Mohammed Abdel Razek, Claude Frasson, and Marc Kaltenbach
In E-learning systems, where both helpers (tutors) and learners are separated geographically, finding a reliable helper is one of the most important challenges. Although helpers could have a lot of useful information about courses to be taught, many learners fail to understand their presentations. A major part of this paper deals with the following challenges: do helpers have information that the learners need? Will helpers present information so that learners can understand? And can we guarantee that these helpers will collaborate effectively with learners? A new technique is filtering according to helpers’ credibilities. We define ”credibility” as the dependability degree of the learners on the information presented by helpers during a learning session. We propose a guide agent, based on the pyramid model, which can group helpers. This makes it possible to recommend reliable ones. Furthermore, we developed a new statistical metric called Precision Probability Value. We have used this metric to measure statistical accuracies rather than the mean absolute error.