The Multi-Agents Architecture for Emotion Recognition: Case of Information Retrieval System

The Multi-Agents Architecture for Emotion Recognition: Case of Information Retrieval System

Mohamed Néji (University of Sfax, National Engineering School of Sfax (ENIS), Sfax, Tunisia), Ali Wali (University of Sfax, National Engineering School of Sfax (ENIS), Sfax, Tunisia) and Adel M. Alimi (University of Sfax, National Engineering School of Sfax (ENIS), Sfax, Tunisia)
Copyright: © 2014 |Pages: 13
DOI: 10.4018/ijsi.2014010106


The author's research focuses on the problem of Information Retrieval System (IRS) that integrates the human emotion recognition. This system must be able to recognize the degree of satisfaction of the user for the result found through its facial expression, its physiological state, its gestures and its voice. This paper is an algorithm for recognizing the emotional state of a user during a search session in order to issue the relevant documents that the user needs. The authors also present the architecture agent of the envisaged system and the organizational model.
Article Preview


The goal of a personalized Information Research System (IRS) is to supply relevant information according to the user characteristics. Indeed, personalize an information to a user in particular, require the arrangement of certain information on this last one of whom the purpose to help the research system to put back to hand him the information for which he looks. This information is often structured in the form of a model called in the literature: model or user profile (Gunes, Jan, & Piccardi, 2004; Berisha-Bohé & Rumpler, 2007).

It is about a database of the user which presents its particulars, his centers of interests and his preferences.

Several works (Boughanem, Chrisment, & Tamine, 2007; Eugene Santos & Nguyen, 2009) proposed systems of presentation and classification of the user profile. Some works classified the information which represents the user in two types: short-term information and the others in the long term, other works (De la Rosa, González, & López, 2002; Silveira Netto Nunes, 2008) proposed a profile based on the history of navigation of the user. We distinguished, absence of a value which represents the weight of every characteristic in the user profile.

Some systems (Alimi, Ben Ammar, Gouardères, & Neji, 2010; Brandherm, Heckmann, Schmitz, Schwartz, & VonWilamowitz-Moellendorf, 2005 ; Alimi, Ben Ammar, & Dammak, 2011) based on the notion of interests, preferences and contexts to get the intention of the user in a dynamic way thanks to information extracted from documents relevant. However, the opaqueness of the preferences and the user context slowed down the efficiency of these systems.

Other systems (Dellaert, Polzin, & Waibel, 1996; Pentland & Roy, 1996) integrated the notion of emotion into the definition of the user profile and judged that the facial expression of the user at the time of the research reflects its opinion towards the found result. These systems approved an important degree of satisfaction but they lose them efficiency when the motivation of the user is disorientated further to an external event such as the appearance of an external actor in the working environment which can disrupt the concentration of the user, the sensation of the fatigue, etc.

In this frame, and to remedy these problems, we propose an intelligent information research system by taking into account with human behavior. The proposed system takes into account, besides the classic information representing in the majority of the user profile, besides the emotional state of the user and its motivation with regard to the task to be realized.

Our contribution thus consists in proposing an intelligent system of research for information based on the human behavior. This system takes into account the emotional state, the motivational state and the static characteristics (particulars, preferences user, etc.).

In fact, the proposed system holds its originality of the fact that he proposes a process of gratitude of the behavior of the user during a session of research the role of which is of:

  • 1.

    Get back the emotional state of the user from several indicators throughout the session of research;

  • 2.

    Supply the emotional state final by applying the fusion technique and this by attributing automatically a weight to every value returned by a sensor;

  • 3.

    Identify the user behavior by merging the final emotional state with a value which represents the motivation of the user during the research session;

  • 4.

    This fusion is adopted by an expert system.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 7: 4 Issues (2019): 1 Released, 3 Forthcoming
Volume 6: 4 Issues (2018)
Volume 5: 4 Issues (2017)
Volume 4: 4 Issues (2016)
Volume 3: 4 Issues (2015)
Volume 2: 4 Issues (2014)
Volume 1: 4 Issues (2013)
View Complete Journal Contents Listing