that can be used to enhance the education process. The experimental results have shown that the data mining model presented in this paper was able to obtain comprehensible and logical feedback from the LMS data describing students’ learning behavior patterns. To further test the effectiveness of the proposed model and to increase the generality of this research, more extensive experiments should be conducted by using larger LMS data sets.

Keywords:

Data mining, e-learning, LMS.