April 2012 Volume 15 Number 2 - Educational Technology & Society

visit courses with related content, based on student course paths in the LMS. In the future we will perform aggregate
rankings on these groups, which are based on the proposed metrics (Enrichment / Interest / UPCS).
Specifically our methodology has the following advantages:
1. It is independent of the LMS platform since it is based on the Apache log files and not the LMS platform itself.
Thus, it can be easily implemented for every LMS.
2. It uses new metrics in order to facilitate the evaluation of each course in the LMS and allows the instructors to
make proper adjustments to their course e-learning material.
3. It uses clustering techniques in order to identify different groups of courses and different groups of students.
4. The MCL graph clustering algorithm is fast and scalable, and is applied for the first time in the field of elearning.
MCL can stack successfully courses into the same cluster based on the students’ course visits and can
distinguish isolated courses.
5. The BioLayout visualization tool helps depict the clustering results and allows the instructor to further explore
the results in an interactive 3D environment.
We received feedback about the methodology by the educators. The educators were informed about the indexing
results and most of them increased the quality and the quantity of their educational material. They improved the
quality by reorganizing the educational material in a uniform, hierarchical and structured way. They also increased
the quantity by embedding additional educational material. An important outcome through the process of informing
the educators about our results was that the ranking of the courses constitutes an important motivation for the
educators to try to improve their educational material. Because of their mutual competition, they each want their
courses to be highly ranked. However, a few educators complained that the organization of their courses does not
assist them in having high final scores in the ranking list. They argued that, for example, the metric interest is heavily
influenced by the number of web pages used to organize the educational material. Thus, courses that have all their
educational material organised in a few pages have a low interest score. They were asked again to re-organize the
material for each course in the LMS according to the order they are taught, in order to facilitate the use by the
students. In the future we plan to evaluate our methodology by using a pre- and post-test design. A statistical
comparison of the metrics of the courses before and after an instructor's exposure to and utilization of the metrics
would offer valuable insights and help systematically evaluate our methodology.
We also plan to further automate the whole procedure, that is, we are developing a plug-in tool to automate the data
pre-processing and clustering steps. This tool will run periodically (each month) and will e-mail to the instructors
course ranking and suggestions. A similar policy was also applied by Feng et al. (2005), where in the long term,
instructors were informed automatically by email about the quality of the content of their courses. We intend the
final tool to offer insights at 2 levels: (i) On-line, with total statistical information such as number of visits per
course (pages and sessions), student trends and activities at their visits, as well as detailed information per student
(student duration per course and activity, student preferences and activities for all courses), and (ii) Off-line, with the
use of data mining techniques such as pre-process, visualization, clustering, classification, regression and
association, and discovering hidden data patterns.
References
Avouris, N., Komis, V., Fiotakis, G., Margaritis, M., & Voyiatzaki, G. (2005). Logging of fingertip actions is not enough for
analysis of learning activities. Proceedings of Workshop Usage Analysis in learning systems (AIED’05), (pp. 1-8), The
Netherlands: IOS Press.
Baraglia, R., & Palmerini, P. (2002). SUGGEST: A web usage mining system, Proceedings of IEEE International Conference on
Information Technology: Coding and Computing (pp. 282–287). Las Vegas, NV: IEEE Computer Society Press.
Binali, H., Chen, Wu, & Potdar, V. (2009). Emotion detection in E-learning using opinion mining techniques. Proceedings of the
3rd IEEE International Conference on Digital Ecosystems and Technologies (pp. 259-264). Perth, WA, Australia: IEEE.
Bing L. (2007). Web data mining: Exploring hyperlinks, contents and usage data. New York, NY: Springer.
Claroline. (2009). Claroline.NET open source eLearning and eWorking platform. Retrieved April 10, 2009 from
http://www.claroline.net
Dongen, S. (2000). Graph clustering by flow simulation (Unpublished doctoral thesis). Utrecht University, Netherlands.
165