Stage Predicting Student Stay Time Length on Webpages of Online Courses Based on Grey Models

Abstract

To provide adaptive learning, an e-learning system needs to gather information about what student state is while the student learns online course. A student state index is the length of time the student stays on a webpage of online course. By predicting student’s stay time length, the e-learning system has potential to dynamically tailor the learning content to the students. A literature review is conducted on power law of learning and knowledge component. We assume that online course consists of knowledge components. A knowledge component crosses several successive web pages. Accordingly, an initial prediction method is proposed based on the two learning curve modes and the grey models. Based on the experimental result of this initial prediction method, construction method of grey models is modified. The results produced by the grey models based on the two construction methods are then compared and analyzed. The results show that prediction of stay time length is possible to certain degree while the students learn knowledge on web pages.