Affective reasoning based on bi-modal interaction and user stereotypes

Abstract

This paper describes a novel research approach for affective reasoning that aims at recognizing user emotions within an educational application. The novel approach is based on information about users that arises from two modalities (keyboard and microphone) and is processed based on a combination of the user stereotypes theory and a decision making theory. The resulting system is called Educational Affective Tutor (EAT). EAT is an educational system that helps students learn geography and supports bi-modal interaction. The main focus of this paper is to show how affect recognition is designed based on and empirical study aimed at finding common user reactions that expressed user feelings while they interacted with computers. © 2008 Springer-Verlag Berlin Heidelberg.