Abstract— Personality and individual differences are effective parameters in human activities such as learning. Thus e-learning environments have begun to consider the personality of the learner in order to enhance his/her performance. In this paper we design a model based on the learner’s personality. We use an expert knowledgebase for selecting suitable Virtual Classmate Agent (VCA). Our expert knowledgebase uses the Mayers-Briggs Type Indicator (MBTI) as personality assessment tools, to select suitable VCA for learner. The VCA selects tactics to interact with the learner and will be able to cooperate intelligently with him/her. The results show that the presence of the VCA leads advancements in the learning process and attractiveness of virtual learning environment. Index Terms— Learning Styles, MBTI Indicator, Personality, Virtual Classmate Agent (VCA), Virtual Learning. I. INTRODUCTION E-learning is usually defined as a type of learning supported by information and communication technology (ICT) that improves quality of teaching and learning. E-learning system is a powerful tool for achieving strategic objectives of the university (teaching, research and serving the society) [3]. E-Learning like all other tools offers advantages such as: access to differentiated online resources, Self-directed learning, and Learning matches learners’ lifestyles, etc. Despite of all the advantages this kind of learning lacks the necessary attractiveness most of the time. It seems that regarding the human characteristics and inserting them in virtual learning environments, it would be possible to show these environments more real. Obviously humans are different in personality and characteristics, and this difference in the characteristics of the individuals is reflected in their daily activities and their works. In the case of education and learning, personality difference between learners plays an important role. The learner’s personality will be effective in his learning style [7]. In virtual learning projects this point should be taken into account that the learners’ personalities are various and the teaching method used for each learner should be different from the other learners. In this paper, section 2 is a review of the previous works and literature. In section 3 personality, learning styles, and MBTI indicator are discussed. Section 4 explains the proposed model. Sections 5 and 6 state the simulation and S.Fatahi is with the Department of Computer Engineering, University of Isfahan, Isfahan, Iran. (e-mail:fatahi@comp.ui.ac.ir ). N. Ghasem-Aghaee is with the Department of Computer Engineering, University of Isfahan, Isfahan, Iran. (e-mail: aghaee@eng.ui.ac.ir ). M.Kazemifard is with the Department of Computer Engineering, University of Isfahan, Isfahan, Iran. (e-mail: kazemifard@eng.ui.ac.ir ). implementation of the VCA respectively and finally Section 7 presents the conclusion and the future works. II. PREVIOUS WORKS In virtual learning systems created up to now, the learner’s emotions received much more attention and the emotional agents were more employed. In a few of these systems personality drew our attention as an independent parameter that some of them are mentioned here: In ERPA architecture by using ID3 algorithm, the learner’s emotional reaction towards an event is predicted (for example taking an exam score) [6]. Chaffar and his colleagues used the Naïve Bayes Classifier method to predict the learner’s emotions [5]. In ESTEL architecture, the Naïve Bayes Classifier method is used to predict the optimized emotional status. In this architecture, in addition to emotion, the learner’s personality is also considered. In this system, a module tries to create and induce an optimized emotional state. For instance, when the learner enters the system, after the identification of learner’s personality, for example extrovert, and recognition of optimal emotional state, such as happiness, an emotion is induced to that learner by showing various interfaces (e.g. music, picture, and etc.) to him [4]. In Passenger software designed by German researchers, cooperative learning methods are used. This software examines a series of emotions for the virtual teacher that is present in the system based on OCC model, and invites the learners to group work. The virtual teacher traces the learners’ activities and helps the learners who are not able to do cooperative activities [13]. Abrahamian and his colleagues designed an interface for computer learners appropriate for the type of their personality using MBTI test and concluded that learning through this interface as a result of using personality characters leads into developments in learning process [1]. In implementation performed by Maldonado and his colleagues, a virtual classmate agent is used. This agent is placed beside the learners, and mostly plays the role of a co-learner and a support. In this project each of the teacher, learner, and classmate has own emotions and the learner’s emotions affected his/her classmate [14]. III. PERSONALITY Personality and individual differences are effective parameters in human activities such as learning. Many definitions are presented for personality that some of them are mentioned here: In Schultz’s view, the unique, relatively constant internal and external aspects of a person’s character that influence his behavior in different situations are called personality [9]. In Brody & Ehrlichmann’s view, personality includes thoughts, feelings, desires, intentions, and action tendencies that contribute to important aspects of individuality [9]. Better to say, personality refers to the sets Design an Expert System for Virtual Classmate Agent (VCA) S. Fatahi , N. Ghasem-Aghaee, and M.Kazemifard Proceedings of the World Congress on Engineering 2008 Vol I WCE 2008, July 2 - 4, 2008, London, U.K. ISBN:978-988-98671-9-5 WCE 2008