Dynamic learner profiling and automatic learner classification for adaptive e-learning environment |
| |
| Authors: | KR Premlatha B Dharani TV Geetha |
| |
| Institution: | Department of Computer Science and Engineering, College of Engineering, Anna University, Chennai, India |
| |
| Abstract: | E-learning allows learners individually to learn “anywhere, anytime” and offers immediate access to specific information. However, learners have different behaviors, learning styles, attitudes, and aptitudes, which affect their learning process, and therefore learning environments need to adapt according to these differences, so as to increase the results of the learning process. In addition, providing the same learning content to all the learners may lead to a reduction in the learner's performance. Hence, there is a need to classify the learners based on their performance and knowledge level. Learner profiles play an important role in making the e-learning environment adaptive. Providing an adaptive learning environment, catering to the changing needs and behavior of the learner can be achieved by evolving dynamic learner profiles. Navigation logs can be used to analyze learners’ behavior over a period of time. In this work, we propose dynamic learner profiling to cater to changing learner behaviors, styles, goals, preferences, performances, knowledge level, learner's state, content difficulty, and feedbacks. Based on the continuous observation of learner preferences and requirements, the learner profile is dynamically updated. Furthermore, we propose an automatic learner classification to construct the learner profile and identify the complexity level of learning content, using the Bayesian belief network and decision tree techniques. We evaluated our system with two traditional adaptive e-learning systems, using static profiles and behavioral aspects, through our performance evaluation method of different learner types. In addition, we compared the actual learners’ data with the system generated results for various types of learners, and showed the increased interest in their learning outcomes. |
| |
| Keywords: | adaptive e-learning learner profile learner behavior classification Bayesian belief network decision tree |
|
|
