Automatic Detection of Tutoring Styles Based on Tutors' Behavior

In e-learning systems, tutors have a significant impact on learners' life to increase their knowledge level and to make the learning process more effective. They are characterized by different features. Therefore, identifying tutoring styles is a critical step in understanding the preference of tutors on how to organize and help the learners. In this context, the authors address the problem of extracting tutoring styles from tutors' behavior. According to this later, tutors are classified automatically into their styles. This technique will be helpful to provide a suitable advice to learners. In the first step, a set of indicators are defined to characterize a tutoring style. In the second one, the accuracy between the tutoring styles obtained from the proposed approach and those defined from a simple questionnaire is investigated. To validate this approach, the authors have collected data from an on line tutoring system LETline, http://www.labstic.com/letline. They present the results of their analysis and discuss some limitations that can be helpful to the researchers working in the tutoring field.