Home Research Topics Education and Social Sciences

Learning the Online Learner's Behavior Using Web Mining

 

Dr. Khalid  Matar

Integration of Web-based learning environments into educational processes is rapidly increasing during the last decade. The models of implementation are varied and include, among others: Websites supporting face-to-face courses, on-line asynchronous instruction, and virtual learning communities focusing on communication among participants. While implementing these models, data on the user's behavior is constantly accumulated in the servers' Web logs. One of the major challenges in this field is to develop and apply innovative Web (usage) mining techniques for analyzing this data and for mapping patterns of learning in the Web-based learning environments. To this end, the log files serve as the bridge between the wishes to "see" the online learners and the situation in practice of having no face-to-face contact with them.

Therefore, the purpose of the proposed study is to explore the behavior of the online learners throughout the learning process using Web mining techniques. This methodology is non-intrusive (in the sense that it is not disturbing/biasing the learning process), objective (i.e., data tell us exactly what happened and not what students said that happened), and scalable to large populations.

The study will be held in four main phases: (a) Characterizing the research environments - for this, a few learning environments will be chosen, representing the varied online learning opportunities in Qatar (e.g., Web-supported university-level instruction, supervised online units in mid-school and high-school level, fully online courses for adults in life-long learning). Each environment logs will be examined for understanding their structure and for determining which information might be extractable from them. (b) Constructing conceptual and computational frameworks for the Web mining algorithms - this stage will include the understanding of which learning variables (to describe the online learner) will be researched and how they will be calculated. (c) Analyzing the variables - this stage will consist of a thorough investigation of the learning variables and possible connections between them. (d) Validating the results - applicable for variables describing cognitive and affective situations (e.g., motivation, anxiety, high-order thinking, strategy development).

The proposed study is aimed on improving the online learning/teaching process by understanding the learner's behavior (and changes in it), and this is its main contribution. By investigating individual differences between students regarding their actual online behavior in varied learning scenarios, the instructor - who cannot actually see the students while they learn online - is being able to respond (either online or offline) to each student according to his or her needs. Furthermore, learning environment designers might use the knowledge about the students in order to improve (and maybe to personalize) the system. Therefore, all the three main actors in the Web-based learning field (students, instructors, software designers) are being benefited.