Font Size: a A A

Modelisation des reactions emotionnelles dans un systeme tutoriel intelligent

Posted on:2010-12-16Degree:Ph.DType:Thesis
University:Universite de Montreal (Canada)Candidate:Chaffar, SoumayaFull Text:PDF
GTID:2448390002479254Subject:Artificial Intelligence
Abstract/Summary:
The emotional factor has been never taken into account in Intelligent Tutoring Systems (ITS) until recently. However, emotions playa crucial role in cognitive processes particularly in learning tasks (Isen, 2000). Indeed, modelling the learner's emotional reaction is fundamental for ITSs in order to help the tutor deciding, when and how to intervene, for helping the learner to achieve learning in the best conditions. In this thesis, our first aim was to develop a method for predicting the learner's emotional reaction at a given time of the learning process. Our second aim was to study the effect of some tutoring actions on the learner's emotional state.;To achieve the second aim, we developed a data structure web course and a virtual tutor using different pedagogical actions to induce positive emotions in the learner. We have conducted an experiment to collect, among other, participants' physiological responses after the tutoring actions. The results of this experimental study showed that certain actions have significant positive effects on the learner's emotional state.;We conclude in this thesis that, it is possible for an ITS to predict the learner's emotional reaction and to induce positive emotions.;Keywords: ITS, learner's emotional reaction, tutoring emotional actions, supervised machine learning, physiological responses;To achieve the first aim, our approach of prediction relays on the causal events which could trigger this emotion and on its determining factors like the personality for example. Thus, we propose to solve this problem by using supervised machine learning algorithms and more precisely those of classification. The J48 algorithm can predict the learner's emotional reaction (positive or negative) after receiving his mark in online evaluation test's, with an accuracy rate of 81.54% by a leave-one-out validation. The BayesNet algorithm can predict the learner's emotional reaction (positive or negative) after the tutor's feedback with an accuracy rate of 70.69% by the same method of validation.
Keywords/Search Tags:Reaction, ITS, Actions, Positive, Tutoring
Related items