Applying Data Mining as a tool to discover the learning behaviour

In online and distance education environments analyzing or observing the student's learning patterns is a huge challenge. To identify the students with learning problems understanding the student learning behaviors can be used as a open bridge by the tutors to provide adaptive and customized feedbacks, guidelines and personalized attention as needed. Most of the available Learning Management Systems (LMS) like moodle or commercial e-learning products provide very limited capability on tracing the student activities such as visit history, discussion boards etc. But the information which is provided can be used for the analysis process since different learning patterns can be identified through the student learning behavior.
 

Jui-Long Hung and Ke Zhang [2008] tried to research on applying data mining techniques on educational data to discover the learning patterns on the students. In their study they tried to differentiate as many online learning behaviors as possible by applying various data mining approaches on data. For this research they selected a undergraduate course in a University and the course module was access by a LMS known as Wisdom Master. The data were collected through server log files where 98 students were selected as the data source for this study.

In the data preprocessing phase the log file data was cleaned by removing all useless, irregular, and missing data from the original LMS common log files and after the initial preprocessing, a session filter was applied to the reduced log file for feature extractions. The purpose of the filter was to aggregate all user requests within a session into a single set of variables. Feature extractions filtered out the following primary variables: user identifier, session identifier, session start date and time, session end date and time, user’s hit count, and session duration in minutes and based on these derived variables (duration and frequency of data of each student) were extracted through calculating or accumulating primary variable data on a daily and weekly basis.

In this research the data mining phase included two sub-phases which were descriptive analysis, and artificial intelligence analysis. Descriptive analysis was used with summarizing, clustering, and association rules techniques to generate an overview on the dataset, to gain an insight into students’ characteristics, and to depict students’ learning patterns and the Artificial intelligence analysis was used for predictive purposes .

Clustering techniques were applied to classify students based on their shared characteristics. They used K-Means clustering mechanism to identify the student clusters based on the behaviors they showed within the LMS environment and Sequential association rules were applied to discover the daily learning patterns of the students in the LMS. Finally they used decision tree algorithm to build the predictive model on the students. According to the predictive model, the frequency of accessing course materials was the most important variable for performance prediction in this study. Also this study concludes that when students participated more actively that is having a higher value on frequency of accessing course materials, number of messages posted, number of messages read, and frequency of synchronous discussions attended they performed better than the others in academically.

Finally in this research paper the authors suggested that instructors would be able to get a quick view of basic learning data, such as login date, frequency, pages visited, etc. However, no functions or features are currently available to help instructors identify learners’ individual or group learning patterns, or to identify successful or less successful learning behavioral patterns, or to identify the predictive learning behaviors or to help identify necessary facilitation needs. Therefore, the researchers of this study strongly suggest that LMS developers should integrate data mining tools to facilitate effective online teaching and learning.

Reference: J. L. Hung and K. Zhang, “Revealing online learning behaviors and activity patterns and making predictions with data mining techniques in online teaching,” MERLOT Journal of Online Learning and Teaching, 2008.